Astronomy

What Is Astronomy?

Astronomy is the study of everything beyond Earth's atmosphere: the Moon and planets, the Sun and the other stars, the galaxies, and the universe taken as a whole. The name joins the Greek astron, star, and nomos, law, and that is still the job: to find the rules the sky obeys. It is also a science that can touch almost nothing it studies. Apart from meteorites, about 380 kilograms of Apollo Moon rock and a few precious grams brought back from asteroids, every fact in this book was read from light and other messengers that crossed space to reach us.

It is the oldest science too. A baboon bone from the Lebombo Mountains of southern Africa, around 43,000 years old, carries 29 notches that some researchers read as a count of the Moon's phases; others are unconvinced. By c. 1600 BC the Nebra sky disc, found in Germany, shows a crescent Moon, a full disc and the Pleiades in gold on bronze. Babylonian scribes logged the planets on clay tablets for centuries and learned to forecast eclipses. Egyptian, Chinese, Maya, Polynesian and Aboriginal Australian sky-watchers built their own systems, for calendars, harvests and the navigation of open ocean.

Today the subject runs on several fronts, each with its own tools. Planetary science sends spacecraft such as Juno to Jupiter and the Perseverance rover to Mars. Stellar astrophysics splits starlight into spectra to read a star's temperature and make-up. Galactic astronomy maps our own Milky Way: between 2014 and 2025 ESA's Gaia satellite measured the positions and motions of nearly two billion stars. Cosmology treats the universe as a single object, from the Big Bang's afterglow charted by the Planck satellite to the three-dimensional galaxy maps of DESI in Arizona.

Light no longer works alone. Neutrinos from Supernova 1987A were caught in underground detectors hours before the explosion was seen, and on 14 September 2015 the LIGO detectors in the United States registered gravitational waves from two black holes merging over a billion light-years away. The James Webb Space Telescope, launched on Christmas Day 2021, and the Vera C. Rubin Observatory in Chile, which released its first images in 2025, push the reach of the eye further each year.

Did you know? Everything any telescope has ever seen, from planets to galaxies, is made of ordinary atoms that account for only about 5 per cent of the universe; the rest is dark matter and dark energy.

Astronomy is not astrology, though the two shared practitioners until the 17th century; Johannes Kepler cast horoscopes to pay the bills. Its practical gifts are real: calendars, navigation, the law of gravity, and even a piece of Wi-Fi, patented in the 1990s by Australian engineers who borrowed techniques from radio astronomy. Amateurs still find comets and supernovae. Anyone who looks up is doing the oldest experiment there is.

The Celestial Sphere

Stand in a field at night and the sky looks like the inside of a vast dome. Greek thinkers such as Eudoxus, around 370 BC, took that impression literally and pictured the stars fixed to a crystal shell. We now know the stars lie at wildly different distances, but the idea survives as the celestial sphere: an imaginary globe of infinite size centred on Earth, onto which every object is projected. It is a fiction, and an extremely useful one.

Project Earth onto the sphere and you get its map. Extend the planet's axis and it pierces the sky at the north and south celestial poles. Earth's equator, pushed outwards, becomes the celestial equator. The point directly overhead is the zenith, the line where sky meets ground is the horizon, and the meridian runs from north to south through the zenith. Across all of this the Sun traces a yearly path called the ecliptic, tilted 23.4° to the celestial equator and crossing it at the two equinoxes.

Positions work like latitude and longitude. Declination runs from +90° at the north celestial pole to −90° at the south. Right ascension is measured eastwards in hours, from 0 to 24, because the sky turns 15° every hour; its zero is the March equinox point. Sirius sits at about 6 hours 45 minutes and −16.7°. Polaris lies about two-thirds of a degree from the north pole, a little more than the width of the full Moon. The height of the pole above your horizon equals your latitude: from London Polaris hangs 51.5° up, from the equator it sits on the horizon, and the south celestial pole has no bright star at all, only faint Sigma Octantis.

The sphere also drifts. Around 130 BC Hipparchus compared his star positions with records made a century and a half earlier and found that they had all shifted. The cause is precession: Earth's axis wobbles like a spinning top, tracing a circle every 25,800 years or so. Thuban in Draco was the pole star when the pyramids of Giza were built, and bright Vega will come close to the role around AD 14,000. The March equinox point, still called the First Point of Aries, now lies in Pisces.

Did you know? Because the sky's grid slowly slides, every star catalogue must state a date; today's standard, J2000, describes the heavens as they stood at noon on 1 January 2000.

Since 1998 the official frame has been anchored not to stars at all but to hundreds of distant quasars, pinpointed by radio dishes thousands of kilometres apart, and Gaia tied it to visible light. Every telescope mount, star chart and planetarium app relies on that frame. The sphere is imaginary, but the map is exact enough to find a galaxy 10 billion light-years away on the first try.

Constellations

To an astronomer a constellation is not a stick figure but a piece of territory. In 1922 the International Astronomical Union adopted an official list of 88, and the Belgian astronomer Eugène Delporte then drew their borders along lines of right ascension and declination; the boundaries were approved in 1928 and published in 1930. Every point of the sky now belongs to exactly one constellation, so a newly found comet, nova or galaxy arrives with a ready-made address.

The names are far older than the borders. Forty-eight come from Ptolemy's Almagest, written in Alexandria around AD 150 and drawing on Greek and Babylonian lore. Dutch navigators charted the southern sky in the 1590s and added a dozen, including the Phoenix and the Toucan. In 1751–52 Nicolas-Louis de Lacaille catalogued the southern stars from the Cape of Good Hope and invented 14 more, named after the instruments of the Enlightenment, such as Telescopium and Microscopium. He also broke Ptolemy's unwieldy ship Argo Navis into Carina, Puppis and Vela.

A constellation is a trick of perspective. The stars of Orion look like a single figure but lie at very different depths: Bellatrix is about 250 light-years away, Betelgeuse about 550, Rigel about 860, and Alnilam, the middle belt star, well over 1,000. We know because the Hipparcos and Gaia satellites measured each star's parallax, the tiny annual shift caused by Earth's orbit. Seen from another star system, the hunter would fall apart. Many familiar shapes are not constellations at all but asterisms: the Plough, or Big Dipper, is only part of Ursa Major, and the Summer Triangle borrows its corners from three different constellations.

The figures also move. Five of the Plough's seven stars belong to the Ursa Major Moving Group, travelling together through space, while Dubhe and Alkaid head elsewhere; over roughly 100,000 years the saucepan will bend out of shape. Sizes vary enormously. Hydra, the largest constellation, covers 1,303 square degrees, while Crux, the Southern Cross, covers just 68, yet it appears on the flags of Australia, New Zealand, Brazil, Papua New Guinea and Samoa.

Did you know? The Sun spends longer each year in Ophiuchus, the Serpent Bearer, than in neighbouring Scorpius, yet Ophiuchus has never been a sign of the zodiac.

Other cultures carved up the same sky differently. Aboriginal Australians trace the Emu in the Sky in the dark dust lanes of the Milky Way, from the Southern Cross to Scorpius, and its position through the year marked the season for gathering emu eggs. Classical Chinese astronomy divided the heavens into some 283 asterisms. The 88 official regions are a convention, but a priceless one: astronomy's postcode system, and the first thing anyone learns when they start to find their way around the dark.

Why the Sky Turns

Every star rises in the east and sets in the west, yet nothing in the night sky is doing the moving. Earth spins from west to east, and we ride along. Measured against the distant stars, one turn takes 23 hours 56 minutes and 4 seconds, a sidereal day. The familiar 24-hour solar day is about four minutes longer because Earth also moves roughly one degree along its orbit each day, so it must turn a little extra to bring the Sun back to the same place. The result is that each star rises about four minutes earlier every night, two hours a month, which is why the winter and summer skies show different constellations.

The speeds are startling. A point on the equator travels at about 1,670 kilometres per hour, and London, at latitude 51.5°, moves at around 1,040. The sky therefore turns 15° every hour. Leave a camera shutter open and the stars draw star trails, concentric arcs around the celestial pole, with Polaris leaving the shortest scratch of all. Stars close enough to the pole never set: from London, Ursa Major and Cassiopeia are circumpolar and can be seen on any clear night of the year.

For most of history the obvious reading was the reverse: a fixed Earth under a turning sky. The direct proof came in 1851, when the French physicist Léon Foucault hung a 28-kilogram ball from a 67-metre wire under the dome of the Panthéon in Paris. A swinging pendulum keeps its plane of swing fixed in space, so the floor turned beneath it. At Paris the plane crept round clockwise by about 11° an hour, completing a circle in roughly 32 hours. At the North Pole it would take one sidereal day; at the equator it would not turn at all. The same spin twists hurricanes anticlockwise north of the equator and clockwise south of it.

Today Earth's rotation is tracked to a fraction of a millisecond by radio telescopes timing distant quasars, and the record shows the planet slowing. Tides raised by the Moon act as a brake, adding roughly 1.8 milliseconds per century to the length of the day; ancient Babylonian and Chinese eclipse records confirm the drift. Clocks have needed 27 leap seconds since 1972, though none since 2016, and the world's timekeepers have agreed to stop adding them by 2035.

Did you know? Fossil corals from about 380 million years ago grew some 400 daily layers per year, so a day back then lasted only about 22 hours.

The turning sky is the reason for night itself, for the rhythm of rising and setting that every calendar counts, and for the circumpolar stars that steered sailors home. Nothing up there is wheeling around us. We are the merry-go-round.

The Seasons

Seasons happen because Earth is tilted. Its spin axis leans 23.4° from upright relative to the plane of its orbit, and it keeps pointing the same way in space, towards Polaris, all year round. In June the northern hemisphere leans towards the Sun; six months later, on the far side of the orbit, the southern hemisphere does. The tilt changes two things at once: how high the Sun climbs, and how long it stays up.

Both matter. At noon in London the Sun stands about 62° above the horizon at the June solstice and only about 15° in December. A low winter Sun spreads the same beam of sunlight over more than three times the ground it covers in June, and its light passes through more atmosphere on the way. London also gets around 16 hours 38 minutes of daylight in midsummer against about 7 hours 50 minutes in midwinter. More hours of stronger light make summer; fewer hours of weak, slanting light make winter.

The most popular myth is that summer comes when we are closest to the Sun. In fact Earth reaches perihelion, its nearest point, around 3 January, at 147.1 million kilometres, and aphelion, its farthest, around 4 July, at 152.1 million. That 3 per cent change in distance alters sunlight by only about 7 per cent, and if distance ruled the seasons, both hemispheres would have summer at the same time. Kepler's second law adds a twist: Earth moves fastest near perihelion, so the northern summer half of the year, from the March to the September equinox, lasts about 186 days and the southern one about 179.

The year has four turning points. At the solstices, around 21 June and 21 December, the noon Sun stands overhead on the Tropic of Cancer or Capricorn, 23.4° from the equator, and beyond the Arctic and Antarctic Circles it never sets or never rises; at the poles each day and night lasts about six months. At the equinoxes, around 20 March and 22 September, day and night are close to equal everywhere. Ancient builders tracked these moments in stone: Stonehenge is aligned on the midsummer sunrise and midwinter sunset, and at Newgrange in Ireland, a tomb built around 3200 BC, the winter-solstice dawn shines down a passage into the inner chamber.

Did you know? Uranus is tilted by 98°, so it rolls around the Sun on its side; each pole gets about 42 years of continuous daylight followed by 42 years of night.

Even Earth's tilt is not fixed. It drifts between 22.1° and 24.5° over about 41,000 years, one of the slow orbital rhythms identified by Milutin Milanković that pace the ice ages, and our large Moon keeps the wobble small. A few degrees of lean decide when crops are sown, when birds migrate and when the ice comes. The tilt, not the distance, writes the calendar.

Phases of the Moon

The Moon makes no light of its own. It reflects about 12 per cent of the sunlight that falls on it, which makes it about as dark as worn tarmac; it only looks bright against the black sky. At every moment the Sun lights exactly half of it. What changes through the month is how much of that lit half is turned towards Earth, and that is all a phase is.

The cycle runs from new Moon to new Moon in 29.53 days, the synodic month. The Moon actually completes an orbit against the stars in 27.32 days, but in the meantime Earth has moved along its own orbit, so the Moon needs about two more days to line up with the Sun again. At new Moon it sits roughly between us and the Sun and is lost in the glare. A thin waxing crescent follows in the western evening sky; the first quarter, half lit, rises around midday and stands high at sunset; the waxing gibbous swells until the full Moon, opposite the Sun, rises at sunset and stays up all night. Then the sequence reverses through waning gibbous, last quarter, rising around midnight, and a waning crescent in the dawn.

A stubborn myth says the phases are Earth's shadow falling on the Moon. They are not: our shadow reaches the Moon only during a lunar eclipse. Another says the Moon has a permanent dark side. It does have a far side, because tidal forces have locked its rotation to its 27.3-day orbit, so the same face always points at us; but the far side gets just as much sunlight. It was first photographed by the Soviet probe Luna 3 in October 1959, China's Chang'e 4 landed there in January 2019, and Chang'e 6 brought back the first far-side samples in June 2024. A gentle rocking called libration lets us see about 59 per cent of the surface over time.

Look at a crescent Moon after sunset and the rest of the disc often glows faintly. Leonardo da Vinci explained this earthshine in his notebook known as the Codex Leicester, around 1510: it is sunlight bounced off Earth, which seen from the Moon would be nearly full when the Moon looks thin to us. How do we know the Moon's distance so well? Since 1969, observatories have fired lasers at reflectors left by Apollo 11, 14 and 15 and timed the echoes, measuring the average 384,400 kilometres to within millimetres.

Did you know? Those laser echoes show the Moon drifting away from Earth by about 3.8 centimetres a year, roughly the rate at which your fingernails grow.

The phases were humanity's first clock. The word month comes from the Moon; the Islamic calendar still runs on twelve lunar months of about 354 days, and the dates of Easter, Passover and Chinese New Year are set by the Moon. The strongest tides arrive near new and full Moon. The Moon changes shape only from where we stand. Geometry, made visible every night.

Eclipses

An eclipse happens when the Sun, Earth and Moon fall into a straight line. In a solar eclipse the Moon passes in front of the Sun, which can only happen at new Moon. In a lunar eclipse Earth's shadow falls across the Moon, which can only happen at full Moon. They do not happen every month because the Moon's orbit is tilted by about 5° to Earth's, so the three bodies line up only when a new or full Moon falls near one of the two crossing points, or nodes. These eclipse seasons come round about every 173 days, giving between four and seven eclipses of both kinds each year.

Total solar eclipses depend on a coincidence. The Sun is about 400 times wider than the Moon and, on average, about 390 times farther away, so the two discs look almost exactly the same size. When the Moon is near its farthest point it is slightly too small to cover the Sun and leaves a blazing ring, an annular eclipse. When it fits, its darkest shadow, the umbra, races across the ground in a track usually less than 270 kilometres wide at well over 1,700 kilometres per hour. Totality at any one spot lasts at most about seven and a half minutes; the longest this century, on 22 July 2009, gave 6 minutes 39 seconds. On 12 August 2026 totality crossed Greenland, Iceland and northern Spain, the first seen from mainland Spain since 1905, and on 2 August 2027 Luxor in Egypt will get more than six minutes. Britain must wait until 23 September 2090.

A lunar eclipse is gentler and safe to watch with the naked eye. Totality can last more than an hour and a half, and the Moon rarely vanishes; it turns coppery red because Earth's atmosphere bends sunlight into the shadow and filters out the blue. The red glow is, in effect, every sunrise and sunset on Earth at once. A solar eclipse, by contrast, must be watched through certified eclipse filters except during the brief moments of totality.

Eclipses have been laboratories for centuries. Babylonian astronomers found the saros, a cycle of 18 years, 11 days and 8 hours after which eclipses repeat. During an eclipse in India in 1868, the French astronomer Jules Janssen recorded a yellow line in the spectrum of the Sun's prominences that Norman Lockyer attributed to a new element, helium. In 1919 expeditions led by Arthur Eddington and Frank Dyson photographed stars beside the eclipsed Sun from Príncipe and Sobral and found their light bent as Einstein's general relativity predicted. Totality also reveals the corona, the Sun's outer atmosphere at one to three million degrees.

Did you know? In 1973 a prototype Concorde raced the Moon's shadow across Africa, giving scientists on board 74 minutes of totality.

The coincidence will not last. The Moon is drifting away by 3.8 centimetres a year, and in several hundred million years it will be too small to cover the Sun. We happen to live in the age of total eclipses, the greatest free show in nature.

Starlight & Magnitude

Astronomers grade brightness with a scale invented more than two thousand years ago. Around 130 BC, tradition says, the Greek astronomer Hipparchus sorted the stars he could see into six classes, with the brightest of "first magnitude" and the faintest of "sixth"; Ptolemy preserved the system in the Almagest. The result is the apparent magnitude scale, and it runs backwards: the smaller the number, the brighter the object, and the very brightest have negative values.

In 1856 the English astronomer Norman Pogson made the scale exact. He noted that a first-magnitude star is roughly a hundred times brighter than a sixth-magnitude one and defined five magnitudes as precisely a factor of 100, so each step is a factor of about 2.512. The scale is logarithmic, which suits it to a sky that spans enormous ranges, and roughly mirrors the way our eyes judge brightness.

On this scale the Sun is −26.7 and the full Moon −12.7: a 14-magnitude gap, which means the Sun outshines the full Moon about 400,000 times. Venus reaches about −4.9 at its best, Jupiter −2.9 and Sirius, the brightest star of the night sky, −1.46. Polaris, at about +2.0, is only around the 48th brightest star, despite its reputation. The naked-eye limit under a truly dark sky is about +6 to +6.5. The Hubble Ultra Deep Field and the deepest images from JWST reach about +30, some four billion times fainter than the faintest star you can see.

Apparent magnitude mixes up two things: how luminous a star really is and how far away it is. So astronomers also use absolute magnitude, the brightness a star would have if placed at a standard distance of 10 parsecs, about 32.6 light-years. The Sun's absolute magnitude is +4.8, which would make it an unremarkable faint star; Rigel's is around −7, tens of thousands of times more luminous. Stars can change, too. Between late 2019 and February 2020 Betelgeuse faded from about +0.5 to around +1.6 in what became known as the Great Dimming; images from Hubble and the Very Large Telescope traced it to a cloud of dust the star had puffed out.

Did you know? Vega was long used as the zero point of the scale, so the handful of stars that outshine it, led by Sirius, Canopus and Arcturus, carry negative magnitudes.

Measurement has moved from the eye to photographic plates, where Henrietta Swan Leavitt at Harvard found in 1912 that certain pulsing stars reveal their true luminosity, and then to electronic detectors. Gaia measured the brightness of nearly two billion stars, and NASA's TESS watches for dips of a fraction of a per cent as planets pass in front of their suns. Combine a star's brightness with its distance and you know how much energy it pours out, the first step in measuring the universe. Hipparchus ranked the sky by eye; we have only made his ranking precise.

Light-Years

A light-year is a distance, not a length of time, whatever science fiction suggests. Light in a vacuum travels at 299,792.458 kilometres per second, and in one year it covers about 9.46 trillion kilometres. Professional astronomers often prefer the parsec, about 3.26 light-years, which comes straight from the geometry used to measure star distances.

The idea depends on two discoveries. In 1676 the Danish astronomer Ole Rømer, working in Paris, noticed that eclipses of Jupiter's moon Io arrived later than predicted when Jupiter was farther from Earth. He concluded that light takes time to travel, the first proof that its speed is finite. The second step was measuring the stars. In 1838 Friedrich Bessel detected the parallax of the star 61 Cygni, the tiny shift in its position as Earth moves around the Sun, and worked out a distance of about 10 light-years; the modern value is 11.4. Since 1983 the metre itself has been defined by the speed of light, and the Gaia satellite has measured parallaxes for some 1.5 billion stars.

Because light takes time to arrive, every view of the sky is a view of the past. The Moon is 1.3 light-seconds away and the Sun 8 minutes 20 seconds, so if the Sun vanished we would not know for over eight minutes. Signals from Neptune take about four hours, and Voyager 1, the most distant spacecraft, is due to pass one light-day from Earth around the end of 2026. Proxima Centauri, the nearest star beyond the Sun, is 4.24 light-years away. We see Betelgeuse as it was around 550 years ago; if it has already exploded, the news is still on its way. The centre of the Milky Way lies about 26,000 light-years off.

The Andromeda Galaxy, faintly visible to the naked eye on a dark autumn night, is about 2.5 million light-years away. The light reaching your eye left it around the time the earliest members of our genus, Homo, were appearing in Africa. Telescopes stretch the gap much further: in 2025 JWST confirmed a galaxy called MoM-z14, seen as it was only about 280 million years after the Big Bang, its light having travelled for more than 13.5 billion years.

Did you know? The edge of the observable universe lies about 46 billion light-years away, far more than its 13.8-billion-year age suggests, because space has kept stretching while the light was in transit.

That stretching means that for the most distant objects, "how far away" has several answers, and astronomers specify which one they mean. For everyday sky-watching, though, the principle is simple. A telescope is a time machine, and the night sky is a gallery of the past, with every star a postcard sent long ago.

Observing Without a Telescope

The best instrument for the night sky is the one you already own, provided you give it time. Step outside from a lit room and your pupils widen within seconds, from about 2 millimetres to as much as 7 in young eyes. The real change is slower and chemical. The rod cells of the retina rebuild their light-sensitive pigment, rhodopsin, over 20 to 30 minutes, until the eye is around a million times more sensitive than in daylight. A single glance at a bright phone screen undoes most of the work, which is why observers carry a dim red torch: rods barely respond to deep red light.

The eye has a further quirk. Its centre, the fovea, is packed with colour-sensing cones and has no rods at all, so faint objects can vanish when you look straight at them. Rods are densest about 20° away from the centre, and the trick of averted vision, looking slightly to one side, can make a faint smudge spring into view. That is how most people first see the Andromeda Galaxy, M31, at magnitude 3.4. The same patience shows the Orion Nebula, M42, as a glow in Orion's sword and resolves the Pleiades, M45, into six or seven stars.

From a truly dark site a few thousand stars are visible at once, out of about 9,000 bright enough for the naked eye across the whole sky. Five planets have been known since antiquity, Mercury, Venus, Mars, Jupiter and Saturn, wandering against the fixed patterns. Meteor showers need no equipment at all: the Perseids peak around 12 August and the Geminids around 14 December, each producing upwards of 100 meteors an hour under ideal skies. The International Space Station crosses the sky in a few minutes, often rivalling Venus in brightness, and thousands of other satellites now drift across the evening sky.

The great enemy is artificial light. In 2016 the World Atlas of Artificial Night Sky Brightness, compiled by Fabio Falchi and colleagues, found that the Milky Way is hidden from more than a third of humanity. A citizen-science study led by Christopher Kyba, using observations from the Globe at Night project, found that skies grew about 10 per cent brighter every year between 2011 and 2022. Observers grade their skies on the nine-step Bortle scale, published by John Bortle in 2001, from class 1, where the Milky Way casts shadows, to class 9, an inner city showing only a few dozen stars. Protected places are fighting back: Galloway Forest Park became Britain's first Dark Sky Park in 2009, and Sark the world's first Dark Sky Island in 2011.

Did you know? The same 2016 atlas found that almost 80 per cent of North Americans and about 60 per cent of Europeans can no longer see the Milky Way from home.

Naked-eye astronomy needs no money, no kit and no qualifications, only a dark place, twenty minutes and a chart. It is how every astronomer in history began. The sky rewards anyone who simply keeps looking.

The First Astronomers

Long before anyone ground a lens, people were measuring the sky with stones, shadows and patience. The first astronomers were farmers, priests and scribes who needed to know when to plant, when the river would flood and when the gods might darken the Sun. Their instruments were the horizon and a long memory, and the result was humanity's first exact science.

At Stonehenge, raised in stages from around 2500 BC, the main axis points to midsummer sunrise in one direction and midwinter sunset in the other. Popular belief credits the Druids, who arrived some two thousand years too late. In Egypt the Great Pyramid of Khufu, built around 2560 BC, is squared to the cardinal points within about four arcminutes, a fifteenth of a degree, probably by sighting stars that circle the pole, and the 365-day civil calendar was tied to the dawn rising of Sirius that heralded the Nile flood. In Germany the Nebra sky disc, buried around 1600 BC, sets the Sun, the Moon and what look like the Pleiades in gold on bronze.

Babylon is where astronomy became arithmetic. From 652 BC scribes kept the Astronomical Diaries, nightly records of the Moon, planets and weather maintained for some six centuries. From such records came the Saros, a cycle of 223 lunar months, about 18 years 11 days and 8 hours, after which eclipses repeat in similar form. Greek astronomers inherited this arithmetic, and the Antikythera mechanism, a shoebox-sized bronze calculator recovered from a shipwreck in 1901 and usually dated to the 2nd or 1st century BC, turned it into gears, with a Saros dial, a 19-year calendar and a pointer for the Moon's phase.

Chinese court astronomers logged the sky as a matter of state. Shang dynasty oracle bones from around 1200 BC record eclipses, and in AD 1054 Song astronomers noted a "guest star" bright enough to see in daylight for 23 days. Its wreckage is the Crab Nebula, M1, still expanding at about 1,500 kilometres per second. Across the Atlantic, Maya astronomers tracked Venus. The Dresden Codex uses a Venus cycle of 584 days against a true average of 583.92, with corrections that kept its tables in step for centuries, and the round El Caracol tower at Chichén Itzá has windows that appear to line up with the planet's extreme setting points.

Did you know? The Antikythera mechanism had at least 30 interlocking bronze gears, and nothing of comparable mechanical complexity is known from anywhere for well over a thousand years afterwards.

We know all this from the stones themselves, surveyed with theodolites and now laser scanners; from thousands of astronomical clay tablets, many held in the British Museum; and from X-ray scans of the Antikythera fragments made in 2005. Every calendar on your phone still runs on a year and a month that skywatchers first measured by eye. The sky was our first clock, and we have never stopped reading it.

Ptolemy's Universe

For some fourteen centuries the best map of the heavens put us at its centre. Claudius Ptolemy, working in Roman Alexandria around AD 150, gathered centuries of Greek and Babylonian observation into a single treatise, the Mathematical Syntaxis. Its Arabic title, al-majisti, "the greatest", became the Latin Almagest, which tells you how it was regarded.

Earth sat motionless in the middle. Around it, in order, moved the Moon, Mercury, Venus, the Sun, Mars, Jupiter and Saturn, with the fixed stars beyond. The trouble was that planets misbehave. Every 26 months or so Mars slows, stops and drifts backwards against the stars for about ten weeks before resuming. Ptolemy's answer was geometry. Each planet rode a small circle, the epicycle, whose centre travelled around a larger one, the deferent. Earth sat a little off the deferent's centre, and the epicycle moved at an even rate not around Earth but around a third point, the equant. Circles on circles produced loops and changing speeds, and the resulting tables matched the sky to within a degree or so for centuries.

The Almagest also held a catalogue of 1,022 stars in 48 constellations, with brightnesses on a six-step scale that survives, refined, as the modern magnitude system. Ptolemy drew on Hipparchus, three centuries earlier, and historians still debate how much of the catalogue he took over. In his Planetary Hypotheses he nested the spheres snugly and put the outer edge of the cosmos at about 20,000 Earth radii, some 130 million kilometres: vast by ancient standards, yet less than Earth's real distance from the Sun.

The book survived because others valued it. Translated into Arabic in 9th-century Baghdad, it was checked and improved by astronomers such as al-Battani, who refined the length of the year, and later by Nasir al-Din al-Tusi at Maragha and Ibn al-Shatir in Damascus, whose devices for removing the equant reappear, strikingly similar, in Copernicus. Gerard of Cremona's Latin translation from the Arabic, made in Toledo around 1175, carried it back into European universities.

Two myths deserve retiring. Medieval scholars did not think Earth was flat; Ptolemy gives arguments for a round Earth and educated Europeans accepted them. Nor did astronomers pile epicycles on epicycles to patch the model before Copernicus; most working tables used one epicycle per planet.

Did you know? Ptolemy's model predicted planetary positions so well that, when Copernicus's Sun-centred version appeared in 1543, its tables were not clearly more accurate.

Ptolemy was wrong at the centre and right about method: describe the sky with mathematics, test it against observation, publish tables others can check. When Mars backs up against the stars, as it does about every two years, we are watching the very puzzle he solved. A useful wrong answer can last fourteen hundred years.

The Copernican Turn

In May 1543 a book printed in Nuremberg reached a dying canon in Frombork, on the Baltic coast of Poland. De revolutionibus orbium coelestium, "On the Revolutions of the Heavenly Spheres", by Nicolaus Copernicus, moved the Sun to the centre and set Earth spinning daily and circling yearly like any other planet. Tradition, based on a letter from his friend Tiedemann Giese, says he saw the first printed copy on the day he died.

The idea was not new. Aristarchus of Samos had proposed a Sun-centred system around 270 BC, known to us mainly from a remark by Archimedes, and Copernicus had circulated a short sketch, the Commentariolus, around 1514. What he finally published was a complete mathematical system, coaxed into print by a young mathematician from Wittenberg, Georg Joachim Rheticus. In it, retrograde motion stopped being a property of the planets and became an effect of perspective: Earth, on its faster inside track, overtakes Mars every 26 months or so, and Mars seems to drift backwards, as a slower car appears to slide back when you pass it on a motorway.

The new arrangement also gave the planets a scale. Because Mercury and Venus never stray more than about 28 and 47 degrees from the Sun, they must orbit inside Earth, and from such angles Copernicus worked out every planet's distance in units of Earth's. His values, Venus at 0.72, Mars at 1.52 and Jupiter at 5.22, are close to modern ones. For the first time the Solar System had proportions rather than merely an order.

Accuracy was another matter. Copernicus kept perfect circles and still needed small epicycles, so his tables were no better than Ptolemy's. There was also a serious objection. If Earth moves, nearby stars should shift to and fro over a year, and no one could see it. Copernicus's answer, that the stars are immensely far away, was right. James Bradley's discovery of the aberration of starlight in 1729 gave direct proof that Earth moves, and in 1838 Friedrich Bessel measured the first stellar parallax: 61 Cygni shifts by about a third of an arcsecond, the width of a coin seen from more than ten kilometres away.

Did you know? Copernicus's remains, buried unmarked in Frombork Cathedral, were identified in 2008 by matching DNA from a skull with hairs found in a book he had owned.

An unsigned preface by the theologian Andreas Osiander presented the book as a calculating device rather than a claim about reality, and it circulated quietly for decades; the Catholic Church placed it on the Index only in 1616, "until corrected". The idea it launched, that we occupy no special place, has since been applied again and again: the Sun is an ordinary star, the Milky Way one galaxy among hundreds of billions. Each time Mars backs up against the stars, you are watching Earth overtake it. Losing the centre gave us the universe.

Galileo's Telescope

The telescope was not Galileo's invention. In October 1608 the spectacle-maker Hans Lipperhey of Middelburg asked the Dutch States General for a patent on a tube that made distant things look near; rivals made the same claim, and the patent was refused. Galileo Galilei, professor of mathematics at Padua, heard the news in 1609, worked out the optics for himself and began grinding lenses: first about three times magnification, then eight or nine, and by the end of the year about twenty.

His first customer was the Venetian state. In August 1609 he showed an instrument of eight or nine times to senators from the bell tower of San Marco, where they could spot ships hours before they reached port, and his salary was doubled. Then he pointed the tube upwards. The design was simple. A convex objective lens gathered light towards a focus about a metre away, and a concave eyepiece intercepted it to give an upright, magnified image. The field of view was tiny, about a quarter of a degree, half the width of the full Moon, and the glass was full of bubbles.

In Sidereus Nuncius, "The Starry Messenger", published in Venice in March 1610, he reported a Moon with mountains, craters and plains, whose heights he estimated from their shadows; thousands of stars invisible to the eye; a Milky Way resolved into countless faint stars; and, from 7 January 1610, four "stars" beside Jupiter that changed places night after night. They were moons, now called Io, Europa, Ganymede and Callisto, names proposed by Simon Marius. Here was a centre of motion that was not Earth.

Late in 1610 came the blow to Ptolemy. Venus shows a full set of phases: small and almost full when far beyond the Sun, large and crescent when close. In Ptolemy's scheme Venus could never appear full, so it must circle the Sun, although the phases also fit Tycho Brahe's compromise, in which the Sun itself circles Earth. Galileo also tracked sunspots, as did Thomas Harriot, Christoph Scheiner and Johannes Fabricius at about the same time.

Did you know? The Englishman Thomas Harriot sketched the Moon through a telescope on 26 July 1609, months before Galileo, but never published his drawings.

The trouble came later. Galileo's Dialogue Concerning the Two Chief World Systems of 1632 led to his trial in 1633 and house arrest at Arcetri until his death in 1642; the muttered "and yet it moves" is almost certainly legend. In 1992 Pope John Paul II formally acknowledged the Church's error. Two of his telescopes survive in the Museo Galileo in Florence, and NASA named its Jupiter orbiter after him. Today a pair of steady binoculars shows the same four moons from a back garden. What changed in 1610 was not the sky but the evidence: from then on, the eyepiece was the court of appeal.

Kepler's Laws

Johannes Kepler had the theory; Tycho Brahe had the numbers. From his island observatory of Uraniborg on Hven, between Denmark and Sweden, Tycho spent two decades from 1576 measuring planetary positions with giant quadrants and sextants to within one or two arcminutes, the best ever achieved with the naked eye. In 1600 Kepler joined him in Prague. Within eighteen months Tycho was dead, and Kepler inherited the data, over the objections of Tycho's heirs.

He had been assigned Mars, the planet whose path strayed most from a circle. Kepler called the campaign his "war with Mars" and filled hundreds of pages with calculation. His best circular model still missed Tycho's observations by up to eight arcminutes, about a quarter of the Moon's width, and Tycho's data were better than that. Rather than fudge, Kepler threw out the circle. The result, published in Astronomia nova in 1609, was two laws. Planets move on ellipses with the Sun at one focus. And a line from the Sun to a planet sweeps out equal areas in equal times, so a planet travels fastest at perihelion, its closest point, and slowest at aphelion.

The ellipses are subtle. Earth's eccentricity is just 0.017: we are about 147.1 million kilometres from the Sun in early January and 152.1 million in early July, so northern winter falls near our closest approach, a reminder that seasons come from tilt, not distance. Earth moves at about 30.3 kilometres per second at perihelion and 29.3 at aphelion, which is why northern winter is a few days shorter than northern summer. Mars, with an eccentricity of 0.093, was lopsided enough to betray the circle.

In Harmonices Mundi, 1619, Kepler found a third law tying the planets together: the square of a planet's period in years equals the cube of its average distance in astronomical units. Jupiter, 5.20 units out, takes 11.86 years, and both 5.20 cubed and 11.86 squared come to about 141. Kepler could not say why, but the regularity cried out for a physical cause, which Newton would supply. His Rudolphine Tables of 1627 were far better than anything before and predicted a transit of Mercury across the Sun in 1631, which Pierre Gassendi duly observed in Paris.

Did you know? Kepler's laws govern every one of the roughly 6,000 exoplanets now confirmed, and NASA's Kepler space telescope, launched in 2009, found more than 2,600 of them.

Kepler's life was hard. He lost his wife and children to disease and spent years defending his mother against a charge of witchcraft. Yet every satellite in Earth orbit and every probe coasting between the planets follows his ellipses to a first approximation. Eight arcminutes, one stubborn mind and the true shape of the heavens.

Newton's Gravity

Kepler described how the planets move; Isaac Newton explained why. In Philosophiae Naturalis Principia Mathematica, published in London in 1687 at the expense of his friend Edmond Halley, Newton set out three laws of motion and a law of universal gravitation: every mass attracts every other with a force proportional to the product of their masses and falling off with the square of the distance between them.

The apple is real enough as a story, told by Newton himself in old age to William Stukeley, but it did not hit him on the head, and the idea grew over two decades rather than in one afternoon at Woolsthorpe during the plague years of 1665 and 1666. The key step was the Moon test. The Moon is about 60 Earth radii away, so if gravity weakens with the square of distance it should feel one 3,600th of the pull an apple feels. An apple drops 4.9 metres in its first second; the Moon, curving around Earth, falls towards us by about 1.4 millimetres each second. The numbers match.

From that single law Newton derived all three of Kepler's laws, explained the tides as the uneven pull of the Moon and the Sun, and predicted that a spinning Earth should bulge at the equator, as French expeditions to Lapland and Peru later confirmed; the equatorial radius is about 21 kilometres greater than the polar. His thought experiment of a cannon on a mountaintop showed why the Moon never lands: fire fast enough, about 7.9 kilometres per second near the surface, and the ground curves away as fast as the ball drops.

The law then made predictions. Halley used it to argue that the comets of 1531, 1607 and 1682 were one object, and forecast its return; it was recovered on Christmas night 1758, after his death. Henry Cavendish measured the gravitational constant with a torsion balance in 1798, in effect weighing the Earth. In 1846 Urbain Le Verrier, analysing the wanderings of Uranus, told Johann Galle in Berlin where to look, and on 23 September Galle found Neptune within about a degree of the predicted spot. Newton also built, in 1668, the first working reflecting telescope, about 15 centimetres long with a metal mirror some 33 millimetres across, free of the colour fringes that plague lenses.

Did you know? Newton's law has one famous failure: Mercury's orbit swings round by 43 arcseconds per century more than it allows, a puzzle solved only by Einstein's general relativity in 1915.

Newtonian gravity still navigates nearly every spacecraft, from the Apollo flights to gravity-assist swings past Jupiter; relativity matters mainly for precision work such as GPS timing. When you watch the Moon rise, you are watching something fall around us at about a kilometre per second. The apple and the Moon obey the same law.

Herschel's Deep Sky

In 1781 a music teacher in Bath doubled the size of the known Solar System. William Herschel, an oboist and organist from Hanover who had settled in England as a young man, ground his own telescope mirrors at his house on New King Street, helped by his sister Caroline. On 13 March 1781, sweeping the sky with a reflector whose mirror was 6.2 inches across, he noticed a "star" that showed a small disc. He first reported a comet, but within months its slow, nearly circular orbit showed it to be a planet, the first found since antiquity, about 19 times farther from the Sun than Earth.

Herschel proposed calling it Georgium Sidus, George's Star, after King George III, who rewarded him with a royal stipend that let him give up music. Astronomers abroad preferred a mythological name, and Johann Bode's suggestion, Uranus, father of Saturn, became universal by the mid-1800s. Herschel went on to find its moons Titania and Oberon in 1787, and Saturn's moons Enceladus and Mimas in 1789 with his new 40-foot telescope at Slough, whose mirror was about 1.2 metres across.

His real target was the deep sky. With Caroline at a desk recording the positions he called out from the eyepiece, Herschel swept the heavens strip by strip and published three catalogues, in 1786, 1789 and 1802, listing about 2,500 nebulae and star clusters, where Charles Messier's famous list had about a hundred. By counting stars in sample fields, his "star gauges", he drew the first map of the Milky Way's shape, a flattened disc, though with the Sun wrongly near its middle. He showed that some double stars orbit one another, proof that gravity works beyond the Solar System, and in 1800, testing filters for viewing the Sun, he placed thermometers beyond the red end of a prism's spectrum and found invisible heat: infrared light.

Caroline was an astronomer in her own right. She discovered eight comets between 1786 and 1797 and several nebulae, among them the Sculptor Galaxy, NGC 253, and in 1828 the Royal Astronomical Society awarded her its Gold Medal. William's son John Herschel took the family telescopes to the Cape of Good Hope from 1834 to 1838 to survey the southern sky, and his General Catalogue of 1864 listed 5,079 objects. Expanded by J. L. E. Dreyer, it became the New General Catalogue of 1888, the NGC whose numbers astronomers still use.

Did you know? In 1787 George III granted Caroline Herschel a salary as her brother's assistant, making her, as far as we know, the first woman ever paid to do science.

ESA's Herschel Space Observatory, which flew from 2009 to 2013 with a 3.5-metre mirror, carried William's infrared discovery into orbit. Look at the Orion Nebula or the Andromeda Galaxy in a star atlas and you will find a Herschel catalogue number beside it. Three Herschels turned the sky from a dome of lights into a place with depth.

The Great Debate

On the evening of 26 April 1920, at the Smithsonian's natural history museum in Washington, two astronomers addressed the National Academy of Sciences on "The Scale of the Universe". Harlow Shapley of Mount Wilson Observatory argued that the Milky Way was enormous and contained everything, the spiral nebulae included. Heber Curtis of Lick Observatory argued that our galaxy was modest and that the spirals were separate "island universes", galaxies in their own right.

Despite its name, the Great Debate was no formal duel: each man gave a talk, and the sharper exchange came in papers published the following year. Shapley's evidence came from globular clusters, which he had found bunched towards Sagittarius. If they outline the Galaxy, its centre lies far from the Sun and the whole system spans perhaps 300,000 light-years. He also leaned on Adriaan van Maanen's measurements, later shown to be wrong, which seemed to show spirals visibly rotating; if they were remote galaxies, their edges would have to move faster than light.

Curtis pointed to Vesto Slipher's spectra, which showed spirals receding at hundreds of kilometres per second, faster than anything known in our galaxy; to dark dust lanes across edge-on spirals; and to novae in the Andromeda nebula far fainter than those nearby, implying great distance. But he placed the Sun near the centre of a galaxy only about 30,000 light-years wide.

The ruler that settled the matter had been made a decade earlier by Henrietta Swan Leavitt, one of the women "computers" measuring photographic plates at Harvard College Observatory. Studying variable stars in the Small Magellanic Cloud, all at roughly the same distance, she showed in 1912 that the longer a Cepheid takes to pulse, the brighter it truly is. Calibrate a few nearby Cepheids and every Cepheid becomes a standard candle. On 6 October 1923, with the 100-inch Hooker telescope, Edwin Hubble found a Cepheid in the Andromeda nebula, M31, on a plate he marked "VAR!". It put Andromeda at about 900,000 light-years, far beyond any possible edge of the Milky Way, and the result was read to the American Astronomical Society on 1 January 1925.

Did you know? In 1925 the Swedish mathematician Gösta Mittag-Leffler wrote to Henrietta Leavitt about nominating her for a Nobel Prize, unaware that she had died of cancer in 1921.

Both men were half right. Shapley correctly put the Sun far from the centre, about 26,000 light-years out, and correctly made the Galaxy big, though too big: its bright disc is about 100,000 light-years across. Curtis was right that the spirals are galaxies. Andromeda is in fact about 2.5 million light-years away; Hubble's figure was low because, as Walter Baade showed in 1952, there are two kinds of Cepheid. On a dark autumn night Andromeda is a faint smudge to the naked eye. One flickering star turned that nebula into a galaxy.

Hubble's Expanding Sky

Once the spirals were known to be galaxies, the next question was what they were doing, and the first answers had already arrived. From 1912, at Lowell Observatory in Arizona, Vesto Slipher took long-exposure spectra of spiral nebulae. Andromeda was approaching at about 300 kilometres per second, but nearly all the others showed redshift, their spectral lines stretched to longer wavelengths, meaning they were moving away, some at more than 1,000 kilometres per second.

In 1929 Edwin Hubble, working with the 100-inch Hooker telescope on Mount Wilson, then the largest in the world, combined velocities, mostly Slipher's, with his own distance estimates for 24 galaxies. The points were scattered, but the trend was plain: the farther a galaxy, the faster it recedes. His colleague Milton Humason, a former mule driver and janitor who became a superb observer, pushed the measurements outward, and by 1931 their joint paper reached galaxies receding at about 20,000 kilometres per second. The straight line held.

The relation is written v = H₀ × d. Hubble's slope was about 500 kilometres per second per megaparsec, a megaparsec being 3.26 million light-years. That was far too high, because his distances were too small by a factor of about seven, and taken at face value it implied a universe only some 2 billion years old, younger than Earth's rocks. Modern measurements give between about 67 and 73, a disagreement now called the Hubble tension. At 70, a galaxy 100 megaparsecs away recedes at about 7,000 kilometres per second, and running the expansion backwards gives an age of roughly 14 billion years.

Hubble himself was cautious, preferring to speak of "apparent velocities". The theory came from others. Alexander Friedmann had shown in 1922 that Einstein's equations allowed an expanding universe, and in 1927 the Belgian priest and physicist Georges Lemaître derived the velocity–distance relation and estimated its slope, in a paper in a little-read Belgian journal. In 2018 the International Astronomical Union recommended calling it the Hubble–Lemaître law. Einstein, who had added a cosmological constant in 1917 to keep his universe static, visited Mount Wilson in 1931 and accepted the expansion.

The expansion is not galaxies flying away from a central explosion. Space itself stretches, so observers in every galaxy see the same pattern and none sits at a centre, while gravity holds smaller systems together: the Milky Way, the Solar System and you are not expanding.

Did you know? Andromeda is one of the few galaxies with a blueshift: it is closing on the Milky Way at about 110 kilometres per second, and the two may merge billions of years from now.

Every redshift logged by today's surveys, from the Sloan Digital Sky Survey to Euclid, builds on that scattered plot of 1929. Light from a distant galaxy arrives stretched by the very space it crossed. Look at any faint galaxy and you are watching the universe grow.

Astronomy Goes Multiband

For most of history astronomy used one band of light, the one our eyes evolved for: wavelengths from about 380 to 750 nanometres, less than one octave of an electromagnetic spectrum that runs from radio waves kilometres long to gamma rays smaller than an atomic nucleus. Earth's atmosphere opens just two wide windows, for visible light and radio, plus slivers of infrared. Everything else must be caught from mountaintops, balloons, rockets or orbit.

The first step beyond the rainbow came in 1800, when William Herschel found infrared heat past the red end of a spectrum, and Johann Ritter found ultraviolet beyond the violet a year later. Radio astronomy began by accident. In 1932 Karl Jansky, a Bell Labs engineer at Holmdel, New Jersey, chasing static that troubled transatlantic radio telephones, found a steady hiss at a wavelength of 14.6 metres that rose and set with the stars and peaked towards Sagittarius, the centre of the Milky Way. Grote Reber built a 9-metre dish in his Illinois back garden in 1937 and mapped the radio sky, and Bernard Lovell's 76-metre dish at Jodrell Bank, completed in 1957, famously tracked the rocket that launched Sputnik.

In 1964 and 1965, also at Holmdel, Arno Penzias and Robert Wilson found an unexplained microwave hiss at 7.35 centimetres, equivalent to about 3.5 kelvin and coming from every direction. They scrubbed out the pigeon droppings, "a white dielectric material" as they put it, and the hiss remained. It was the cosmic microwave background, the afterglow of the Big Bang, since measured at 2.725 kelvin by COBE and mapped in detail by Planck. Penzias and Wilson shared the 1978 Nobel Prize in Physics.

High-energy astronomy needed space. In 1962 a rocket experiment led by Riccardo Giacconi found the first X-ray star, Scorpius X-1, and NASA's Chandra X-ray Observatory, launched in 1999, sees gas at millions of degrees in galaxy clusters and around black holes. American Vela satellites, built to watch for nuclear tests, caught the first gamma-ray bursts in 1967, and the Fermi telescope has mapped the gamma-ray sky since 2008. In the infrared, IRAS in 1983 and Spitzer from 2003 to 2020 led to the James Webb Space Telescope, launched on 25 December 2021, whose 6.5-metre mirror works about 1.5 million kilometres from Earth.

Did you know? ALMA, an array of 66 radio antennas on Chile's Chajnantor plateau at about 5,000 metres, can work as a single telescope up to 16 kilometres across.

Each band shows different physics. Radio traces cold hydrogen and the jets of quasars; infrared pierces dust into stellar nurseries and catches the most distant galaxies, their light stretched by expansion; X-rays expose the hottest gas. The Crab Nebula, M1, shines in every band from radio to the most energetic gamma rays. Modern astronomy is multiband by default, and the view through your own eye, lovely as it is, is a single key on a very long keyboard.

The Sun's Family

The solar system is one star and everything its gravity holds. The Sun carries 99.86 per cent of the total mass; the eight planets, five recognised dwarf planets, more than 400 moons and millions of asteroids and comets share the remaining sliver. Even that sliver is lopsided. Jupiter alone is 318 times the mass of Earth, about two and a half times all the other planets put together, and together with Saturn it makes up over nine-tenths of the mass of all the planets.

The family sorts itself by distance. Four small rocky planets, Mercury, Venus, Earth and Mars, sit within about 1.5 astronomical units of the Sun, one AU being Earth's average distance of 149.6 million kilometres. Beyond the asteroid belt come the gas giants Jupiter and Saturn, then the ice giants Uranus and Neptune at 19 and 30 AU. Past Neptune lie the Kuiper Belt, home of Pluto, and far beyond it the Oort Cloud of comets, which may stretch to 100,000 AU. All eight planets orbit in the same direction and in nearly the same plane, tilted to one another by no more than Mercury's 7 degrees.

That flatness is a fossil. It records a spinning disc of gas and dust, the solar nebula, that collapsed 4.57 billion years ago, an idea first sketched by Immanuel Kant in 1755 and Pierre-Simon Laplace in 1796. The date comes from the oldest grains in meteorites such as Allende, which fell on Mexico in 1969, and the picture is no longer theory alone: in 2014 the ALMA telescope array in Chile photographed rings and gaps in the disc around the young star HL Tauri, a solar system caught in the act of forming.

Measuring the family took centuries. Copernicus put the Sun at the centre in 1543, and Kepler's laws gave the proportions of the orbits but not their size. In 1672 Giovanni Cassini and Jean Richer timed Mars from Paris and French Guiana at once and turned the angle into a distance. Radar echoes bounced off Venus in 1961 fixed the scale more precisely than ever before, and in 2012 the AU was defined as exactly 149,597,870,700 metres. The masses follow from orbits: how fast a moon circles its planet tells you how heavy the planet is.

Did you know? The popular claim that all eight planets fit between Earth and the Moon is only half true: side by side they span about 392,000 km, more than the average lunar distance, so they fit only when the Moon is near its farthest, around 405,000 km.

Our own position is ordinary in one way and lucky in another. Earth sits where sunlight keeps water liquid, inside a family whose giants have steered comets and asteroids for 4.5 billion years. Every planet we can see with the naked eye, five of them, is a sibling made from the same cloud.

Mercury

Mercury is the smallest planet and the closest to the Sun, a grey, cratered ball 4,879 kilometres across, barely larger than our Moon. It races around its orbit in 88 days at an average of 47 kilometres a second, faster than any other planet, which is why the Romans named it after their fleet-footed messenger. The orbit is the most stretched of the eight: Mercury swings in to 46 million kilometres from the Sun and back out to 70 million, so the Sun in its sky grows and shrinks as the year turns.

Its rotation is stranger still. For decades astronomers assumed Mercury kept one face to the Sun, as the Moon does to Earth; Giovanni Schiaparelli drew maps on that basis in the 1880s. In 1965 Gordon Pettengill and Rolf Dyce bounced radar off the planet with the giant Arecibo dish in Puerto Rico and found it spins once every 58.6 days. That is exactly two-thirds of its year, a 3:2 spin-orbit resonance: three turns for every two laps. Combine the two motions and one solar day, from noon to noon, lasts 176 Earth days, two full Mercury years.

With almost no atmosphere to hold heat, the surface swings between extremes. At noon near the equator it reaches about 430°C, hot enough to melt lead; through the long night it falls to around minus 180°C. Yet Mercury is not the hottest planet; Venus, almost twice as far out, is hotter under its blanket of carbon dioxide. And at Mercury's poles, crater floors never see sunlight at all. Radar from Arecibo hinted at ice there in 1991, and NASA's MESSENGER orbiter confirmed it in 2012.

Did you know? Averaged over time, Mercury rather than Venus is the planet closest to Earth, and in fact the closest planet to every other planet, because it never strays far from the Sun.

Most of what we know comes from three spacecraft. Mariner 10 flew past three times in 1974 and 1975. MESSENGER orbited from 2011 until it ran out of fuel and crashed into the surface in 2015, revealing a huge iron core filling about 85 per cent of the planet's radius, a magnetic field, and cliffs hundreds of kilometres long that show Mercury has shrunk as its interior cooled. The European and Japanese BepiColombo mission, launched in 2018, is due to settle into orbit in late 2026. Mercury also helped prove Einstein right: its orbit swings round by an extra 43 arcseconds a century, a puzzle Urbain Le Verrier noted in 1859 and general relativity explained in 1915.

Mercury is hard to see because it hugs the Sun, visible only low in twilight, and many people never knowingly spot it. When you do, you are looking at the planet that tested our theory of gravity.

Venus

Venus is Earth's near twin in size, 12,104 kilometres across against our 12,742, with 82 per cent of Earth's mass and a similar rocky make-up. There the resemblance ends. Its surface sits at about 465°C, day and night, pole to equator, hotter than Mercury and hot enough to melt lead. The air pressing down on that surface is 92 times Earth's, the pressure a submarine would feel some 900 metres under the sea. Venus is the hottest planet in the solar system, and the reason is its air.

That atmosphere is 96.5 per cent carbon dioxide, a crushing ocean of gas that lets sunlight in and traps the heat that tries to radiate back out. This runaway greenhouse effect was proposed in the early 1960s, notably by a young Carl Sagan, before any spacecraft arrived. On Earth most carbon is locked in limestone and other rocks; on Venus, lacking oceans to dissolve it, it stayed in the sky. Above the hot lower air lies a deck of sulphuric acid clouds from about 48 to 70 kilometres up, reflecting about three-quarters of the sunlight that strikes it and making Venus the brightest object in our sky after the Sun and Moon.

The planet turns the wrong way, and slowly. One rotation takes 243 Earth days, longer than its 225-day year, and runs backwards compared with most planets, so the Sun rises in the west. The cloud tops ignore this lethargy, whipping around the planet in about four days in a pattern called superrotation.

Did you know? Venus's surface has been photographed only by Soviet landers; Venera 13 survived 127 minutes in 1982 and sent back the first colour pictures of its orange-lit, slab-strewn ground.

How we know comes from a long list of firsts. Galileo saw Venus show phases like the Moon in 1610, strong evidence that it circles the Sun. NASA's Mariner 2 made the first successful flyby of another planet in 1962 and measured the heat. The Soviet Venera programme dropped probes through the clouds: Venera 7 sent the first signal from another planet's surface in 1970 and Venera 9 the first pictures in 1975. NASA's Magellan orbiter mapped about 98 per cent of the surface by radar between 1990 and 1994, revealing volcanic plains, the great volcano Maat Mons and surprisingly few craters, a sign of a surface repaved within the last billion years. In 2020 a team reported phosphine in the clouds as a possible sign of life; follow-up work has largely failed to confirm it.

New missions are planned for the 2030s: NASA's VERITAS and DAVINCI and ESA's EnVision. Their question matters here. If Venus once had oceans, as some models suggest, it shows how a world like ours can go wrong. The bright evening star is a warning hung in the western sky.

Earth & Moon

Earth is the only planet known to have liquid water on its surface, active plate tectonics and life, and it has a companion unusually large for its size. The Moon is 3,474 kilometres across, about 27 per cent of Earth's width, and orbits on average 384,400 kilometres away, a gap you could fill with 30 Earths in a row. Only Pluto and Charon are closer in size among major pairs, which is why some astronomers call Earth and Moon a double world.

The leading explanation for the Moon's birth is the giant impact hypothesis, proposed in the mid-1970s by William Hartmann and Donald Davis and, independently, by Alastair Cameron and William Ward. Around 4.5 billion years ago a Mars-sized body, now called Theia, struck the young Earth a glancing blow. Molten debris thrown into orbit gathered into the Moon. The evidence came home in Apollo sample bags: Moon rocks have almost the same oxygen isotope ratios as Earth rocks, they are poor in iron and water, and they show that the young Moon was covered by an ocean of magma. The details are still argued over, but no rival idea fits as well.

The Moon shapes life here. Its gravity raises two tidal bulges in the oceans, and as Earth spins beneath them friction acts as a brake. Our day is slowly lengthening and the Moon is drifting away, about 3.8 centimetres a year, roughly the speed fingernails grow. We know because laser pulses fired from observatories bounce off reflectors left by Apollo 11, 14 and 15 and two Soviet Lunokhod rovers, timing the round trip of about 2.5 seconds to within millimetres. The Moon also steadies Earth's 23.4-degree tilt, keeping our seasons fairly stable over millions of years.

Did you know? Growth rings in fossil shells and corals record ancient day lengths: about 400 million years ago a day lasted roughly 22 hours, and in the last age of the dinosaurs about 23 and a half.

Exploration has come in waves. The Soviet Luna 3 photographed the far side in 1959. In December 1968 Apollo 8's crew saw the first Earthrise, and between 1969 and 1972 twelve Apollo astronauts walked on the surface and returned 382 kilograms of rock and soil. China's Chang'e 4 made the first landing on the far side in 2019, and Chang'e 6 brought far-side samples home in 2024. NASA's Artemis II is the first crewed flight to the Moon since Apollo 17, a step toward landing again near the south pole, where shadowed craters hold water ice.

Without the Moon, Earth's tilt might wobble wildly and tides would be far weaker. The pale disc that lights our nights is part of the machinery that made this planet a stable home.

Mars

Mars is a cold desert world half Earth's width, 6,779 kilometres across, with 38 per cent of our gravity and a day about 40 minutes longer than ours. Its red colour is literal rust: iron oxide dust coats almost everything, and storms can lift it into planet-wide haze. Its air is 95 per cent carbon dioxide but so thin, around 6 millibars, that surface pressure is only about 0.6 per cent of Earth's. Liquid water would boil and freeze at once in the open. Yet the planet is covered in the marks of water that once flowed.

The scenery is the grandest in the solar system. Olympus Mons, a shield volcano, rises about 22 kilometres above the surrounding plains, roughly two and a half times the height of Everest, and its base is about 600 kilometres across. Valles Marineris, a system of canyons, runs about 4,000 kilometres along the equator, nearly the width of the United States, and drops up to 7 kilometres deep; the Grand Canyon would be a side gully. Both were revealed by NASA's Mariner 9, the first spacecraft to orbit another planet, in 1971.

Mars has long attracted wishful seeing. In 1877 Giovanni Schiaparelli mapped dark lines he called canali, channels, and Percival Lowell turned them into canals built by Martians. Mariner 4's grainy flyby pictures in 1965 showed a cratered, Moon-like world instead. The same year that Schiaparelli drew his lines, Asaph Hall found the two small moons, Phobos and Deimos, at the US Naval Observatory.

Did you know? Mars has seasons and polar caps much like Earth's because its axis is tilted by 25 degrees, close to our 23.4, though each season lasts nearly twice as long in a 687-day year.

Robots have since read the rocks closely. Vikings 1 and 2 landed in 1976. Curiosity, in Gale crater since 2012, found that an ancient lake once offered water and chemistry that microbes could have used. Perseverance landed in Jezero crater in 2021, beside the remains of a river delta, and is sealing rock cores in tubes for a possible future return to Earth; its little helicopter Ingenuity made 72 flights. In 2024 it drilled a rock nicknamed Cheyava Falls whose leopard-spot patterns could be a sign of ancient microbes or could be chemistry alone. Orbiters show layered sediments, ancient shorelines and buried ice, and many scientists think Mars had rivers and lakes, perhaps seas, more than three billion years ago before it lost its magnetic field and most of its air to the solar wind.

Mars is the world we are most likely to visit next, and the best place to test whether life ever began twice in one solar system. When it shines orange-red in the night sky, you are looking at a planet that had its chance at being wet and lost it.

The Asteroid Belt

Between Mars and Jupiter, mostly from 2.1 to 3.3 astronomical units from the Sun, circles the main asteroid belt: more than a million catalogued rocky and metallic bodies and countless smaller ones. They are leftovers from the age of planet-building, pieces that never gathered into a world because Jupiter's gravity kept stirring them, raising their speeds so that collisions shattered rather than merged. The old idea that the belt is the wreck of an exploded planet, proposed by Heinrich Olbers in 1802, has long been abandoned. There was never enough material: the entire belt holds only about 3 per cent of the Moon's mass, and the four largest bodies hold about half of that.

The first was found by accident and design together. On 1 January 1801 Giuseppe Piazzi, at Palermo Observatory, spotted a moving point he named Ceres. It was lost behind the Sun, then recovered thanks to an orbit calculated by the 24-year-old Carl Friedrich Gauss. Pallas, Juno and Vesta followed by 1807, and each was briefly called a planet until there were too many. Ceres, 940 kilometres across, became a dwarf planet in 2006.

Jupiter has also carved the belt's structure. In 1866 Daniel Kirkwood noticed gaps where few asteroids orbit, at distances where an asteroid would circle the Sun exactly three times for each Jupiter orbit, or five times for every two. Those regular tugs pile up and throw the rocks out of such orbits, some onto paths that cross Earth's. Jupiter also corrals other groups: the Hildas in a 3:2 rhythm, and the Trojans, two swarms that share Jupiter's orbit 60 degrees ahead and behind.

Did you know? Despite the films, the belt is overwhelmingly empty: spacecraft cross it with no special precautions, and from the surface of a typical asteroid the nearest sizeable neighbour would usually be too far away to see with the naked eye.

Spacecraft have turned points of light into places. Pioneer 10 first crossed the belt safely in 1972. Galileo flew past Gaspra in 1991 and Ida in 1993, finding Ida had its own tiny moon, Dactyl. NASA's Dawn, driven by ion engines, orbited Vesta in 2011 and 2012 and Ceres from 2015 to 2018; it found Vesta's huge Rheasilvia impact basin, source of a family of meteorites that land on Earth, and bright salt deposits in Ceres's Occator crater left by briny water from below. Lucy, launched in 2021, flew past the small belt asteroids Dinkinesh and Donaldjohanson on the way to the Trojans, which it is due to visit from 2027 to 2033, and Psyche is heading for a metal-rich asteroid of the same name.

Most meteorites in museum cases come from the belt, so we hold pieces of it in our hands. They are the unbuilt bricks of the planets, preserving the chemistry of the solar system's first few million years, including, in some, the amino acids that life uses. The belt is not a minefield; it is an archive.

Jupiter

Jupiter is the giant of the planets: 139,820 kilometres across, eleven times Earth's width, and 318 times its mass, about two and a half times all the other planets combined. It is made mostly of hydrogen and helium, like the Sun, and it spins faster than any other planet, once every 9 hours 56 minutes, which flattens it visibly at the poles and draws its clouds into bright zones and darker belts that circle the globe.

There is no surface to stand on. Beneath ammonia clouds whose tops sit near minus 145°C, the hydrogen grows denser and hotter with depth until gas grades into liquid. Thousands of kilometres down, at millions of times Earth's air pressure, it becomes metallic hydrogen, its electrons free to flow, and currents in that churning conductor generate the strongest planetary magnetic field in the solar system. NASA's Juno spacecraft, orbiting since 2016, measured the gravity field closely enough to find that the core is not a neat ball of rock but a dilute core, heavy elements smeared through a large part of the interior, perhaps by a giant impact long ago.

The Great Red Spot is a storm wider than Earth, though it has been shrinking: it was more than twice as long in the late 19th century. It has been watched continuously since at least 1831; whether the spot Giovanni Cassini reported in 1665 was the same storm is doubtful. Jupiter is often called a failed star, but that is a myth. It would need about 80 times its mass to ignite hydrogen fusion. Its role as Earth's shield is also debated: its gravity flings many comets away, but it nudges others inward too. In July 1994 the world watched fragments of Comet Shoemaker-Levy 9 strike Jupiter, leaving dark scars bigger than Earth.

Did you know? Jupiter's moon Ganymede, 5,268 km across, is larger than the planet Mercury and the only moon known to generate its own magnetic field.

When Galileo turned his telescope on Jupiter in January 1610 he found four moons circling it, the first objects ever seen orbiting something other than Earth. Io, the innermost, is the most volcanic body known, with hundreds of active volcanoes heated by tidal squeezing. Europa hides a salty ocean beneath its ice shell, possibly holding twice the water of Earth's oceans. Ganymede and Callisto may hold buried oceans too. NASA's Galileo orbiter studied the system from 1995 to 2003 and dropped a probe into the atmosphere. ESA's JUICE, launched in 2023, is due to arrive in 2031, and NASA's Europa Clipper, launched in 2024, in 2030, to ask whether those oceans could support life.

Jupiter is easy to find, the brightest point in the night sky after Venus for much of the year, and a pair of steady binoculars shows the four Galilean moons as tiny stars in a line. Watch them shift from night to night and you see what Galileo saw.

Saturn's Rings

Saturn's rings are the most spectacular structure in the solar system and one of the flimsiest. The main rings span about 280,000 kilometres from edge to edge, roughly three-quarters of the distance from Earth to the Moon, yet over most of their area they are only about 10 metres thick. Shrink them to the thickness of a sheet of paper and they would still be about 3 kilometres across. They are made of countless separate pieces of nearly pure water ice, from grains of dust to boulders the size of houses, each on its own orbit around the planet.

Seeing what they were took half a century. Galileo's small telescope showed Saturn with strange ears in 1610, and in 1612 the ears vanished, because the rings had turned edge-on to Earth. In 1659 Christiaan Huygens published the answer: a thin, flat ring that touches the planet nowhere. In 1675 Giovanni Cassini found the dark gap now called the Cassini Division. In 1859 James Clerk Maxwell proved mathematically that a solid ring would tear itself apart, so it must be a swarm of particles, and in 1895 James Keeler confirmed it with a spectroscope, showing the inner parts orbit faster than the outer.

The rings are named by letter in order of discovery, which is why they run D, C, B, A, F, G and E outward. The B ring is the broadest and densest; the Cassini Division, about 4,600 kilometres wide, is cleared by a resonance with the moon Mimas; the small moon Pan sweeps the Encke Gap in the A ring; and the narrow F ring is kept in line by two shepherd moons, Prometheus and Pandora. Far outside, the faint E ring is fed by ice geysers erupting from the moon Enceladus, discovered by the Cassini spacecraft in 2005.

Did you know? The rings seem to disappear roughly every 15 years when Earth passes through their plane; the latest such crossing came in March 2025, and Saturn looked almost ringless in small telescopes.

Most of the detail comes from NASA's Voyager flybys of 1980 and 1981 and the Cassini orbiter, which studied Saturn from 2004 to 2017. In its Grand Finale Cassini dived 22 times through the gap between the rings and the planet, weighing the rings at about 40 per cent of the mass of Mimas. Their cleanness suggested they might be young, perhaps a few hundred million years old, though others argue old rings could stay bright, and the debate is open. Saturn is also eating them: ice falls into its atmosphere as ring rain, and on some estimates the rings could be largely gone in a few hundred million years.

Saturn has 274 known moons, the most of any planet, and is so light for its size that it is less dense than water. Even a small telescope shows the rings, and almost everyone who sees them for the first time gasps. We happen to live in the age when Saturn wears them.

The Ice Giants

Uranus and Neptune are a class of their own, the ice giants. Both are about four times Earth's width, 51,118 and 49,528 kilometres across, and around 15 and 17 times its mass. Unlike Jupiter and Saturn they are not mostly hydrogen. Beneath an atmosphere of hydrogen, helium and methane lies a deep, hot, slushy mantle of water, ammonia and methane, the substances planetary scientists call ices even though they are fluid, around a rocky core. Methane in the upper air absorbs red light, which is why both look blue-green.

Uranus was the first planet discovered with a telescope. On 13 March 1781 William Herschel, a musician turned astronomer, spotted it from his garden in Bath and at first took it for a comet. It is the strange one: its axis is tilted by about 98 degrees, so it rolls around the Sun on its side, perhaps knocked over by a giant impact. Over its 84-year orbit each pole gets about 42 years of continuous sunlight followed by 42 years of night. It gives off almost no internal heat, and its atmosphere has the coldest temperatures measured on any planet, about minus 224°C. Its faint rings were found in 1977 when they blocked the light of a star.

Neptune was found with mathematics. Uranus was drifting from its predicted path, and Urbain Le Verrier in France, like John Couch Adams in England, calculated where an unseen planet must be pulling it from. On 23 September 1846 Johann Galle at the Berlin Observatory found Neptune within about one degree of Le Verrier's position. Its orbit takes 165 years, so it completed its first lap since discovery only in 2011. Despite receiving little sunlight, Neptune has the fastest winds in the solar system, around 2,100 kilometres an hour. Its largest moon, Triton, orbits backwards and is probably a captured Kuiper Belt world.

Did you know? Galileo saw Neptune in December 1612 and January 1613, more than two centuries before its discovery, but recorded it in his notes as a faint star.

Each planet has been visited only once, by Voyager 2: Uranus in January 1986 and Neptune in August 1989, when it found the Great Dark Spot, a storm that had vanished by the time the Hubble Space Telescope looked in 1994, and nitrogen geysers on Triton. The famous deep-blue Voyager images of Neptune were contrast-enhanced; a 2024 reanalysis led by Patrick Irwin at Oxford showed both planets are a similar pale greenish-blue. The James Webb Space Telescope now images their rings and moons, and a Uranus orbiter and probe was the top flagship priority of the 2022 US planetary science survey.

Ice giants may be the most common kind of planet in the galaxy, judging by the many Neptune-sized worlds found around other stars, yet ours remain the least explored. The two faint blue dots at the edge of the solar system are a reminder of how much is still unseen.

Pluto & the Kuiper Belt

Beyond Neptune the solar system does not stop; it changes character. The Kuiper Belt, a broad ring of icy bodies roughly 30 to 50 astronomical units from the Sun, holds perhaps 100,000 objects wider than 100 kilometres, leftovers from planet formation that never grew. Pluto is the best known. On 18 February 1930 Clyde Tombaugh, a 24-year-old assistant at Lowell Observatory in Arizona, found it by flicking between photographic plates and spotting a dot that moved. An 11-year-old Oxford girl, Venetia Burney, suggested the name.

Pluto turned out to be small. In 1978 James Christy discovered its moon Charon, half Pluto's width, and the pair's orbits revealed Pluto's mass: less than a fifth of the Moon's. Pluto is 2,377 kilometres across, smaller than our Moon, on a tilted, stretched orbit that takes 248 years and brought it inside Neptune's path between 1979 and 1999. Then the neighbourhood filled in. On 30 August 1992 David Jewitt and Jane Luu found the first other Kuiper Belt object, and on 5 January 2005 Mike Brown, Chad Trujillo and David Rabinowitz spotted Eris in images taken in 2003, about the same size as Pluto and 27 per cent more massive. Either both were planets, or neither. In August 2006 the International Astronomical Union voted in Prague to make Pluto a dwarf planet, because it has not cleared its orbit of other bodies.

The demotion did nothing to dull the place. NASA's New Horizons flew past on 14 July 2015 at 12,500 kilometres and found an active world: a heart-shaped region, Tombaugh Regio, whose western lobe, Sputnik Planitia, is a basin about 1,000 kilometres wide filled with slowly churning nitrogen ice; mountains of water ice several kilometres high; and layers of blue haze. On 1 January 2019 it passed Arrokoth, a 36-kilometre snowman of two lobes that merged gently, the most primitive object ever explored.

Did you know? Pluto has not completed a single orbit since its discovery: its year lasts 248 Earth years, so its first full lap since 1930 ends around 2178.

Farther still, the solar wind ends. Voyager 1 crossed the heliopause, the boundary where the Sun's wind gives way to interstellar gas, in August 2012 at about 122 AU, and Voyager 2 followed in 2018. Beyond lies the Oort Cloud, proposed by Jan Oort in 1950 to explain long-period comets: perhaps trillions of icy bodies from about 2,000 out to 100,000 AU or more. That outer edge is more than a third of the way to the nearest star, Proxima Centauri, at 268,000 AU. No one has ever seen the cloud directly, and at its present speed Voyager 1 would need around 300 years to reach its inner edge.

The Kuiper Belt and Oort Cloud send us comets, and with them a deep-freeze sample of the material the planets were made from. Pluto lost its planet title but gained something better: a reputation as the gateway to the largest region of the Sun's family.

Anatomy of the Sun

The Sun is an ordinary star seen from unusually close. It is a ball of hydrogen and helium plasma 1.39 million kilometres across, 109 times the width of Earth, holding 333,000 Earth masses and about 99.8 per cent of all the mass in the Solar System. It is not on fire. Nothing burns; instead, nuclear fusion in the centre releases about 380 trillion trillion watts, enough every second to cover humanity's present energy use for several hundred thousand years.

The Sun is built in shells. The core, the inner quarter of the radius, reaches about 15 million °C and a density around 150 times that of water. Outside it lies the radiative zone, so dense that light cannot travel straight: each photon is absorbed and re-emitted countless times, and energy takes something like 100,000 years to work its way out. At about 70 per cent of the radius the gas begins to turn over like a simmering pan. In this convective zone hot plasma rises, cools and sinks, and the tops of the cells show on the surface as granules around 1,000 kilometres wide that live for about ten minutes.

What we call the surface, the photosphere, is simply the layer where the gas becomes transparent, a skin a few hundred kilometres thick at about 5,500 °C. Above it the thin chromosphere glows pink during total eclipses, and beyond that the corona streams out for millions of kilometres at one to three million degrees. Why the outer atmosphere is hundreds of times hotter than the surface is the coronal heating problem; the leading suspects are magnetic waves and countless tiny magnetic explosions.

Nobody can look inside, yet the interior is well mapped. In 1960 Robert Leighton found the surface rising and falling with a period of about five minutes, and helioseismology now reads millions of such sound waves, from the ground network GONG and from the SOHO and Solar Dynamics Observatory spacecraft, to measure the depth of the convective zone and how the inside rotates. Neutrinos made in the core fly straight out in about two seconds: Ray Davis first caught them in a tank of dry-cleaning fluid deep in a South Dakota mine from 1968. Close up, NASA's Parker Solar Probe swept to within about 6.1 million kilometres of the surface on 24 December 2024, nearer than any spacecraft before, while the four-metre Daniel K. Inouye Solar Telescope on Maui resolves surface features about 30 kilometres across.

Did you know? Seen from space the Sun is white, not yellow: its light is a near-even blend of all colours, and our atmosphere scatters some of the blue away to leave the familiar yellowish disc.

The sunlight on your face left the photosphere 8 minutes and 20 seconds ago, but the energy it carries was released in the core long before the first farmers sowed a field. The corona's endless outflow, the solar wind, is the weather our satellites fly through. Sunlight is old news by the time it reaches us: tens of millennia in the making, eight minutes in the post.

Nuclear Fusion

Every star is a balancing act. Gravity pulls the gas inward; the heat and pressure of nuclear fusion push back. In the Sun's core, at about 15 million °C and under a pressure some 250 billion times that at sea level, bare hydrogen nuclei, which are simply protons, are squeezed together into helium. The finished helium nucleus weighs about 0.7 per cent less than the four protons that made it, and that missing mass leaves as energy, exactly as Einstein's E = mc² of 1905 demands.

The route is the proton–proton chain. First two protons collide and one turns into a neutron, making deuterium and spitting out a positron and a neutrino. This step relies on the weak nuclear force and is fantastically slow: a given proton waits on average around ten billion years for it. The deuterium then grabs another proton within about a second to become helium-3, and two helium-3 nuclei merge into helium-4, handing two protons back to the pool. Strictly, the protons should never touch at all, since 15 million degrees is far too cool to beat their electrical repulsion. They get through by quantum tunnelling, the effect George Gamow explained in 1928 and Robert Atkinson and Fritz Houtermans applied to stars a year later.

The scale is hard to hold in the mind. The Sun turns around 600 million tonnes of hydrogen into helium every second and loses about four million tonnes of mass as pure energy. Yet the core's power density is feeble, roughly 280 watts per cubic metre. In stars more than about 1.3 times the Sun's mass a different route dominates, the CNO cycle, in which carbon, nitrogen and oxygen nuclei act as catalysts. Hans Bethe worked out both processes in 1938 and 1939 and won the 1967 Nobel Prize for it.

The theory has been tested directly. Arthur Eddington suggested in 1920 that stars shine by turning hydrogen into helium, soon after Francis Aston weighed helium and found it lighter than four hydrogen atoms. Neutrinos from the very first step of the chain were detected by the Borexino experiment under Italy's Gran Sasso mountain in 2014, and CNO neutrinos in 2020, showing that the Sun runs about 1 per cent on that cycle. Some 60 billion solar neutrinos pass through each square centimetre of your skin every second, at night too, because the Earth barely slows them.

Did you know? Litre for litre, a resting human body gives off more heat than the core of the Sun; the Sun outshines us only because its core is around 20,000 times the volume of the Earth.

On Earth, copying the Sun is hard because we cannot borrow its gravity. Laboratories confine plasma with magnets, as at the ITER project under construction in southern France, or crush fuel pellets with lasers: in December 2022 the National Ignition Facility in California released more fusion energy than its lasers delivered to the target, a first. The Sun, about halfway through its ten-billion-year hydrogen budget, has another five billion years to go. Every warm afternoon is a little mass, gone for good, turned into light.

Space Weather

The Sun is not a steady lamp. Its equator spins once in about 25 days and its poles in about 35, so its magnetic field is wound up, tangled and knotted until loops burst through the surface. The knots show as sunspots, regions around 1,500 degrees cooler than their surroundings because the strong field chokes the upwelling heat. Their number rises and falls in the solar cycle of about 11 years, spotted by the German amateur Heinrich Schwabe in 1843. At each cycle the overall field flips, so the full magnetic cycle, found by George Ellery Hale, lasts about 22 years.

When tangled fields snap and reconnect they release solar flares, bursts of X-rays and ultraviolet that cross to Earth in eight minutes. Often they come with coronal mass ejections, clouds of magnetised plasma weighing billions of tonnes, hurled out at hundreds to more than 2,000 kilometres per second, which take from under a day to about four days to reach us. When a CME's magnetic field points against Earth's, it couples into our magnetosphere and drives a geomagnetic storm.

The benchmark is the Carrington event. On 1 September 1859 the English astronomer Richard Carrington, sketching sunspots, saw two beads of white light flare and fade; about 17 hours later the storm hit, auroras glowed as far south as Cuba and telegraph operators took shocks while some lines carried messages on the storm's current alone. In March 1989 a smaller storm collapsed Quebec's power grid in about 90 seconds, leaving six million people without power for nine hours. In February 2022 a modest storm swelled the upper atmosphere enough to drag down 38 of 49 freshly launched Starlink satellites, and on 10 and 11 May 2024 the strongest storm in two decades painted auroras over southern England, Spain and Mexico.

A common picture has solar particles raining straight down to light the aurora. In fact the storm loads energy into the long magnetic tail Earth trails behind it, and electrons accelerated there spiral down the field lines to strike oxygen about 100 to 300 kilometres up, glowing green and, higher still, red, while nitrogen adds blue and purple. Watching for trouble is now routine. SOHO's coronagraphs have seen CMEs leave the Sun since 1996, NOAA's DSCOVR spacecraft at the L1 point, 1.5 million kilometres sunward, gives roughly 15 to 60 minutes' warning of an arriving cloud, and the UK Met Office has run a round-the-clock space weather centre since 2014.

Did you know? Between about 1645 and 1715 astronomers recorded almost no sunspots at all; the lull, now called the Maunder Minimum, overlapped some of the coldest decades of the Little Ice Age.

Cycle 25, which began in December 2019, peaked around late 2024 and is winding down towards a minimum expected near 2030, though big storms often strike on the declining side. For most of us space weather is a gift: green curtains caught on a phone camera from a suburban garden. For grid engineers, airlines and satellite operators it is a forecast checked daily, because the star that warms us also, now and then, throws things.

Stellar Nurseries

Stars are born in the coldest places in the galaxy. Giant molecular clouds, made mostly of hydrogen molecules with a seasoning of dust, sit at 10 to 20 degrees above absolute zero and can hold from ten thousand to a few million times the mass of the Sun, spread across tens to hundreds of light-years. They are thin by earthly standards, a few hundred molecules in each cubic centimetre, emptier than any laboratory vacuum, but there is so much of them that their own gravity matters.

A clump collapses when its gravity beats the pressure of its gas, a threshold worked out by James Jeans in 1902. A nudge helps: the shock of a nearby supernova, the wind of a hot young star or the squeeze of a spiral arm. The densest cores fall inward over around 100,000 years, spinning faster as they shrink, like a skater pulling in her arms, and flattening into a disc around a hot protostar. Material that cannot fall in is flung out along twin jets from the poles, punching glowing knots called Herbig–Haro objects into the surrounding gas. For a star like the Sun, slow contraction then continues for some tens of millions of years until the core reaches about 10 million °C and hydrogen fusion ignites. Clumps below about 8 per cent of the Sun's mass never get there and become brown dwarfs.

Visible light cannot see into a nursery, because dust blocks it, but infrared and radio waves slip through. The Spitzer and Herschel space telescopes mapped clouds threaded with long filaments; the ALMA array in Chile's Atacama Desert photographed the young star HL Tauri in 2014 with rings and gaps already carved in its disc; and since 2022 the James Webb Space Telescope has looked through the dust at jets, cavities and newborn stars in sharp detail.

The best known nursery is visible from a garden. The Orion Nebula, M42, about 1,350 light-years away, is the fuzzy middle "star" of Orion's sword, lit by the hot Trapezium stars at its heart. In the 1990s Hubble revealed dozens of protoplanetary discs, nicknamed proplyds, silhouetted against its glow. In the Eagle Nebula, M16, some 6,500 light-years off, the Pillars of Creation, photographed by Hubble in 1995 and again by Webb in 2022, are columns of gas several light-years tall, eroded at their tips by starlight. A popular claim that a supernova has already destroyed them, based on a 2007 Spitzer study, has never been confirmed.

Did you know? Stars are rarely born alone: most form in clusters, and the Sun's siblings, born from the same cloud 4.6 billion years ago, have long since scattered around the galaxy.

The Milky Way still makes roughly one to two Suns' worth of new stars every year, and most of the newborns are small red dwarfs. Every atom in your body passed through a cloud like these before the young Earth gathered it up. When you find M42 in winter binoculars you are looking at a maternity ward with the lights on.

The Main Sequence

Look at any star in the night sky and the odds are that it is quietly doing what the Sun does: fusing hydrogen into helium in its core. This long, stable phase is the main sequence, and stars spend about 90 per cent of their lives on it. A star joins on the day fusion ignites and stays in rough balance, gravity against pressure, until the hydrogen in its core runs low. The Sun is 4.6 billion years into a stay of about ten billion.

One number decides almost everything: mass. A star twice the Sun's mass is not twice as bright but more than ten times as bright, because luminosity climbs roughly as mass to the power 3.5. Bigger stars therefore burn through their fuel ferociously. A star of ten solar masses lasts some 20 million years; the Sun, ten billion; a red dwarf of a tenth of a solar mass, should shine for several trillion years. Stars range from about 0.08 solar masses, the lower limit for hydrogen fusion, to well over 100.

Astronomers sort stars by their spectra into the classes O, B, A, F, G, K and M, from blue-white stars above 30,000 degrees to red ones below about 3,700, a scheme largely built at Harvard by Annie Jump Cannon, who classified hundreds of thousands of stars by eye. In 1925 Cecilia Payne showed in her doctoral thesis that stars are made overwhelmingly of hydrogen and helium, a result so unexpected that she was persuaded to describe it as probably spurious. Masses come from binary stars: watching two stars orbit each other and applying Kepler's third law weighs them. The European Space Agency's Gaia mission, launched in 2013, has measured distances to nearly two billion stars.

The night sky offers a tour of the sequence. Sirius A, the brightest star, is an A-type star of about twice the Sun's mass shining 25 times as brightly, 8.6 light-years away. The Sun is a modest G2 star. Proxima Centauri, the nearest star at 4.24 light-years, is an M dwarf of about an eighth of the Sun's mass, invisible without a telescope. At the other extreme, the brightest Trapezium star in the Orion Nebula, Theta-1 Orionis C, has around 30 solar masses and will be gone within a few million years. Red dwarfs are the silent majority, about three in every four stars in the Milky Way, yet barely any can be seen with the naked eye.

Did you know? No red dwarf has ever died of old age: their lifetimes run to trillions of years, and the universe is only 13.8 billion years old.

The main sequence also explains why the Sun has been kind to us. A much heavier star would have died before life got going; a much lighter one would offer a dim, flare-prone glow. The Sun has brightened by about 30 per cent since its birth and will keep slowly brightening, but for now it sits near the middle of the table, steady enough for four billion years of biology. Mass is destiny, and ours was just about right.

The HR Diagram

Around 1911 the Danish astronomer Ejnar Hertzsprung, and independently in 1913 the American Henry Norris Russell, plotted the brightness of stars against their colour and found that the points did not scatter at random. They crowded into a few distinct regions. The result, the Hertzsprung–Russell diagram, is astronomy's periodic table: a single chart on which a star's position reveals its size, its stage of life and often its mass.

By convention the horizontal axis shows surface temperature, or spectral class, with hot blue stars on the left and cool red ones on the right, an awkward backwards order kept from history. The vertical axis shows luminosity, usually on a logarithmic scale from a ten-thousandth of the Sun to a million Suns. Because a star's output depends on its surface area and on the fourth power of its temperature, a cool star can be very bright only if it is enormous, so lines of equal radius slant across the chart, with giants at the top right and dwarfs at the bottom left.

About nine stars in ten lie on the main sequence, the diagonal band running from hot, bright stars such as Spica down to dim red dwarfs like Proxima Centauri, with the Sun, at 5,800 kelvin and one solar luminosity, about halfway along. Above it to the right sit the red giants, such as Arcturus and Aldebaran, tens of times the Sun's width. Across the very top run the supergiants: in Orion, red Betelgeuse at about 3,600 kelvin and blue-white Rigel at about 12,000 kelvin both shine with something like 100,000 Suns' worth of light. Far below lie the white dwarfs, hot but tiny. Sirius B, the faint companion of the brightest star in the sky, packs about the Sun's mass into a ball roughly the size of Earth.

The diagram is a snapshot, not a racetrack. A star does not slide down the main sequence as it ages; it sits at a point set by its mass, then leaves, swelling up and to the right towards the giants and ending, for most, among the white dwarfs. Plot a single star cluster, whose members share an age, and the main sequence is eaten away from the top: that turn-off point dates the cluster, and puts the oldest globular clusters at around 12 billion years or more. In 2018 the Gaia mission's second data release plotted more than four million nearby stars with such precise distances that new fine structure appeared, including a split in the white dwarf sequence.

Did you know? In late 2019 and early 2020 Betelgeuse faded to well under half its usual brightness; observations later showed this "Great Dimming" was caused by a dusty cloud the star had puffed out, not a sign of imminent explosion.

For a backyard observer the diagram is hidden in plain sight. Orion alone holds a red supergiant, a blue supergiant and a nursery of newborn stars, and the colours you can see by eye, Betelgeuse's orange against Rigel's icy white, are temperatures you are reading directly. Every star in the sky has an address on one chart.

Red Giants & White Dwarfs

The Sun is in middle age, and its future is unusually well understood, because we can see other stars further along the same road. Even now it is about 30 per cent brighter than at its birth, and in around a billion years it will be roughly 10 per cent brighter still, enough, many models suggest, to drive off Earth's oceans long before anything dramatic happens to the Sun itself.

In about five billion years the hydrogen in the core runs out. The core, now helium ash, contracts and heats while hydrogen burns in a shell around it, and the outer layers swell enormously: the Sun becomes a red giant, cooler at the surface, a couple of thousand times as luminous and perhaps 200 times its present width. Mercury and Venus will be swallowed. Whether Earth follows is genuinely debated, because the swollen Sun will lose much of its mass, loosening its grip and letting the planets drift outward. When the core reaches about 100 million °C, helium ignites in the helium flash and starts fusing into carbon and oxygen, and the star settles down for around 100 million years before swelling once more.

In its last swollen phase the Sun will shed its outer layers in pulses, exposing a hot core whose ultraviolet light makes the expelled gas glow. This is a planetary nebula, a name from William Herschel in the 1780s, because the round glows looked like planets in his telescope; they have nothing to do with planets. The Ring Nebula, M57 in Lyra, is the classic example; the Helix Nebula, NGC 7293, about 650 light-years away, is among the nearest; and the Southern Ring, NGC 3132, was one of the James Webb Space Telescope's first images in July 2022.

What remains is a white dwarf: a cinder of about half the Sun's mass crushed into a sphere the size of Earth, so dense that a teaspoonful would weigh several tonnes. It shines only by stored heat. Gravity is held off not by fusion but by electrons that, by quantum rules, refuse to be packed any closer, a pressure with a ceiling: in 1930 the young Subrahmanyan Chandrasekhar calculated that no white dwarf can exceed about 1.4 solar masses. The first known was Sirius B, inferred by Friedrich Bessel from the wobble of Sirius in 1844, glimpsed by Alvan Graham Clark in 1862. In 2019 Gaia data revealed white dwarfs crystallising as they cool, their carbon and oxygen freezing into solid cores.

Did you know? No white dwarf has yet cooled into a cold, dark "black dwarf": that takes vastly longer than the 13.8 billion years the universe has existed.

About 97 per cent of the stars in the Milky Way, the Sun among them, will end this way. It is the quiet majority fate: no explosion and no black hole, just a long cooling glow surrounded by the drifting remains of its planets. The carbon and oxygen that earlier giants puffed into space are already in your body. The Sun will go out gently, which is the most any of us can ask.

Supernovae

A supernova is a star exploding, and for a few weeks it can rival the light of an entire galaxy of billions of stars. There are two main recipes. In a core-collapse supernova, a star of more than about eight solar masses runs out of fuel and its core implodes. In a Type Ia supernova, a white dwarf in a binary system gains enough mass, from a companion or a merger, to approach the 1.4-solar-mass limit and is destroyed in a runaway thermonuclear blast.

A massive star builds its own bomb. After hydrogen and helium it fuses carbon, neon, oxygen and silicon, each stage faster than the last, until it resembles an onion of burning shells around an iron core. For a star of about 25 Suns, hydrogen burning lasts some seven million years, carbon a few centuries and silicon about a day. Iron is the dead end, because fusing it absorbs energy rather than releasing it. Once the iron core passes about 1.4 solar masses it collapses in well under a second, from roughly the size of Earth to a ball some 20 kilometres across, at up to a quarter of the speed of light. Infalling matter rebounds off the newborn neutron star, and a flood of neutrinos helps drive the shock outward through the star.

Most of the energy is invisible. About 99 per cent leaves as neutrinos, and on 23 February 1987 detectors proved it. When SN 1987A went off in the Large Magellanic Cloud, about 168,000 light-years away, the Kamiokande II detector in Japan, IMB in the United States and Baksan in the Soviet Union caught about two dozen neutrinos in 13 seconds, a few hours before the light was seen. It was the first supernova visible to the naked eye since 1604, and in 2024 the James Webb Space Telescope found strong evidence of the neutron star hidden in its debris.

History is studded with them. The supernova of 1006 in Lupus, the brightest ever recorded, was noted by observers from China and Japan to Egypt and Europe. Tycho Brahe's star of 1572 in Cassiopeia showed that the heavens could change, and Kepler's of 1604 in Ophiuchus is the last seen in our own galaxy. In May 2023 the Japanese amateur Koichi Itagaki discovered SN 2023ixf in the Pinwheel Galaxy, M101, 21 million light-years away, the closest core-collapse supernova in a decade. Type Ia explosions, which peak at nearly the same brightness, serve as standard candles for measuring the expanding universe.

Did you know? A supernova would need to be within roughly 50 light-years to seriously damage Earth's ozone layer, and no star that close is massive enough to explode.

The Milky Way should host one or two core-collapse supernovae a century, most hidden from us by dust, so one is arguably overdue. Betelgeuse, some 550 light-years away, will probably explode within the next 100,000 years, bright enough to see in daylight and perfectly safe. The oxygen you are breathing and the calcium in your bones were made in stars and scattered by explosions like these. We are, quite literally, fallout.

Neutron Stars & Pulsars

A neutron star is what is left when the core of a massive star collapses and the explosion does not finish the job. Squeeze around 1.4 Suns into a sphere some 20 to 25 kilometres across, the size of a city, and electrons are forced into protons until the star is mostly neutrons, packed as tightly as in an atomic nucleus. A sugar cube of the material would weigh hundreds of millions of tonnes, roughly as much as the whole human race.

Walter Baade and Fritz Zwicky proposed such objects in 1934, only two years after the neutron was discovered, but few expected to see one. Then in 1967 Jocelyn Bell, a graduate student at Cambridge working with Antony Hewish on a new radio array covering four and a half acres, noticed a small patch of "scruff" on her chart recordings. It was a pulse repeating every 1.337 seconds, so regular that it was half-jokingly labelled LGM-1, for Little Green Men. Finding more of them in other parts of the sky ruled out aliens: they were pulsars. Hewish shared the 1974 Nobel Prize in Physics with Martin Ryle, and the omission of Bell, now Dame Jocelyn Bell Burnell, is still argued over.

A pulsar is a spinning neutron star whose magnetic field, a trillion times Earth's or more, is tilted away from its spin axis. Charged particles stream along the magnetic poles and beam radiation outward, and each time a beam sweeps across Earth we see a tick, like a lighthouse. The Crab Pulsar, at the heart of the remnant of the supernova of 1054, turns about 30 times a second; PSR J1748−2446ad, in the globular cluster Terzan 5, spins 716 times a second, its equator moving at around a fifth of the speed of light. Magnetars, with fields up to a thousand times stronger still, unleash flares such as the one from SGR 1806−20 on 27 December 2004.

Pulsars are superb clocks, and astronomers use them as instruments. In 1974 Russell Hulse and Joseph Taylor found a pulsar orbiting another neutron star; over the following years the orbit shrank exactly as Einstein's theory predicts if the pair is radiating gravitational waves, earning them the 1993 Nobel Prize. NASA's NICER X-ray telescope on the International Space Station has measured neutron stars about 25 kilometres across, and the heaviest known weigh more than twice the Sun. In 2023 the NANOGrav collaboration, timing dozens of fast-spinning pulsars for 15 years, reported evidence of a background hum of gravitational waves washing through the galaxy.

Did you know? The plaques on Pioneer 10 and 11 and the Voyager golden records show the Sun's position relative to 14 pulsars, each tagged with its period, so that any finder could work out where we live.

More than 3,000 pulsars are now catalogued, many found with great radio dishes such as Parkes in Australia and FAST in China. They are cosmic metronomes, steady enough that they could one day guide spacecraft across the Solar System like a galactic satnav. A dead star, it turns out, keeps excellent time.

Black Holes

A black hole is a region where gravity is so strong that nothing, not even light, can climb out. Its boundary is the event horizon, a point of no return rather than a surface. Its size depends only on its mass: the radius is about 3 kilometres for every solar mass, so a black hole with the Sun's mass would be about 6 kilometres across, and one with the Earth's mass about the size of a marble. At the centre, general relativity predicts a singularity of infinite density, a sign that the theory has reached its limits.

The idea is older than the name. In 1783 the English clergyman John Michell imagined "dark stars" too massive for light to escape. In 1916, Karl Schwarzschild found the first exact solution of Einstein's equations, and it contains the horizon. The term "black hole" was popularised by John Wheeler in 1967. Stellar black holes form when the cores of very massive stars collapse; supermassive ones, of millions to billions of solar masses, sit at the centres of galaxies.

Black holes are seen by what they do to their surroundings. In the early 1970s Cygnus X-1, an X-ray source orbiting a blue supergiant, became the first widely accepted black hole; it is now thought to weigh about 21 Suns. Since the 1990s the teams of Reinhard Genzel and Andrea Ghez have tracked stars whipping round the centre of the Milky Way, including S2, which completes an orbit every 16 years, proving that Sagittarius A\ holds about four million solar masses in a tiny volume; they shared the 2020 Nobel Prize in Physics with Roger Penrose. On 14 September 2015 the twin LIGO detectors recorded GW150914, the gravitational waves from two black holes of about 36 and 29 solar masses merging 1.3 billion light-years away. And in 2019 the Event Horizon Telescope, a planet-wide network of radio dishes, released the first image of a black hole's shadow, in the galaxy M87, followed by Sagittarius A in 2022.

Black holes are not cosmic vacuum cleaners. Far from the horizon their gravity is exactly that of any other object of the same mass; replace the Sun with a black hole of equal mass and Earth's orbit would not change. The danger lies close in. Fall into a stellar black hole and tides would stretch you like spaghetti well before the horizon; fall into a supermassive one and you might cross the horizon unharmed, while distant watchers would see you slow, redden and fade. Stephen Hawking showed in 1974 that black holes should slowly leak radiation, though for stellar ones the glow is far too faint ever to detect.

Did you know? The nearest black hole known, Gaia BH1, lies about 1,560 light-years away and was found in 2022 only by the wobble it gives a Sun-like companion star.

For all their menace, black holes are among the simplest objects in nature, described completely by mass, spin and electric charge. Point binoculars at Cygnus on a summer night and you are looking towards one: invisible by definition, betrayed by the company it keeps.

The Milky Way

On a dark, moonless night a pale river of light arches across the sky. That river is the Milky Way, our own galaxy, seen from inside its disc. Every star you can see with the naked eye belongs to it, and so do between 100 and 400 billion more. The Greeks called it galaxias kyklos, the milky circle, which is where the word galaxy comes from.

Seen from outside, it would be a barred spiral. A central bar of old stars, roughly 10,000 light-years long on each side of the centre, sits inside a flat disc about 100,000 light-years across but only around 1,000 light-years thick: in proportion, thinner than a vinyl record. Spiral arms, the Perseus and Scutum-Centaurus arms chief among them, wind out from the bar, picked out by young blue stars and glowing gas. Around it all lies a near-spherical halo of sparse old stars, about 150 globular clusters and a much larger cloud of dark matter that brings the galaxy's total mass to roughly a trillion Suns, though estimates still vary.

The Sun lives in a modest spur called the Orion Arm, about 26,000 light-years from the centre. It orbits at around 230 kilometres per second, so even at that speed one lap, a galactic year, takes about 230 million years. The Sun has made only about 20 laps since it formed. At the hub sits Sagittarius A\*, a black hole of about four million solar masses, hidden from optical telescopes behind some 30 magnitudes of dust but plain in radio, infrared and X-rays.

Working this out took four centuries. In 1610 Galileo turned his telescope on the milky band and found it was made of countless faint stars. In 1785 William Herschel counted stars in different directions and drew the galaxy as a flattened disc, with the Sun near the middle. In 1918 Harlow Shapley used the distances to globular clusters to show that the Sun is far off-centre, towards the edge of a much bigger system. Radio astronomers mapping the 21-centimetre glow of hydrogen in the 1950s traced the spiral arms through the dust, and between 2013 and 2025 the European Space Agency's Gaia spacecraft measured the positions and motions of nearly two billion stars, revealing a disc that is warped, rippling and still settling from ancient collisions.

Did you know? The last time the Sun stood at this point in its orbit, around 230 million years ago, the very first dinosaurs were only just appearing.

From the city the Milky Way has all but vanished; around a third of humanity can no longer see it at all. Find a dark site in late summer, look towards Sagittarius, and you are looking straight at the crowded heart of the galaxy. It is the only galaxy we will ever see from the inside.

Galaxy Types

Galaxies come in a handful of shapes, and the shape says a great deal about a galaxy's life. In 1926 Edwin Hubble, sorting photographic plates from the 100-inch telescope on Mount Wilson, proposed a simple scheme. A decade later, in his book The Realm of the Nebulae, he drew it as a tuning fork, and astronomers still use its labels a century on.

On the handle sit the ellipticals, smooth balls of stars graded from E0, perfectly round, to E7, a flattened cigar. They hold little cold gas, make few new stars and glow with the reddish light of old ones. The biggest live in the hearts of galaxy clusters: M87 in the Virgo Cluster holds perhaps a trillion stars and a black hole of 6.5 billion suns. Where the fork divides sit the lenticulars, S0, discs with no visible arms, like the Spindle Galaxy. Then the prongs split into spirals, Sa to Sc, from tightly wound arms around a big bulge to loose arms around a small one, and barred spirals, SBa to SBc, whose arms trail from the ends of a straight bar of stars. The Whirlpool Galaxy, M51, is a classic spiral; NGC 1300 is the textbook barred one; and the Milky Way is a barred spiral too. Anything that fits nowhere is an irregular, such as the Small Magellanic Cloud or the battered starburst galaxy M82.

Hubble called ellipticals "early" and spirals "late" types, and the words stuck, but the tuning fork is not a family tree and galaxies do not march along it. The truth runs closer to the reverse: collisions between spirals can scramble their discs into ellipticals. Colour tells the story. Spirals are blue because their gas still forms hot, short-lived stars; ellipticals are "red and dead" because their gas has been used up, blown out or heated too much to collapse.

Classifying galaxies by eye became a public pastime in 2007, when Chris Lintott and colleagues launched Galaxy Zoo. More than 150,000 volunteers made over 50 million classifications in its first year, and the project turned up new kinds of object along the way. Today machine learning sorts the hundreds of millions of galaxies imaged by Euclid and the Vera C. Rubin Observatory, while the James Webb Space Telescope finds surprisingly orderly discs in the young universe.

Did you know? The faintest known galaxies, ultra-faint dwarfs such as Segue 2, may contain only around a thousand stars, fewer than many star clusters.

Dwarfs are by far the most common galaxies of all, though giants like ours hold most of the stars. The fork's real lesson is that shape records history: a calm spiral has lived quietly, an elliptical bears the scars of mergers. Every galaxy wears its biography on its face.

Andromeda

The Andromeda Galaxy, catalogued as M31, is the most distant thing most people can see with the naked eye. From a dark site it is a faint oval smudge in the constellation Andromeda, and its light has travelled about 2.5 million years to reach you: it left before our own species existed. The Persian astronomer al-Sufi described it as a "small cloud" in his Book of Fixed Stars around 964. Long-exposure photographs show that it stretches about three degrees across the sky, six times the width of the full Moon.

For centuries no one knew what it was. Many astronomers thought spiral "nebulae" were gas clouds within the Milky Way. Then, in October 1923, Edwin Hubble found a Cepheid variable, a star whose pulsing rhythm reveals its true brightness, on a plate taken with the 100-inch Hooker telescope. He crossed out the "N" for nova he had first written beside it and wrote "VAR!". The star's faintness put Andromeda far beyond the Milky Way, and the universe suddenly became a place of countless galaxies.

Andromeda is a big spiral, its bright disc roughly 150,000 light-years across, with a vast halo of faint stars and gas reaching a million light-years or more. It holds around a trillion stars, perhaps several times as many as the Milky Way, yet measurements of the two galaxies' dark matter suggest their total masses are similar, each of order a trillion Suns. In 2015 the Panchromatic Hubble Andromeda Treasury, a mosaic made with the Hubble Space Telescope, resolved more than 100 million of its individual stars. Andromeda has its own court of satellites, including the compact elliptical M32 and M110.

Andromeda is approaching us at about 110 kilometres per second. In 2012 Hubble measured its sideways drift and the answer seemed clear: a head-on collision in about 4.5 billion years. In 2025, however, a team led by Till Sawala combined Hubble and Gaia data with the pull of the Triangulum Galaxy and the Large Magellanic Cloud, and found that the uncertainties are large. Their simulations gave only about a fifty-fifty chance of a merger within the next ten billion years. If it happens, the two spirals will settle into one giant elliptical, already nicknamed "Milkomeda".

Did you know? Even if the two galaxies collide, the stars are so far apart that almost none would ever hit each other; the Sun would most likely just be flung onto a new orbit.

Andromeda is the closest large galaxy we can study from outside, and so the best stand-in for a portrait of our own. Look at it through binoculars on an autumn night and you are seeing, more or less, what aliens there would see when they look at us.

Galaxy Mergers

Galaxies are not islands. Pulled together by gravity, they meet, tear at each other and eventually merge, and over billions of years these mergers have built most of the big galaxies we see. A collision unfolds far too slowly to watch, typically over a billion years or more, so astronomers study it the way a naturalist might study a forest: by photographing many galaxies caught at different stages and putting the snapshots in order.

The first act is a close pass. As two galaxies swing by each other, the near side of each is pulled harder than the far side, and these tidal forces fling out long streamers of stars and gas. In 1972 the brothers Alar and Juri Toomre ran simple computer models and showed that the strange tails and bridges in Halton Arp's 1966 Atlas of Peculiar Galaxies were exactly what gravity predicts. The pair known as the Mice, NGC 4676, show this stage beautifully, each trailing a single long tail.

The second act is fireworks. The stars themselves almost never collide, but the vast clouds of gas do. They crash, compress and collapse, and star formation can surge to ten or even a hundred times its normal rate, a starburst. The Antennae Galaxies, NGC 4038 and NGC 4039, about 45 million light-years away, are the classic example: their cores are still separate, but thousands of new star clusters blaze in the wreckage, each a pocket of hot blue stars. Gas funnelled to the centre also feeds the central black holes, which can switch on as quasars. Then, after further passes, the galaxies coalesce. Their discs are scrambled, the orderly orbits replaced by a swarm, and the remnant often settles into a smooth elliptical galaxy. The two central black holes sink together and, finally, merge in a burst of gravitational waves that the planned space observatory LISA is designed to hear.

Our own galaxy is a serial merger. In 1994 Rodrigo Ibata, Gerry Gilmore and Mike Irwin discovered the Sagittarius Dwarf Galaxy being torn apart on the far side of the Milky Way, its stars strewn in a stream that wraps around the sky. In 2018 Gaia data revealed a far older collision: the Gaia-Enceladus galaxy, swallowed around 8 to 11 billion years ago, whose stars still move on telltale elongated orbits through the halo.

Did you know? JWST's very first images, released in July 2022, included Stephan's Quintet, a group of galaxies in which shock waves from a collision are heating gas to millions of degrees.

Mergers explain why the universe is not full of identical spirals. They light starbursts, feed black holes and reshape galaxies wholesale. Some of the atoms in your body were forged in stars born in galaxies that the Milky Way swallowed long ago. We are made, in part, from wreckage.

Supermassive Black Holes

Almost every large galaxy has a supermassive black hole at its centre, holding from about a million to tens of billions of times the mass of the Sun. Ours, Sagittarius A\*, weighs about 4.3 million Suns; its event horizon is about 25 million kilometres across, small enough to fit inside the orbit of Mercury. The giant in M87, 55 million light-years away, weighs 6.5 billion Suns and has a horizon wider than the orbit of Neptune. The heaviest known, such as the one powering the quasar TON 618, are estimated at tens of billions of solar masses.

How do you weigh something invisible? By watching its neighbours. Radio astronomers Bruce Balick and Robert Brown found the compact radio source at the Milky Way's centre in 1974. From the 1990s, Reinhard Genzel's team at the European Southern Observatory and Andrea Ghez's at the Keck telescopes tracked stars circling it. The star S2 sweeps round every 16 years at up to 7,650 kilometres per second, which only works if four million solar masses are packed into a region smaller than the Solar System. Genzel and Ghez shared the 2020 Nobel Prize in Physics.

The most direct evidence is a picture. The Event Horizon Telescope links radio dishes from Hawaii to the South Pole, observing at a wavelength of 1.3 millimetres and timing the signals with atomic clocks so that the network acts as a single dish the size of the Earth. Its first campaign, in April 2017, gathered several petabytes of data, too much for the internet, so the hard drives were flown to processing centres. In April 2019 the collaboration revealed a glowing ring around the dark shadow of M87's black hole, and in May 2022 a similar ring around Sagittarius A*, about the size, as seen from Earth, of a doughnut on the Moon.

The puzzle is how such monsters grow. In 2000 two teams found that a black hole's mass tracks the speed of the stars in its galaxy's central bulge: black holes seem to weigh about a few tenths of a per cent as much as their bulges, as if galaxy and black hole grow together. Energy blasted out by a feeding black hole can heat or expel gas and stop star formation, a process called feedback. Yet the James Webb Space Telescope keeps finding black holes of millions of solar masses less than 500 million years after the Big Bang, which is uncomfortably early.

Did you know? The black hole in M87 is so large that light would take more than a day to cross its event horizon.

Our black hole is quiet now, sipping rather than gorging, though it flares in X-rays almost daily. The giant Fermi bubbles above and below the galactic plane suggest it was far busier a few million years ago. Even the calm heart of our galaxy has a temper.

Quasars

A quasar is a supermassive black hole caught in the act of feeding. Gas spiralling inwards forms a flat accretion disc, heated by friction to tens or hundreds of thousands of degrees, and it shines so brightly that a single quasar can outshine all the hundreds of billions of stars in its host galaxy by a factor of a hundred or more. Infalling matter can turn around a tenth of its mass into energy, more than ten times the efficiency of the nuclear fusion that powers the Sun.

Quasars were a mystery for years. In the late 1950s radio surveys at Cambridge found sources that looked, in visible light, like faint blue stars, hence the name, short for quasi-stellar radio source. In 1962 Cyril Hazard and colleagues used the Parkes radio dish in Australia to time the Moon passing in front of the source 3C 273, pinning down its position. In 1963 Maarten Schmidt took its spectrum with the 200-inch Hale telescope at Palomar and realised the strange lines were ordinary hydrogen lines, redshifted by 16 per cent. That put 3C 273 about 2.4 billion light-years away, so it had to be around four trillion times as luminous as the Sun.

What astonished astronomers was the size of the engine. Many quasars flicker over days or weeks, and nothing can vary faster than light can cross it, so the glowing region must be no bigger than the Solar System. The unified model explains much of the variety. Around the disc lies a doughnut of dust, the torus, and from the poles twin jets of particles can blast out at nearly the speed of light, sometimes for hundreds of thousands of light-years. Look straight down a jet and you see a dazzling, flickering blazar; look through the torus and the core is hidden, leaving only a radio galaxy or an obscured active nucleus. Carl Seyfert had catalogued milder versions in nearby spiral galaxies as early as 1943.

Quasars are fossils of a hungrier universe. Their numbers peaked about ten billion years ago, when galaxies were rich in gas and collided often; today most supermassive black holes, including our own, are quiet. The most distant known quasars shine from less than 700 million years after the Big Bang, already powered by black holes of a billion Suns. Surveys such as the Sloan Digital Sky Survey have now catalogued close to a million quasars.

Did you know? Because quasars are so distant that they barely move across the sky, they serve as the fixed points of the International Celestial Reference Frame, the grid against which star positions are measured.

Quasars are also cosmic flashlights. Their light passes through intervening gas clouds on its way to us, and each cloud leaves its fingerprint in the spectrum, mapping matter we could never otherwise see. The brightest objects in the universe help us chart its darkest corners.

Dark Matter's Grip

If you whirl a ball on a string, the faster it goes, the harder the string must pull. Stars orbiting a galaxy work the same way, with gravity as the string. Most of a spiral galaxy's light comes from its crowded centre, so astronomers expected stars in the outer disc to orbit slowly, as the outer planets do around the Sun: Neptune creeps along at about 5 kilometres per second, Mercury races at 48. Measure the speeds across a galaxy and plot them against distance, and you have its rotation curve. The prediction was a curve that rises and then falls away.

It does not fall. In the 1970s Vera Rubin and Kent Ford, at the Carnegie Institution in Washington, used a sensitive new image-tube spectrograph built by Ford to measure the Doppler shifts of glowing gas across the Andromeda Galaxy, and then across dozens of other spirals. The speeds stayed stubbornly high all the way to the edges. At the same time, radio astronomers including Morton Roberts and Albert Bosma traced the 21-centimetre glow of hydrogen far beyond the visible discs and found the same: the curves stay flat. Something unseen must be pulling, and there must be far more of it than of stars.

The idea was not new. In 1933 the Swiss astronomer Fritz Zwicky measured the speeds of galaxies in the Coma Cluster and found them moving so fast that the cluster should fly apart unless it held hundreds of times more mass than its light suggested. He called the missing material dunkle Materie, dark matter. For forty years few took it seriously. Rubin's flat curves changed that, and every independent test since has agreed: the bending of light around galaxy clusters, the pattern of ripples in the cosmic microwave background measured by the Planck satellite, and simulations that can only build galaxies in time if dark matter is there.

Put together, the evidence says that dark matter outweighs ordinary matter by about five to one. It does not emit, absorb or reflect light and barely interacts at all except through gravity. Each galaxy sits inside a huge, roughly spherical halo of it extending far beyond the stars; the Milky Way's reaches out hundreds of thousands of light-years. A minority of physicists argue that gravity itself needs modifying instead, an idea called MOND proposed by Mordehai Milgrom in 1983, but it struggles to explain clusters and the cosmic microwave background.

Did you know? Standard estimates of the dark matter density near the Sun mean a volume the size of the Earth would hold only around a kilogram of it.

What dark matter actually is remains unknown. In 2025 the observatory named after Vera Rubin opened its eyes in Chile, ready to map dark matter's grip across billions of galaxies. Most of the matter in the universe is something we have never touched.

Galaxy Clusters

Galaxies gather in groups, and the biggest gatherings are galaxy clusters: hundreds to thousands of galaxies, packed into a region a few million to ten million light-years across and bound by their own gravity. A rich cluster weighs up to about a thousand trillion Suns. Clusters are the largest objects in the universe that have pulled together and settled, rather than still drifting apart with cosmic expansion.

The nearest big one is the Virgo Cluster, about 54 million light-years away, with well over a thousand member galaxies. At its heart sits the giant elliptical M87, whose black hole the Event Horizon Telescope imaged in 2019. Farther off, about 320 million light-years away, lies the Coma Cluster, with more than a thousand bright galaxies, where Fritz Zwicky first found the signature of dark matter in 1933. In 1958 George Abell catalogued 2,712 rich clusters from the photographic plates of the Palomar Sky Survey, and Abell numbers are still in daily use.

The galaxies are the least of it. In the 1970s the Uhuru X-ray satellite discovered that clusters glow in X-rays, because the space between the galaxies is filled with thin gas at 10 to 100 million degrees. This intracluster medium holds several times more mass than all the stars combined. Most of the mass, roughly 85 per cent, is dark matter, which is why the galaxies move so fast without escaping. The breakdown is remarkably consistent: a few per cent stars, around an eighth hot gas, the rest invisible.

A cluster's gravity is strong enough to bend light. Galaxies behind a cluster appear stretched into arcs and multiple images, an effect called gravitational lensing that lets astronomers weigh the cluster and also magnify the faint galaxies behind it. The first JWST deep field, released in July 2022, looked through the cluster SMACS 0723 to galaxies more than 13 billion light-years away. The most famous case is the Bullet Cluster, two clusters that collided about 150 million years ago. In 2006 Douglas Clowe and colleagues showed with Chandra X-ray images and lensing maps that the hot gas had been slowed in the crash, while most of the mass, the dark matter, had passed straight through: a direct sign that dark matter is real stuff, not a quirk of gravity.

Did you know? In 2003 the Chandra X-ray Observatory found sound waves rippling through the gas of the Perseus Cluster: a B flat 57 octaves below middle C, the deepest note known in the universe.

Counting clusters of different masses at different epochs is one of the sharpest tests of dark energy, and surveys from eROSITA, Euclid and Rubin are now counting hundreds of thousands. Clusters are the cosmos's great cities, and like real cities their character lies mostly in what lies between the buildings.

The Local Group

The Milky Way does not travel alone. It belongs to the Local Group, a family of galaxies about 10 million light-years across, a name Edwin Hubble coined in 1936. Two giant spirals dominate it, the Milky Way and Andromeda, about 2.5 million light-years apart, and between them they hold most of the group's stars and its total mass of a few trillion Suns. Everything else orbits one of the two, or the gap between them.

The third-largest member is the Triangulum Galaxy, M33, a loose spiral about 2.7 million light-years away with perhaps 40 billion stars. Sharp-eyed observers can just glimpse it without a telescope from a truly dark site. It may be a satellite of Andromeda, or on its first fall towards it. Then come the Large and Small Magellanic Clouds, about 160,000 and 200,000 light-years away, bright enough to be seen with the naked eye from the Southern Hemisphere. They were known to peoples of the southern skies long before Ferdinand Magellan's fleet reported them during its voyage of 1519 to 1522, and al-Sufi mentioned the larger one around 964. The Large Cloud alone weighs perhaps a tenth as much as the Milky Way, enough to tug our galaxy's disc.

Most members are dwarf galaxies, faint swarms of a few thousand to a few billion stars. Andromeda keeps M32 and M110 close by; the Milky Way has dozens, including the Sagittarius Dwarf, being torn apart on the far side of the galactic centre, and the faint Fornax and Sculptor dwarfs. Many were found only recently. Since 2005 the Sloan Digital Sky Survey and the Dark Energy Survey have turned up dozens of ultra-faint dwarfs, and in 2018 Gaia revealed Antlia 2, a ghostly satellite as wide as the Large Magellanic Cloud but thousands of times fainter. The census now stands at more than a hundred known or candidate members, and the Vera C. Rubin Observatory is expected to add many more.

How do we know the group is bound? Nearly every other galaxy in the universe is rushing away from us as space expands, but within the Local Group the pull of gravity wins. Andromeda is approaching rather than receding, and measurements of galaxies near the group's edge mark a boundary, about 3 million light-years from its centre, inside which galaxies fall back rather than flowing away. The group itself sits on the outskirts of the Virgo Supercluster, part of a larger flow called Laniakea.

Did you know? Dark-matter simulations predicted far more satellite galaxies around the Milky Way than were known, a puzzle called the missing satellites problem; the steady discovery of ultra-faint dwarfs has since closed much of the gap.

Cosmic expansion will one day carry every other galaxy beyond our sight, but the Local Group will stay together. Far in the future, astronomers here may see only one merged galaxy and its dwarfs, and wonder whether anything else exists. The neighbourhood is home, for good.

The Cosmic Web

On the largest scales, galaxies are not sprinkled evenly through space. They are strung along vast filaments, gathered into sheets called walls and knotted together where filaments meet in dense clusters, the nodes. Between them lie voids, bubbles of near-emptiness up to hundreds of millions of light-years across. The whole pattern is called the cosmic web, and it is the largest structure in nature, a foam on a scale where even galaxy clusters are mere dots.

The web was mapped one slice at a time. In 1981 Robert Kirshner and colleagues found the Boötes Void, a hole more than 250 million light-years wide containing only a handful of galaxies. In 1986 Valérie de Lapparent, Margaret Geller and John Huchra at the Harvard-Smithsonian Center for Astrophysics plotted the redshifts of about 1,100 galaxies in a thin wedge of sky. The result looked like a slice through soap suds, with galaxies lining the walls of bubbles, and a stick-figure shape in the middle that became known as the "stickman". In 1989 the same team found the Great Wall, a sheet of galaxies about 500 million light-years long. Bigger surveys followed: the Anglo-Australian 2dF survey measured some 220,000 galaxies, and in 2003 J. Richard Gott and colleagues found the Sloan Great Wall, about 1.4 billion light-years long, in Sloan Digital Sky Survey data.

The web grows from almost nothing. The cosmic microwave background shows that the early universe was smooth to about one part in 100,000. Slightly denser regions pulled in more matter, mostly dark matter, which collapsed first into sheets and filaments and then into knots, while emptier regions emptied further. Ordinary gas followed, falling along the filaments into the clusters where galaxies grow. In 2005 the Millennium Simulation, run by Volker Springel and colleagues with ten billion particles, reproduced the web in startling detail from known physics, and the term itself had been coined by Richard Bond, Lev Kofman and Dmitri Pogosyan in 1996.

The filaments themselves are hard to see, since most of their gas is thin and hot. In 2019 astronomers using the MUSE instrument on the Very Large Telescope caught filament gas glowing faintly in the early universe, and X-ray surveys have detected warm gas bridging nearby clusters. In 2014 Brent Tully and colleagues traced the motions of galaxies to map Laniakea, the supercluster containing the Milky Way, about 520 million light-years across. Today ESA's Euclid, launched in 2023, and the DESI survey in Arizona are charting the web across billions of light-years in three dimensions.

Did you know? In 2020 an astrophysicist, Franco Vazza, and a neurosurgeon, Alberto Feletti, compared the cosmic web with the network of neurons in the brain and found their statistical structure strikingly similar.

The Milky Way sits in a modest filament near the edge of a void. Every galaxy you could ever see is a bead on these threads, spun by gravity from ripples present in the first moments. We live on a thread in a web of light.

The Big Bang

The Big Bang was not an explosion in space but the beginning of an expansion of space itself. Around 13.8 billion years ago everything in today's observable universe was packed into a state hotter and denser than anything since, and it has been expanding and cooling ever after. It happened everywhere at once. There is no centre to point to and no edge to stand on: every galaxy sees the others rushing away, as every raisin in a rising loaf sees its neighbours recede.

Run the film forwards and the physics is surprisingly well understood. One second in, the universe was about 10 billion degrees, a soup of protons, neutrons, electrons, neutrinos and light. Within the first few minutes protons and neutrons fused into nuclei, leaving matter about 75 per cent hydrogen and 25 per cent helium by mass, with a whiff of lithium. For the next 380,000 years the cosmos was an opaque, glowing fog; then, at about 3,000 degrees, electrons settled into atoms and light streamed free. Gravity took over, gathering gas into the first stars a hundred million years or so later, and then into galaxies, ours among them.

The idea grew out of Einstein's general relativity. Alexander Friedmann found expanding solutions to its equations in 1922, and Georges Lemaître, a Belgian priest and physicist, argued in 1927 and 1931 that the universe had grown from a "primeval atom". Edwin Hubble's 1929 measurements of receding galaxies gave it evidence. The name came from its sharpest critic: Fred Hoyle, champion of a rival steady-state universe, used the phrase "Big Bang" on BBC radio in 1949, and it stuck.

Three pillars of evidence hold it up. Galaxies recede at speeds in proportion to their distance. A faint microwave glow, predicted by Ralph Alpher and Robert Herman in 1948 and found by accident in 1965, fills the sky at 2.7 kelvin. And the amounts of hydrogen, helium and deuterium measured in the most pristine gas clouds match the recipe calculated for those first minutes. In 2018 the European Space Agency's Planck satellite fixed the age at 13.79 billion years, give or take about 20 million.

Did you know? Around 1 per cent of the snowy static on an old analogue television tuned between channels was the afterglow of the Big Bang.

What the theory does not describe is a cause, or a "before". It explains what happened from a tiny fraction of a second onwards; earlier than about 10⁻⁴³ seconds, the Planck time, known physics stops working and nobody yet knows what replaces it. It is also not a theory of a bang at a place, a myth kept alive by fireball illustrations. Most of the hydrogen in a glass of water was made in the universe's first second, so you have been carrying the Big Bang around all along.

The Cosmic Afterglow

The cosmic microwave background, or CMB, is the oldest light there is. It set out about 380,000 years after the Big Bang, when the universe had cooled to around 3,000 degrees, cool enough for electrons and protons to combine into hydrogen atoms. Until then light had pinballed between free electrons like headlights in fog; afterwards it flew straight. That moment, the surface of last scattering, surrounds us in every direction, and its light has been travelling towards us ever since.

Expansion has stretched that light by a factor of about 1,100, from a dull orange glow to microwaves. Today it is a near-perfect blackbody, the spectrum of a body at a single temperature, of 2.7255 kelvin, peaking at a wavelength of about a millimetre. Every cubic centimetre of space, including the one in front of your nose, holds about 400 of its photons.

Ralph Alpher and Robert Herman predicted a relic glow of a few kelvin in 1948, but the prediction was largely forgotten. In 1964 Arno Penzias and Robert Wilson at Bell Labs in Holmdel, New Jersey, were preparing a 6-metre horn antenna for radio astronomy and found a stubborn hiss of about 3.5 kelvin coming from every direction, day and night. They evicted a pair of nesting pigeons and scrubbed out what they politely called white dielectric material; the noise stayed. About 50 kilometres away at Princeton, Robert Dicke's group was building an instrument to look for exactly this signal. On hearing the news, Dicke is said to have told his colleagues, "Boys, we've been scooped." Penzias and Wilson shared the 1978 Nobel Prize.

Space missions turned the hiss into a portrait. NASA's COBE showed in 1990 that its spectrum is the most perfect blackbody ever measured, and in 1992 found ripples in temperature of about one part in 100,000, earning John Mather and George Smoot the 2006 Nobel Prize. NASA's WMAP (2001 to 2010) and ESA's Planck (2009 to 2013) mapped those ripples in fine detail. They are frozen sound waves from the early plasma: the most common spots span about one degree, twice the width of the full Moon, which shows that space is geometrically flat and fixes the recipe of about 5 per cent atoms, 27 per cent dark matter and 68 per cent dark energy.

Did you know? The sky is slightly warmer in one direction, by about three thousandths of a degree, because the Sun and Earth are moving at around 370 kilometres per second relative to the afterglow.

The ripples are the seeds of everything that followed: slightly denser patches grew, under gravity, into the galaxy clusters and filaments of the cosmic web. Ground telescopes such as the Atacama Cosmology Telescope and the South Pole Telescope now read finer detail and the light's faint polarisation. Every direction you look ends, eventually, at this wall of light, and beyond it the young universe is hidden in its own glowing fog.

The Expanding Universe

The expanding universe is the discovery that the distances between galaxies are growing, not because galaxies fly apart through space but because the space between them stretches. Picture galaxies as dots inked on a rubber sheet that is being pulled evenly at every edge. Every dot sees every other dot moving away, and a dot twice as far off retreats twice as fast. That simple proportion is Hubble's law: recession speed equals distance multiplied by the Hubble constant, around 70 kilometres per second for every megaparsec, a distance of about 3.26 million light-years.

The evidence came in pieces. From 1912 Vesto Slipher at Lowell Observatory in Arizona measured the spectra of spiral nebulae and found most of them shifted to the red, some receding at more than 1,000 kilometres per second. In 1927 Georges Lemaître combined such speeds with rough distances and showed that Einstein's equations predicted exactly this kind of expansion. In 1929 Edwin Hubble, working with Milton Humason at the 100-inch Hooker telescope on Mount Wilson, published the speed-distance relation that made it famous. In 2018 the International Astronomical Union recommended calling it the Hubble–Lemaître law. Hubble's own value was about 500, seven times too high, because his distances were badly underestimated.

The modern rate sounds slow. A distance of one megaparsec grows by about 70 kilometres every second, but that is only about one part in 14 billion each year. It does not stretch you, your house, the Earth or the Milky Way, because gravity and the forces between atoms hold bound objects together; expansion only shows in the gaps between groups and clusters of galaxies. Close to home, gravity wins outright: the Andromeda Galaxy is approaching us at about 110 kilometres per second.

At about 14 billion light-years, the Hubble distance, the recession speed passes the speed of light. Relativity is not broken, because nothing is moving through space faster than light; space itself is growing, and relativity puts no limit on that. Run the expansion backwards and every distance shrinks towards zero about 13.8 billion years ago, which is the Big Bang seen from the other end.

Did you know? The most distant matter we can see is now about 46 billion light-years away, even though the universe is only 13.8 billion years old, because space kept stretching while its light was on the way.

Measuring the exact rate is still a live fight. The Hubble Space Telescope and JWST, using Cepheid stars and supernovae, find about 73; the cosmic microwave background, read by Planck, implies about 67. Whoever proves right, the picture is the same. The galaxies are not fleeing from us; everyone is sitting still in a universe that keeps making more room.

Redshift

Redshift is the stretching of light to longer, redder wavelengths. Astronomers write it as z: the amount by which a wavelength has grown, divided by its original length. A galaxy at z = 0.1 has light stretched by 10 per cent; at z = 1 every wavelength has doubled; at z = 3 it has quadrupled. "Red" is shorthand. Ultraviolet light can be shifted into the visible, and visible light into the infrared, which is why the deepest views of the universe are infrared ones.

There are three causes. Motion through space produces a Doppler shift, the effect that drops the pitch of a passing siren. Gravity reddens light that climbs out of a gravity well, an effect Robert Pound and Glen Rebka measured in 1959 by firing gamma rays up a 22.5-metre tower at Harvard. For distant galaxies the dominant cause is cosmological redshift: the light's waves are stretched along with the space they travel through. So z is also a record of history. Light from a galaxy at z = 1 set out when every cosmic distance was half of today's, about 8 billion years ago.

The tool that makes this work is the spectral line. In 1814 Joseph von Fraunhofer catalogued hundreds of dark lines in sunlight, which turned out to be fingerprints of particular elements; hydrogen's red line always sits at 656.3 nanometres in the laboratory. Christian Doppler predicted in 1842 that motion would shift such frequencies, and in 1868 William Huggins used the shift of lines in the spectrum of Sirius to estimate its motion. In 1912 Vesto Slipher measured the first galaxy speed, and it was a blueshift: the Andromeda Galaxy approaches the Sun at about 300 kilometres per second, much of that our own orbit around the Milky Way.

Today redshifts come by the million. The Dark Energy Spectroscopic Instrument on Kitt Peak, Arizona, takes 5,000 spectra at a time with robotic optical fibres, and imaging surveys estimate rougher photometric redshifts from galaxies' colours. In 2024 JWST's NIRSpec spectrograph confirmed the galaxy JADES-GS-z14-0 at a redshift of about 14.2: its ultraviolet light, emitted at 0.12 micrometres, now arrives as infrared at about 1.8 micrometres. The cosmic microwave background itself sits at z of about 1,100.

Did you know? A galaxy at z = 14 is seen as it was less than 300 million years after the Big Bang, when the universe was about 2 per cent of its present age.

A common picture treats redshift as galaxies racing away like ambulances. That is right for nearby stars and galaxies, but beyond a few hundred million light-years the shift is mostly space itself stretching, and the light arrives with less energy than it set out with. Every galaxy's colour carries a date stamp, and a spectrograph is the instrument that reads it.

Dark Energy

Dark energy is the name for whatever is making the expansion of the universe speed up. The gravity of matter should slow expansion, as Earth's gravity slows a ball thrown upwards; in 1998 astronomers found the ball accelerating. In the cosmic inventory, dark energy makes up about 68 per cent of everything, dark matter about 27 per cent, and ordinary atoms, every star, planet and gas cloud ever seen, just under 5 per cent.

The discovery came from exploding stars. A Type Ia supernova, the blast of a white dwarf, reaches a nearly standard peak brightness, so its apparent faintness reveals its distance across billions of light-years. Two rival teams raced to compare those distances with redshifts: the Supernova Cosmology Project led by Saul Perlmutter at Berkeley, and the High-z Supernova Search Team led by Brian Schmidt, with Adam Riess. Both found the distant supernovae about 25 per cent fainter than a slowing universe allowed: farther away, so expansion must have sped up. The three leaders shared the 2011 Nobel Prize in Physics.

Independent evidence soon piled up. The microwave background, measured by WMAP and Planck, shows space is flat, which needs far more energy than matter supplies. Baryon acoustic oscillations, a slight excess of galaxy pairs about 500 million light-years apart left by sound waves in the young universe, act as a cosmic ruler; the Sloan Digital Sky Survey and the 2dF survey first detected it in 2005. Together they show that the acceleration began roughly 5 to 6 billion years ago, when matter had thinned out enough for the push to win.

What it is remains unknown. The simplest candidate is Einstein's cosmological constant, which he added to his equations in 1917 to hold the universe still and later dropped; the story that he called it his "biggest blunder" comes second-hand from George Gamow, and historians doubt it. A constant energy of empty space, equivalent to the mass of a few hydrogen atoms per cubic metre, does not dilute as space grows, so its share rises. Quantum theory predicts a vacuum energy up to 10¹²⁰ times larger, often called the worst prediction in physics. The alternative is a changing field, nicknamed quintessence. In 2024 and 2025 results from DESI hinted that dark energy may be weakening over time; the signal is intriguing but not yet at the level physicists accept as a discovery.

Did you know? All the dark energy inside a volume the size of the Earth adds up to the energy equivalent of about 6 milligrams of mass, less than a housefly weighs.

ESA's Euclid, the Rubin Observatory and NASA's planned Roman Space Telescope will measure the expansion history in finer detail than ever. Whatever they find will rewrite physics, and settle how the universe ends. For now, most of the universe is something we can measure precisely and cannot name.

Inflation

Inflation is the leading idea for what happened in the universe's first instant: a burst of runaway expansion that doubled the size of space again and again, at least 85 times, ending around 10⁻³² seconds after it began. A region smaller than a proton swelled by a factor of at least 10²⁶. On that scale a bacterium would grow to around 10,000 light-years across. This was far faster than light, which is allowed, because nothing moved through space; space itself grew.

It was invented to fix three puzzles. The horizon problem: opposite sides of the microwave sky have the same temperature to one part in 100,000, yet in an ordinary Big Bang they could never have been in contact to even out. The flatness problem: the geometry of space is balanced on a knife-edge between open and closed, which needs absurd fine-tuning at the start. And the missing magnetic monopoles that theories of particle physics predicted in abundance. Alan Guth, then a young physicist at SLAC in California, saw in 1979 that a brief exponential expansion would solve all three: a tiny patch evens out, then gets stretched so flat and so large that its edges and any monopoles lie far beyond view. Alexei Starobinsky had proposed a related model in Moscow, and in 1982 Andrei Linde, and independently Andreas Albrecht and Paul Steinhardt, fixed a flaw in how inflation ends.

The engine is usually pictured as a field, the inflaton, stuck in a high-energy state that behaves like an extreme version of dark energy. When it decays, its energy pours into particles and heats the universe: the hot Big Bang begins. Quantum jitters in the field, stretched to cosmic size, become slight differences in density, and gravity later turns those into galaxies.

The theory made predictions that have passed. Space should be flat, and Planck finds it flat to within about 0.2 per cent. The ripples in the microwave background should be random and nearly equal in strength on all scales but slightly stronger on large ones; WMAP and Planck measured exactly that tilt, a spectral index of about 0.965. The missing test is a background of primordial gravitational waves, which would leave a swirling B-mode pattern in the light's polarisation. In March 2014 the BICEP2 team at the South Pole announced it had been seen; within a year, joint work with Planck showed that dust in our own galaxy could explain the signal.

Did you know? If inflation is right, the largest structures in the universe, from galaxy clusters to the cosmic web, began as quantum fluctuations smaller than an atomic nucleus.

The Simons Observatory in Chile and Japan's planned LiteBIRD satellite will push the search further. Critics, now including Steinhardt, argue that inflation is flexible enough to fit almost any result and that it implies an untestable multiverse. Inflation remains the best account of the first instant, but the universe has not yet signed the confession.

Making the Elements

Every atom has a birthplace. The Big Bang made almost nothing but hydrogen and helium; nearly everything else in the periodic table, from the carbon in your cells to the gold in a wedding ring, was forged later in stars and stellar catastrophes, then scattered into space to be built into new stars, planets and people. This is nucleosynthesis, and read the right way the periodic table is a map of cosmic events.

The first forge ran for about 20 minutes. Between roughly ten seconds and twenty minutes after the Big Bang the whole universe was a fusion reactor cooling fast, and it left matter about 75 per cent hydrogen and 25 per cent helium by mass, with traces of deuterium and lithium. Ralph Alpher and George Gamow published the first calculations in 1948, adding Hans Bethe's name for the pun of "alpha, beta, gamma". The reactor stalled at lithium because no stable nucleus has a mass of 5 or 8, and the universe cooled before it could leap the gap.

Stars took over. They fuse hydrogen into helium, then helium into carbon through the triple-alpha process; in 1953 Fred Hoyle predicted that carbon must have a particular excited energy state for this to work, and it was found that year at Caltech, in the laboratory of William Fowler. Massive stars go on to build oxygen, neon, magnesium and silicon, up to iron. There the profit stops: iron-56 and its neighbours have the most tightly bound nuclei, so fusing beyond them costs energy instead of releasing it. Exploding white dwarfs supply much of the iron in the galaxy.

Heavier elements are built by adding neutrons. The slow s-process, in dying red giants, takes thousands of years and makes about half of them, including strontium, barium and lead. The rapid r-process needs a flood of neutrons, found where neutron stars collide and perhaps in rare supernovae, and makes gold, platinum and uranium. The plan was laid out in 1957 by Margaret and Geoffrey Burbidge, Fowler and Hoyle, in a paper still known as B²FH. Proof that stars work as forges came in 1952, when Paul Merrill found technetium in the spectra of red giants: it has no stable form and its longest-lived kind lasts only about four million years, so it must be made inside them now. In 2017 the light of the neutron-star merger GW170817 showed signs of freshly made heavy elements, and strontium was identified in its spectrum in 2019. Lithium, beryllium and boron are a special case, mostly chipped off heavier nuclei by cosmic rays.

Did you know? Around six of every ten atoms in your body are hydrogen, made in the first minutes of the universe and around 13.8 billion years old.

One puzzle remains: the oldest stars hold about three times less lithium than the Big Bang should have made. The rest of the story is settled enough to hang in a classroom. The periodic table on the wall is also a family tree of the cosmos.

Mapping the Universe

Cosmology became an exact science by counting. A redshift survey measures where each galaxy sits on the sky and how far its light is stretched, turning a flat picture into a three-dimensional map; imaging surveys add billions more galaxies with distances estimated from their colours. Each map tests the whole cosmic recipe at once: how much matter there is, how fast structure grows and how dark energy behaves.

The first slice was a shock. In 1986 Valérie de Lapparent, Margaret Geller and John Huchra at the Harvard–Smithsonian Center for Astrophysics plotted about 1,100 galaxies in a thin wedge of sky. Instead of a random scatter, galaxies lay on the walls of great bubbles around empty voids, and the Coma Cluster formed the body of what looked like a stick figure. In 1989 the same survey revealed the Great Wall, a sheet of galaxies some 500 million light-years long. The 2dF Galaxy Redshift Survey on the Anglo-Australian Telescope measured about 221,000 galaxies between 1997 and 2002, and the Sloan Digital Sky Survey, running since 2000 with a dedicated 2.5-metre telescope in New Mexico, has gathered spectra for millions.

Precision came with a ruler. In 2005 teams from Sloan, led by Daniel Eisenstein, and from 2dF found a slight excess of galaxy pairs about 500 million light-years apart, the frozen imprint of sound waves from the cosmic microwave background era. Planck's maps of that background, released in 2013 and refined in 2018, fixed the universe's main parameters to around a per cent, and galaxy maps now check them at later times. Since 2021 the Dark Energy Spectroscopic Instrument (DESI) on the 4-metre Mayall telescope at Kitt Peak has used 5,000 robotic fibres to build the largest 3D map yet, with more than 14 million galaxies and quasars in its 2025 release.

The next maps go deeper and wider. ESA's Euclid, launched in July 2023, is imaging about a third of the sky to measure the shapes of around 1.5 billion galaxies, whose slight distortions trace dark matter. The Vera C. Rubin Observatory on Cerro Pachón in Chile, with an 8.4-metre mirror and a 3.2-gigapixel camera, the largest digital camera ever built for astronomy, released its first images in June 2025 and is starting a ten-year Legacy Survey of Space and Time expected to catalogue around 20 billion galaxies.

Did you know? The Sloan Great Wall, found in Sloan data in 2003, is a chain of galaxies about 1.4 billion light-years long, more than 13,000 times the width of the Milky Way's disc.

No map shows the universe as it is now. The deeper a survey reaches, the further back in time it sees, so each atlas is also a history book read from the last page backwards. Cartography, once the art of coastlines, now charts most of the observable cosmos.

The Cosmic Calendar

The cosmic calendar squeezes the history of the universe into a single year. Carl Sagan popularised it in The Dragons of Eden in 1977 and in the television series Cosmos in 1980, when the universe was thought to be about 15 billion years old. With today's figure of 13.8 billion years, each day of the calendar stands for about 37.8 million years, each hour for about 1.6 million, and each second for about 437 years.

January is crowded. The Big Bang strikes at midnight on 1 January; at 00:14 the universe turns transparent and releases the light now seen as the cosmic microwave background. The first stars light up within the first week, and the Milky Way's oldest stars are shining before the month is out, with its disc taking shape through the spring. Then comes a long, quiet stretch in which generations of stars live, die and seed the gas with heavier elements. The Sun does not form until 1 September, and the Earth follows a day or so later.

The dates come from different clocks. The age of the universe is read from the microwave background and the expansion rate, measured by Planck. The Sun's age comes from the radioactive decay of uranium and lead in the oldest meteorite grains, which formed 4.567 billion years ago. The dates of life come from fossils, dated by the rocks around them. On that evidence the earliest microbes appear in late September, though the oldest traces are debated, and oxygen begins to fill the air around 29 October.

Most of what we think of as natural history crowds into December. Animals with shells and skeletons burst onto the scene in the Cambrian on 17 December. Plants reach dry land around 20 December, dinosaurs appear on Christmas Day, and an asteroid ends their reign early on the morning of 30 December. Our species, Homo sapiens, turns up at about 23:48 on New Year's Eve. Farming begins at 23:59:33, and Galileo raises his telescope to the sky less than a second before midnight.

Did you know? All of recorded history, from the first writing in Mesopotamia to this morning's news, fits into the last 12 seconds of 31 December.

The calendar is a scale model, not a prophecy. The year is not ending at midnight: the universe will go on for vastly longer than it has existed, and on the same scale the Sun has about four more months before it swells into a red giant. A human lifetime is about a fifth of a second. Every human story ever told has happened in the time it takes to say "Happy New Year".

How It All Ends

How the universe ends depends mostly on dark energy, the unknown something that is accelerating cosmic expansion. The gravity of all the matter can no longer stop that expansion on its own, so the outcome turns on whether dark energy stays the same, grows stronger or fades away. Whichever it is, the timescales dwarf everything so far. On the longest reckonings, the universe's 13.8 billion years are a few opening seconds.

The favoured fate is heat death, sometimes called the Big Freeze. If dark energy is a cosmological constant, as most data have long suggested, expansion goes on accelerating for ever. In roughly 100 billion years every galaxy beyond our own merged neighbourhood will have slipped beyond view, and any astronomers then would see a single giant galaxy in an empty dark, with little sign that a Big Bang ever happened. Star formation dwindles as gas runs out; Fred Adams and Gregory Laughlin estimated in 1997 that the last stars will fade around 100 trillion years from now. Black holes then dominate, slowly evaporating through the quantum glow predicted by Stephen Hawking in 1974: one with the mass of the Sun in about 10⁶⁷ years, the largest in around 10¹⁰⁰. What remains is a thin, cold radiation bath with nothing left to happen. The term "heat death" comes from 19th-century thermodynamics: not a universe that burns, but one where no heat can flow because everything is the same temperature.

The violent alternative is the Big Rip. In 2003 Robert Caldwell, Marc Kamionkowski and Nevin Weinberg showed that if dark energy grows denser over time, expansion would eventually overpower the forces holding everything together. In their worked example, about 22 billion years from now, galaxies are torn apart some 60 million years before the end, the Solar System comes unbound about three months before it, and atoms are ripped apart in the final instant. Current measurements push any rip, if there is one, far further off.

The reverse is the Big Crunch. If dark energy weakens and eventually turns negative, expansion could stall and reverse, pulling everything back into a hot, dense state. Hints from DESI in 2024 and 2025 that dark energy may be weakening revived interest, though no recollapse is established. Stranger still is vacuum decay: measurements of the Higgs boson and top quark suggest our vacuum may not be the lowest possible energy state, and a bubble of a truer vacuum could one day appear and spread at nearly the speed of light. Estimates put the expected wait far beyond 10¹⁰⁰ years.

Did you know? A small red dwarf such as Proxima Centauri may shine for around four trillion years, so the last stars to die will be the faintest ones.

Closer to home, the timetable is clearer: in about five billion years the Sun swells into a red giant. Euclid, DESI and Rubin are now measuring dark energy finely enough to narrow the cosmic forecast. The universe almost certainly has an ending, and we happen to live in its brief, bright opening chapter.

Worlds Beyond

For most of history the planets of other stars were a matter of faith. Giordano Bruno argued for countless worlds in the 1580s, and astronomers assumed them for centuries, but nobody could prove a single one. The problem is contrast. A planet is a dim speck lost in the glare of a star typically a billion times brighter, and even the nearest stars are light-years away. An exoplanet, any planet orbiting a star other than the Sun, had to be caught indirectly, by its effect on its star.

The first ones came as a shock. In 1992 Aleksander Wolszczan and Dale Frail, timing the pulsar PSR B1257+12 with the Arecibo radio telescope, found its metronome-perfect pulses arriving early and late in a pattern only orbiting planets could explain. These worlds circle the collapsed core of an exploded star some 2,300 light-years away. Then on 6 October 1995 Michel Mayor and Didier Queloz of the University of Geneva announced 51 Pegasi b, a giant of about half Jupiter's mass circling an ordinary Sun-like star every 4.2 days. Found with the ELODIE spectrograph at the Haute-Provence Observatory in France, it won them a share of the 2019 Nobel Prize in Physics.

What followed was an avalanche. In 1999 the first planet seen crossing the face of its star, HD 209458 b, proved these were real gas giants. NASA's Kepler space telescope, launched in 2009, stared at about 150,000 stars in Cygnus and Lyra and found more than 2,600 confirmed planets before retiring in 2018; TESS took over that year, scanning almost the whole sky. In 2016 a planet turned up around Proxima Centauri, the nearest star of all, and in 2017 seven Earth-sized worlds were revealed around TRAPPIST-1. NASA's Exoplanet Archive passed 6,000 confirmed planets in 2025, with thousands more candidates awaiting checks.

Did you know? The first exoplanets ever confirmed orbit a dead star: the pulsar PSR B1257+12, which spins about 160 times a second and sweeps its planets with beams of radiation.

The numbers now say something profound. Gravitational microlensing surveys, which catch planets by the way their gravity briefly magnifies a background star, suggest that the typical star in the Milky Way hosts at least one planet. Planets therefore outnumber stars, with hundreds of billions in our galaxy alone. They come in a bewildering variety, from lava worlds to puffed-up giants less dense than cork, and most systems look nothing like ours. A common myth casts exoplanets as rare jewels; in fact the rare thing, so far, is a system shaped like the Sun's.

Each method favours different worlds. Transits catch planets lined up edge-on; wobbles weigh the heavy and the close; direct imaging photographs young giants far from their stars; microlensing finds cold worlds and even rogues adrift between the stars. Together they turn every point of light overhead into a probable solar system. On a clear night, most of the stars you can see have planets, and one of those planets is the one you are standing on.

The Transit Method

Most known exoplanets were found by their shadows. When a planet's orbit happens to be lined up edge-on to us, it passes across the face of its star once every orbit, blocking a sliver of light. This is the transit method: watch thousands of stars at once, measure their brightness with exquisite precision, and look for dips that repeat like clockwork. The depth of the dip gives the planet's size, since the fraction of light blocked equals the area of the planet's disc divided by the star's. The spacing between dips gives the length of its year.

The dips are tiny. Jupiter crossing the Sun would dim it by about 1 per cent; Earth would dim it by only 0.008 per cent, roughly 84 parts in a million, for about 13 hours once a year. Seen from a random direction, there is only about a 1 in 200 chance that an Earth-like orbit lines up at all. That is why surveys watch enormous numbers of stars, and why the best of them work from space, above the shimmer of the atmosphere.

The first transit was caught in 1999, when David Charbonneau and colleagues, and independently Greg Henry, watched HD 209458 b, a gas giant about 160 light-years away, dim its star by around 1.5 per cent. NASA's Kepler, launched in 2009, turned the trick into an industry. It stared at one patch of sky in Cygnus and Lyra, measuring about 150,000 stars every half hour for four years, and found thousands of planets, including Kepler-186f, the first Earth-sized world confirmed in a habitable zone. After two of its reaction wheels failed it carried on as K2 until its fuel ran out in 2018. TESS, launched in April 2018, now surveys almost the whole sky in strips, favouring nearby, bright stars that are easier to follow up. ESA's PLATO mission is planned to join it from around the end of 2026.

Did you know? Transits can reveal planets that never cross their star at all: an unseen sibling's gravity makes a transiting planet arrive minutes early or late, which is how Kepler-19c was found in 2011.

Transits tell more than size. Combine the size with a mass from the wobble method and you get density, enough to tell iron and rock from water or gas. And during a transit, a thin ring of starlight filters through the planet's atmosphere, picking up the fingerprints of the molecules in it. Hubble detected sodium this way in HD 209458 b in 2002; in 2022 JWST found carbon dioxide in the atmosphere of WASP-39 b, the first clear detection of the gas on a planet beyond the solar system.

The method has a bias worth remembering: it favours big planets close to small stars, because they transit often and deeply. The census is lopsided, and astronomers correct for it statistically. The alignment also cuts both ways. From some nearby stars, an observer watching the Sun would see Earth's own transit each year, a regular wink announcing that a small world circles in our star's habitable zone.

The Wobble Method

A planet does not simply orbit its star. Both swing around their shared centre of mass, the barycentre, like a dancer whirling a much lighter partner. The star traces a tiny loop, moving towards us for half the orbit and away for the other half. That motion shifts its light by the Doppler effect: slightly bluer as it approaches, slightly redder as it recedes. By measuring this radial velocity, astronomers can detect and weigh planets they cannot see.

The signals are absurdly small. Jupiter drags the Sun around at about 12.5 metres per second, roughly a sprinter's speed, over its 12-year orbit. Earth moves it at just 9 centimetres per second, about the pace of a giant tortoise. Jupiter's pull shifts the wavelength of sunlight by a few parts in a hundred million, Earth's by a few parts in ten billion, so the spectrographs that measure such shifts must be extraordinarily stable: sealed in vacuum tanks, held at constant temperature and calibrated against glowing lamps, iodine gas or laser frequency combs.

This method announced the exoplanet age. Bruce Campbell, Gordon Walker and Stephenson Yang reported a possible planet around Gamma Cephei in 1988 but were too cautious to claim it; it was confirmed only in 2003. Then in 1995 Michel Mayor and Didier Queloz used the ELODIE spectrograph in France to find 51 Pegasi wobbling at about 56 metres per second every 4.2 days, the signature of a giant hugging its star. Geoff Marcy and Paul Butler in California quickly confirmed it and went on to find dozens more. The HARPS spectrograph at La Silla in Chile, working from 2003, reached about 1 metre per second and in 2016 helped find Proxima b around the nearest star. ESPRESSO on the Very Large Telescope now pushes towards 10 centimetres per second.

Did you know? The barycentre of the Sun and Jupiter lies just outside the Sun's surface, so strictly speaking Jupiter does not orbit the Sun's centre at all.

The wobble's rhythm gives the planet's period, its size gives the mass, and its shape reveals whether the orbit is circular or stretched. There is one catch. Because we usually do not know how the orbit is tilted, the method yields only a minimum mass. Pair it with a transit, which fixes the tilt, and the true mass emerges. Stars also cheat: starspots, pulsations and the boiling of their surfaces produce jitters of a metre per second or more that can mimic planets, and untangling them is now the field's hardest problem.

The prize is a true Earth twin, which demands catching 9 centimetres per second over years. The planned Terra Hunting Experiment aims to watch a few dozen Sun-like stars almost nightly for a decade to try. Meanwhile the wobble supplies the masses of most nearby worlds, including the small planets teased out around Barnard's Star in 2024 and 2025. Seen from afar, our own Sun is quietly flinching at nine centimetres a second, and in principle someone could be measuring it.

Hot Jupiters

Before 1995, theory told a neat story. Giant planets form far from their stars, beyond the snow line, where ice is plentiful enough to build massive cores, so gas giants should sit in wide, slow orbits like Jupiter's 12-year circuit at 5.2 times Earth's distance from the Sun. Then the first planet found around a Sun-like star broke the rule. 51 Pegasi b has about half Jupiter's mass but orbits just 0.05 astronomical units from its star, eight times closer than Mercury is to the Sun, and its year lasts 4.2 days. It is the archetype of the hot Jupiters.

These are some of the most extreme places known. Tides at such close range almost certainly lock them with one face permanently towards the star, giving a scorching day side and a cooler night side, with jet streams of thousands of kilometres an hour between them. KELT-9b, the hottest known, has a day side of around 4,300 °C, hotter than many stars. On WASP-76b, the ESPRESSO spectrograph on ESO's Very Large Telescope found in 2020 that iron vaporises on the day side and appears to condense as the gas crosses into night. HD 189733 b, deep blue in Hubble's eyes, may have winds carrying droplets of molten glass. HD 209458 b trails a comet-like tail of hydrogen boiling off its heated atmosphere.

How did they get there? Most astronomers think they formed farther out and migrated inward. Some spiralled in through the gas disc around the young star, trading momentum with it, a scenario worked out for 51 Pegasi b by Douglas Lin, Peter Bodenheimer and Derek Richardson in 1996. Others may have been flung onto stretched orbits by sibling planets, then had their orbits rounded off by tides. Measurements made during transits of the angle between a hot Jupiter's orbit and its star's spin show that many orbit tilted or even backwards, scars of violent histories. A minority may simply have formed where they are.

Hot Jupiters dominated early discoveries because they are easy. Big, close and quick, with wobbles of tens to hundreds of metres per second, they shout in both the wobble and the transit methods. Yet surveys show they are genuinely rare, found around only about 1 per cent of Sun-like stars; their early fame was a selection effect. Some are doomed. WASP-12b, circling its star in about 26 hours, has an orbit shrinking by roughly 29 milliseconds a year and should be swallowed within a few million years.

Did you know? The hot Jupiter TrES-2b reflects less than 1 per cent of the light that falls on it, making it darker than coal.

Their real legacy is a lesson in humility. Hot Jupiters proved that planetary systems are dynamic, with giants that can wander across a system and scatter whatever lies in their path. Our own Jupiter may once have drifted inward and then back out, a proposal called the Grand Tack, and in doing so shaped the small rocky worlds of the inner solar system. Had it kept going, Earth might never have formed.

Super-Earths & Mini-Neptunes

If you picked a planet at random from the Milky Way, it would probably be a kind our solar system lacks. Kepler's census showed that the most common planets around Sun-like stars, at least in orbits of up to a few months, are between one and four times Earth's width. Earth is the largest rocky planet we have; Neptune, at almost four Earth radii, is the smallest giant. In between lies an empty gap in our own family and a crowded one everywhere else.

These planets split into two populations. In 2017 Benjamin Fulton and colleagues, using precise star sizes from the California-Kepler Survey, found a shortage of planets around 1.5 to 2 Earth radii, now called the radius valley or Fulton gap. Below it sit super-Earths, mostly bare rock and iron, some several times heavier than Earth. Above it sit mini-Neptunes, rocky cores wrapped in thick envelopes of hydrogen and helium, puffed up by a few per cent of their mass in gas. The leading explanation is that young planets close to their stars lost their light atmospheres, stripped by stellar X-rays and ultraviolet light or by heat leaking from their own cores, while those with more gas or more distance kept them. Leslie Rogers had estimated in 2015 that most planets bigger than about 1.6 Earth radii are not rocky.

Each is an experiment we cannot run at home. 55 Cancri e, almost twice Earth's width and eight times its mass, circles its star in under 18 hours with a day side hot enough to melt rock; JWST data in 2024 hinted that it holds an atmosphere, perhaps fed by a magma ocean. GJ 1214 b, a mini-Neptune 48 light-years away, is cloaked in haze so thick that its spectrum is nearly featureless. Then there is K2-18 b, 124 light-years away and 2.6 times Earth's width, orbiting in its red dwarf's habitable zone. Nikku Madhusudhan of Cambridge has proposed it as a hycean world, a planet-wide ocean under a hydrogen sky, though other teams argue it may be a gas-rich mini-Neptune or even have a molten surface.

Did you know? Statistically, the Milky Way's most typical planet is a kind we cannot visit or study up close, because the Sun never made one.

Size and mass alone are ambiguous. A planet of 2.5 Earth radii and 8 Earth masses could be rock under a thin hydrogen veil, or rock beneath a deep layer of water, or something stranger; several mixtures fit the same density. That is why JWST's spectra of their atmospheres matter. The balance of water, methane, ammonia and carbon dioxide helps tell an ocean world from a gas dwarf.

Why our system skipped them is itself a clue. One idea is that the young Jupiter blocked the inward drift of pebbles that might have built super-Earths close to the Sun, leaving the small rocky planets we have. Whatever the cause, the planet under your feet is the exception, and the galaxy's favourite kind of world is one we can only study from light-years away.

The Habitable Zone

Around every star lies a band of orbits where a planet with the right kind of atmosphere could keep liquid water on its surface: the habitable zone, nicknamed the Goldilocks zone. Too close and oceans boil away; too far and they freeze solid. The term was popularised by Su-Shu Huang in 1959 and made quantitative by James Kasting and colleagues in 1993, who modelled how a planet's carbon dioxide and water vapour respond to starlight. In 2013 Ravi Kopparapu and colleagues updated the calculation, placing the Sun's conservative zone at roughly 0.99 to 1.7 times Earth's distance, with Earth uncomfortably close to its inner edge.

The zone's position depends on the star. A hot, bright F-type star pushes it out beyond the distance of Mars; a dim red dwarf pulls it in to a few hundredths of an astronomical unit, where a year lasts days or weeks. Proxima b, 4.2 light-years away, orbits in its star's habitable zone every 11.2 days at just 0.05 astronomical units. TRAPPIST-1's zone lies well inside the size of Mercury's orbit. Kepler-186f, announced in 2014, was the first Earth-sized planet confirmed in a habitable zone; Kepler-452b, announced in 2015, circles a Sun-like star in a 385-day year.

The zone also moves with time. Stars brighten as they age, and the Sun is about 30 per cent more luminous than when Earth formed. In roughly a billion years its growing glare will push the inner edge beyond Earth's orbit, and our oceans will begin to escape to space. Life on Earth occupies a window in the zone's slow outward march.

Did you know? Venus sits just inside the inner edge of the Sun's habitable zone, and its surface, at about 465 °C, is hot enough to melt lead.

The habitable zone is a first filter, not a promise, and a popular myth treats it as a label for life. Our own system proves the point. Mars orbits within the zone yet is a frozen desert, because it was too small to keep its atmosphere once its magnetic field faded. Atmosphere, magnetism, plate tectonics, tides and stellar flares all decide the rest. Red dwarfs, about three-quarters of the galaxy's stars, are convenient targets with brutal habits: violent flares, and planets likely to be tidally locked, with one face in permanent day. Conversely, water can exist far outside any habitable zone. Europa and Enceladus hide oceans beneath their ice, kept liquid by tidal flexing.

Even with these caveats the numbers are large. Analyses of Kepler data suggest that a substantial fraction of Sun-like stars host a roughly Earth-sized planet in the zone, and red dwarfs may add tens of billions more across the galaxy. The habitable zone tells telescopes where to look first. For us it is a reminder that a blue sky is a fortunate, and temporary, address.

TRAPPIST-1

Forty light-years away in Aquarius lies a star so faint that you need a large telescope to see it at all. TRAPPIST-1 is an ultracool red dwarf only slightly wider than Jupiter, with about 9 per cent of the Sun's mass and a surface around 2,300 °C, glowing mostly in infrared. Around it circle seven rocky planets, each roughly the size of Earth. The whole system would fit comfortably inside the orbit of Mercury: the outermost planet, h, orbits at about 0.06 astronomical units, and its year lasts under 19 days.

The planets were found by their transits. In 2016 Michaël Gillon of the University of Liège and colleagues announced three, using TRAPPIST, a 60-centimetre robotic telescope at La Silla in Chile whose name nods to the Trappist beers of Belgium. In February 2017 NASA's Spitzer Space Telescope, after watching the star almost continuously for about 20 days, showed there were seven. Their orbits form a precise resonance chain: for every eight orbits of b, c completes five, and simple ratios continue down the line. So delicate a chain suggests the planets formed farther out and migrated inward together, gently.

Because the planets transit often and the star is small, the dips are deep and the timing exquisite. Tiny variations in when each planet crosses, measured over years, give their masses. They turn out to have similar densities, a little lower than Earth's, consistent with rocky bodies holding slightly less iron or some water. Three planets, e, f and g, lie in the habitable zone, with d at its warm inner edge.

The James Webb Space Telescope has started asking the obvious question: do they have air? In 2023 its MIRI instrument measured the day side of b at about 230 °C, close to what a bare dark rock would reach, and c at about 110 °C, ruling out a thick, Venus-like carbon dioxide atmosphere on either. The star's habits count against them: TRAPPIST-1 flares frequently and was brighter in its youth, which may have stripped the inner worlds. The verdict on e, the best candidate for an Earth-like climate, is still being gathered transit by transit.

Did you know? From the surface of one TRAPPIST-1 planet, a neighbour passing close by could look larger in the sky than the full Moon does from Earth.

TRAPPIST-1 is a laboratory for the most important statistical question in the search for life. Red dwarfs make up most of the galaxy's stars, and the smallest live for trillions of years, far longer than the Sun's 10-billion-year lifetime. If planets like these can keep their atmospheres, most of the habitable real estate in the universe orbits dim red suns, and it will last almost forever.

Biosignatures

We will not visit the surfaces of other stars' planets any time soon, so the search for life there is a matter of chemistry at a distance. A biosignature is a substance or pattern that life produces and that is hard to make otherwise. On Earth the clearest is the atmosphere itself: about 21 per cent oxygen, plus methane at under two parts per million. The two react with each other in sunlight, and methane lasts only about a decade in our air, so their coexistence means something is constantly replenishing both. James Lovelock argued in 1965 that this kind of chemical disequilibrium would betray life on any planet.

The idea has been tested on the one living world we know. In December 1990 the Galileo spacecraft, swinging past Earth on its way to Jupiter, turned its instruments on its home planet, and in 1993 Carl Sagan and colleagues reported what an alien observer would conclude. They found abundant oxygen, methane far out of equilibrium, a sharp rise in reflected light just beyond the red, now called the red edge of vegetation, and narrowband radio signals that no natural process explains. Life, and even technology, was detectable.

For exoplanets the main tool is transmission spectroscopy. During a transit, starlight passing through a planet's atmosphere loses specific colours absorbed by its molecules, and JWST, launched in December 2021, reads them with instruments such as NIRSpec and MIRI. In 2023 a team led by Nikku Madhusudhan reported methane and carbon dioxide in the atmosphere of K2-18 b, a possible ocean world 124 light-years away, plus a tentative hint of dimethyl sulphide, a gas that on Earth comes mainly from marine plankton. A 2025 follow-up claimed a stronger signal, but independent reanalyses found the evidence weak or absent, and the argument continues.

Did you know? Seen from another star, Earth's air may have advertised life for around two billion years, ever since oxygen from cyanobacteria built up, long before any animals existed.

No single molecule proves life, and false alarms are part of the method. Oxygen can accumulate without life when ultraviolet light splits water on a planet losing its hydrogen to space, a particular risk around red dwarfs. Methane leaks from volcanoes and from water reacting with hot rock. In 2020 a reported detection of phosphine on Venus caused a sensation, then a long dispute over the data. Astronomers now talk of a ladder of confidence, from "consistent with life" to "no other explanation", and expect any first claim to be fought over for years.

Better tools are coming. ESO's Extremely Large Telescope, planned for first light around the end of this decade, may probe oxygen on nearby rocky planets, and NASA's proposed Habitable Worlds Observatory aims to image Earth-like planets around Sun-like stars and take their spectra directly. The question is no longer whether we could recognise a living world, but how sure we would need to be. Somewhere in the starlight already reaching us, the answer may be written in a few missing colours.

The Drake Equation

In the autumn of 1961 a young radio astronomer, Frank Drake, hosted a small meeting at the National Radio Astronomy Observatory in Green Bank, West Virginia, to discuss whether we could detect intelligent life elsewhere. To organise the agenda he wrote down the factors that would decide how many civilisations in our galaxy might be sending signals right now. The result, now called the Drake equation, multiplies seven terms: the rate at which suitable stars form, the fraction with planets, the number of habitable planets per system, the fractions of those on which life, intelligence and detectable technology arise, and L, the length of time such civilisations keep transmitting.

About a dozen people attended, including Carl Sagan, Philip Morrison and the chemist Melvin Calvin, who learned during the meeting that he had won the Nobel Prize in Chemistry. They called themselves the Order of the Dolphin, after the attendee John Lilly's work on dolphin intelligence. Their own guesses gave an answer roughly equal to L: if civilisations last a thousand years, perhaps a thousand are out there; if they last a million, perhaps a million.

Six decades on, astronomy has filled in the front of the equation. The Milky Way turns roughly one to two Suns' worth of gas into new stars each year. Kepler and its successors show that most stars have planets and that a meaningful fraction have rocky worlds in their habitable zones. The astronomical terms, pure guesses in 1961, are now known to within factors of a few.

Did you know? Drake never expected his equation to be solved: he wrote it simply as a way to organise the agenda of a meeting.

The back half remains guesswork. We have one example of life, one of intelligence and one of technology, and a sample of one cannot give a fraction. Life appeared on Earth fairly quickly, perhaps within a few hundred million years, which some take as a hint that it is easy; intelligence capable of building radios took about four billion years, and only one species out of millions managed it. L is the most humbling term of all, since our own civilisation has been radio-loud for barely a century. Honest answers still range from far less than one, meaning we are probably alone in the galaxy, to many thousands or more.

The equation is often mocked as a way of multiplying ignorance, but that misses its purpose. It turns a vague question into a research programme, showing which numbers astronomy, biology and history would each have to supply. It also sharpens the Fermi paradox, Enrico Fermi's lunchtime question of 1950: if civilisations are common, where is everybody? Frank Drake died in 2022. His equation, still unsolved, remains the best map we have of our own uncertainty, with our own future written into its final term.

SETI

The Search for Extraterrestrial Intelligence, or SETI, looks for technology rather than biology: radio transmissions, laser flashes or other signs that someone, somewhere, built something. Its scientific starting gun was a two-page paper in Nature in 1959, in which Giuseppe Cocconi and Philip Morrison pointed out that radio telescopes could already detect a transmitter like our own across interstellar distances. They suggested listening near 1,420 megahertz, the natural frequency of hydrogen, the universe's commonest element.

A year later Frank Drake tried it. Project Ozma, in 1960, used an 85-foot dish at Green Bank to listen to two nearby Sun-like stars, Tau Ceti and Epsilon Eridani, for about 150 hours. It heard nothing, and set the pattern. On 15 August 1977 Ohio State University's Big Ear telescope recorded a strong, narrow signal from the direction of Sagittarius, lasting the 72 seconds a source took to drift through its beam; the volunteer Jerry Ehman circled it on the printout and wrote "Wow!" beside it. Despite many attempts it has never been seen again. Proposed natural explanations, including a flaring hydrogen cloud suggested in 2024, remain unconfirmed.

The searches since have grown enormously. The SETI Institute, founded in 1984, built the Allen Telescope Array in northern California, which began observing in 2007. Breakthrough Listen, launched in 2015, uses the Green Bank Telescope, the Parkes dish in Australia known as Murriyang, and the MeerKAT array in South Africa to survey a million nearby stars and the centres of a hundred nearby galaxies. China's 500-metre FAST dish has joined in. Modern receivers sift billions of radio channels at once, and machine-learning tools, such as one trained in 2023 by Peter Ma and colleagues, now flag candidates among millions of human-made interference signals.

Did you know? A 2018 estimate by Jason Wright and colleagues found that all SETI searches to date had examined about as much of the possible search space as a large hot tub holds of water, compared with all of Earth's oceans.

That is the key to reading the silence. The cosmic haystack has many dimensions: direction on the sky, radio frequency, signal strength, type of signal and time, since a beacon might be switched on only occasionally. Silence from so small a sample says very little, a point made for decades by Jill Tarter, the SETI pioneer who inspired the heroine of Contact. There have been false alarms too: BLC1, a signal picked up in 2019 from the direction of Proxima Centauri, turned out to be interference from electronics on Earth.

A popular myth says aliens could be watching our old television. In fact our broadcast leakage is weak and would be hard to detect even a few light-years away, though powerful planetary radars carry much farther. SETI is a modest bet with the largest imaginable payoff, and every night the telescopes listen a little longer. Most of the sky's whispers are still unheard.

How Telescopes Work

A telescope is a light bucket with a sharp edge. Its first job is to collect far more light than the pupil of your eye, which opens to about 7 millimetres on a dark night; its second is to bring that light to a crisp focus, where an eye, a camera or a spectrograph can use it. There are only two ways to do this. A refractor bends light through a lens. A reflector bounces it off a curved mirror.

What matters most is the aperture, the diameter of the lens or mirror. Light grasp grows with the area, so with the diameter squared: a 1-metre mirror gathers about 20,000 times as much light as a dark-adapted eye, and an 8-metre giant more than a million times. Aperture also sets sharpness. Light spreads as it passes any opening, so the finest detail a telescope can split is roughly the wavelength divided by the diameter. For visible light a 1-metre telescope could in principle resolve about a seventh of an arcsecond, the width of a coin seen from 30 kilometres.

Refractors came first, in the Netherlands in 1608, and Galileo turned one on the sky a year later. Their weakness is colour. A simple lens bends blue light more than red, smearing every star into a fringed blur, a fault called chromatic aberration. Two-element achromatic lenses, patented by John Dollond in 1758, tamed it, and refractors grew to the 40-inch (1.02-metre) lens at Yerkes Observatory in Wisconsin, completed in 1897. That was effectively the end of the road. A lens can only be held by its rim, so a big one sags under its own weight, and light must pass through it, so the glass has to be flawless all the way through.

Did you know? Isaac Newton's first reflector of 1668 was only about 15 centimetres long, with a mirror of polished speculum metal some 3 centimetres across, yet it rivalled refractors many times its length.

Mirrors have none of those problems. A mirror reflects every colour to the same point, only its front surface needs to be perfect, and it can be propped up from behind. Newton put a small flat mirror inside the tube to send light out to the side; Laurent Cassegrain's design of 1672 sent it back through a hole in the main mirror instead, the layout used by Hubble and most professional telescopes today. When Léon Foucault and Carl August von Steinheil learned to coat glass with silver in the 1850s, the reflector won the size war for good.

One popular myth deserves to go: magnification hardly matters. Any telescope can magnify enormously by swapping the eyepiece, but it only shows more if it has gathered more light and focused it sharply. Astronomers rate an instrument by aperture, image quality and the detector behind it, never by zoom. When you look through a backyard telescope, you are using the same physics as the giants on the mountaintops: a wider bucket, and a truer curve.