The Formation Of The Earth: A Prehistoric Sleep Documentary

Full transcript

In the Jack Hills of Western Australia, locked inside a metamorphosed sandstone conglomerate, sit grains of zircon smaller than a grain of table salt. Those crystals are almost the entire physical record of the formation of the Earth. The oldest has been dated to four point four oh four billion years ago, give or take eight million. Nearly everything else from that time has been remelted, subducted, or shattered. So the question has to be asked backwards. From grains like these, how did one collapsing fragment of a cold molecular cloud become a rocky world with a single large moon, a global ocean, and an atmosphere with oxygen in it?

If these journeys through deep time help you rest, subscribe so the next ancient world can find you at bedtime. The evidence runs in a sequence, and the sequence is the story. A cold cloud fragmenting, a disc flattening, dust gathering into a planet, a Mars-sized body arriving, a magma ocean freezing over, a first ocean condensing out of the sky, the first traces of life in the rock record, and oxygen finally entering the air. The Jack Hills are low, dry ridges in the Narryer Gneiss Terrane, and the rock that holds the crystals is not itself as old as the crystals are.

It is a conglomerate, a sandstone made from older sand, later cooked and pressed until its grains recrystallized. The zircons inside it were already ancient when that sand was laid down. They had weathered out of some earlier rock that no longer exists, been carried by moving water, been buried, and been sealed inside the new stone like seeds in a loaf. The rock they came from is gone. The grains are what remain. Zircon endures what almost nothing else endures. It resists heat, resists chemical attack, and survives being tumbled down rivers for millions of years.

It also builds a clock into itself. When a zircon crystal grows, it accepts uranium into its structure and largely excludes lead. So lead found inside such a grain long afterwards has arrived by radioactive decay. The proportion gives an age. This is why a mineral no larger than a speck of sand can be spoken about with a decimal point, and why the Hadean, the first and oldest eon of Earth's history, is known mostly through a scattering of granular crystals from a single locality. The eon takes its name from Hades, the underworld of Greek myth, and the American geologist Preston Cloud coined the term to label the stretch of time before the earliest known rocks.

The name refers to conditions, not to legend. The planet had only just finished accreting, and its surface is thought to have been molten lava. The British geologist W.B. Harland later proposed an almost synonymous term, the Priscoan, from a Latin word for ancient, and older texts sometimes call the same interval the Pre-Archean. By international convention, the Hadean begins with the formation of the planet about four point six billion years ago and ends four point zero three one billion years ago at the age of the oldest known intact rock formations.

That word, formation, does a great deal of work in geology and needs watching. A geological formation is a formally named body of rock, a stratum with a name and mapped boundaries. The formation of the Earth is an event, or rather, a long process of events. Both meanings appear in the same story. In southern Guyana, in the Iwokrama Formation of the Guyana Shield, zircon cores have been dated at four point two two billion years ago, sitting inside much younger rock that gathered them up like inherited furniture.

The formation is the rock unit. The four point two two billion year date belongs to the grains. Through the last decades of the twentieth century, geologists identified a handful of Hadean materials in three parts of the world: western Greenland, northwestern Canada, and Western Australia. Before those fragments turned up, the early Earth belonged almost entirely to modelers, to people working with equations about heat and viscosity rather than with objects that could be held under a microscope, the arrival of real Hadean grains changed the argument from what must have happened to what the samples allow.

What the samples allow is quietly startling. The four point four zero four billion-year crystal is a slight outlier, and the oldest consistently dated zircons fall closer to four point three five billion years, which is roughly two hundred million years after Earth is thought to have formed. Studies of these grains suggest that liquid water may have existed between about four point four and four point zero billion years ago. That is very soon after the planet assembled and much earlier than geologists once assumed for a world whose surface had recently been an ocean of melt.

A study of Australian Hadean rock published in two thousand and eight found minerals that some researchers read as evidence for plate tectonics as early as four billion years ago, roughly six hundred million years after Earth formed. Those readings remain contested, and the reasons are worth holding lightly through everything that follows. The direct evidence is limited because the zircons come largely from one place. Geophysical models of the Hadean are underconstrained, capable of painting a general picture but not of settling details, and they remain controversial among geologists.

Some argue that the minerals taken as signatures of early plate tectonics could instead have been produced by meteorite impacts, which were plentiful at the time. Even the date of Earth's own formation comes with a spread. It is variously given as about four point five billion years ago, as four point five four billion years ago, with an uncertainty of one percent, as four point five four plus or minus zero point zero four billion years, and as four point six billion years for the opening of the Hadean.

The spread is real, and it is honest to leave it that way. The atoms in those crystals and in the rock around them and in the ocean and the air have a history older than the planet. A few hundred thousand years after the Big Bang, thirteen point eight billion years ago, the expanding universe cooled to the point where atoms could form at all. Hydrogen and helium date from that moment, along with a trace of lithium. As expansion continued and things cooled further, atoms lost enough kinetic energy and dark matter coalesced sufficiently for protogalaxies to gather, and further accretion built those into galaxies.

Galaxies then grew by merging with one another and by drawing in smooth gas, and inside them, the same patient accumulation formed stars. Accretion is the word for that whole habit of the universe, matter attracting more matter, usually gas, and settling into a disk as it falls. Galaxies, stars, and planets are all made this way. Everything heavier than the light elements, though, had to be manufactured afterwards inside stars. Fred Hoyle argued that evolved stars called red giants create many of the heavier elements in their cores and that when a red giant casts off its outer layers, those elements are recycled into new star systems.

Supernovae eject others outright. The route from a hydrogen universe to a rocky planet runs through the interiors of stars that died before the Sun existed. Arthur Stanley Eddington reached the first half of that idea from an unexpected direction. His confirmation of Albert Einstein's theory of relativity led him to the realization that the sun's energy comes from nuclear fusion in its core, hydrogen becoming helium. In 1935, he went further and suggested that other elements might also form within stars.

The suggestion turned out to describe the supply chain for every mountain range and every ocean basin that would ever exist. There is a more specific fingerprint, and it sits inside meteorites. Studies of ancient meteorites reveal traces of stable daughter nuclei of short-lived isotopes, among them iron-60, which forms only in exploding short-lived stars. Its presence indicates that one or more supernovae went off nearby before Earth's material gathered. The distribution of iron- 60 across the solar system is highly homogeneous, and that even spread points to the supernova and to the injection of its debris happening well before nebular dust accreted into planetary bodies.

The signature was stirred in thoroughly, and then the stirring stopped. Stars form inside giant clouds of cold molecular hydrogen. A typical one runs to roughly three hundred thousand times the mass of the Sun and about sixty-five light-years across, which is twenty parsecs, and the cloud that produced the solar system was of that scale. Cold is the operative word. This is hydrogen in molecular form, dark and slow, the kind of material that shows up in photographs as an absence rather than a glow.

Such clouds are the ordinary nurseries of the galaxy, and the Orion Nebula is the standard example of the type. Over millions of years, these clouds are prone to collapse and fragmentation. The fragments come out at roughly one parsec across, about three point two six light-years, and those fragments collapse further into small, dense cores measuring between zero point zero one and zero point one parsec, which is two thousand to twenty thousand astronomical units. Dense in this setting is a relative term.

The particle number density of such a core runs to something like ten thousand to a hundred thousand particles per cubic centimeter. Air at sea level holds two point eight times ten to the nineteenth particles in the same volume. The densest part of the nursery is emptier than the best vacuum ever produced in a laboratory, and it is still enough to build a star. One of those collapsing fragments is called the presolar nebula, and it made the solar system. Its mass was just over that of the Sun.

Its composition was close to the Sun's composition today. Hydrogen, together with helium and trace lithium from Big Bang nucleosynthesis, made up about ninety- eight percent of its mass. The remaining two percent was the heavier material created by nucleosynthesis in earlier generations of stars and thrown into the interstellar medium late in those stars' lives. Every rock, every drop of seawater, every zircon in the Jack Hills was drawn from that two percent. A cloud like this does not fall in on itself simply because it is heavy.

Internal pressure holds it up, and the collapse has to be triggered by something, whether internal or external. The bipolar outflows of young stars can do it, blowing material into denser knots. So can a supernova explosion whose shockwave sweeps through and compresses relatively dense regions. The nebular hypothesis places the solar system's own collapse most likely at the edge of a bubble blown by a Wolf-Rayet star. Some scientists have gone as far as naming the hypothetical star whose death created the presolar nebula, calling it Kohoutek, though the name marks an inference rather than an object anyone has seen.

The idea that a planetary system condenses out of a rotating cloud is old. The nebular hypothesis was first developed in the eighteenth century by Emanuel Swedenborg, Immanuel Kant, and Pierre-Simon Laplace, and its later development pulled in astronomy, chemistry, geology, physics, and planetary science together. None of it could have begun before heliocentrism was generally accepted, a process that started with Nicolaus Copernicus in 1543 and continued through the scientific revolution. The term Solar System itself is first recorded in 1704, which is a reminder of how recently the arrangement of the sky became a thing with parts.

The hypothesis fell out of favor and then back into it. Its most serious difficulty was an apparent failure to explain why the Sun holds so little of the system's angular momentum compared with the planets. From the early 1980s onward, studies of young stars showed them surrounded by cool disks of dust and gas, exactly as the model predicted, and the model was accepted again. Since the dawn of the Space Age in the 1950s and the discovery of exoplanets in the 1990s, it has been challenged and refined against a growing catalog of other systems forming elsewhere.

About 4.6 to 4.5 billion years ago, the pre-solar nebula began to contract, possibly set off by the shockwave from that nearby supernova. A shockwave arriving from one side does more than squeeze. It also imparts a turn, and the nebula would have begun to rotate. From then on, rotation governed the geometry of everything that followed. As the cloud fell inward, it lost potential energy, heated up, and gained kinetic energy. Conservation of angular momentum did the rest. Material falling toward a spinning center cannot fall straight in, so the cloud flattened along its rotation axis and spread out perpendicular to it, becoming a disk.

This is the same behavior that turns a slow, ragged swirl into a tight spiral as it drains, and the same behavior observed today around newly forming stars. Angular momentum, gravity, and inertia together shaped the protoplanetary disk out of which the planets would later be assembled. At the very center, the gas carried relatively little angular momentum, and so it could fall almost directly. That material underwent fast compression and formed a hot hydrostatic core, a region no longer contracting, holding only a small fraction of the original mass.

The core was the seed of the star. Around it, the infalling envelope kept spinning up as it drew closer, and its rotation eventually built the disk. The initial collapse of a protostellar nebula of about one solar mass takes around 100,000 years, which is a brief interval by the standards of everything else in this account, roughly the time it takes a mountain range to lose a few tens of meters to weather. Statistically, the Sun was almost certainly not born alone. Stars of this kind form in clusters numbering from a few tens to a few thousand, and two kinds of cluster are possible.

A rich cluster resembles the Orion Nebula, crowded with hot, massive stars pouring out ultraviolet light and X-rays. A poor cluster sits buried deep inside its molecular cloud with no massive neighbors nearby. The difference leaves a trace in the outer solar system. In a rich cluster, the sun's disk would have been truncated to only about fifty astronomical units unless the sun sat on the outskirts, and the presence of distant objects, such as the dwarf planet Sedna, makes that scenario unlikely.

The siblings, whatever they were, drifted apart long ago and cannot be identified from here. Meanwhile, the first solid matter in the system was already condensing out of the disk. The oldest inclusions found in meteorites are thought to trace that first solid material, and they date to 4568.2 million years, which serves as one definition of the age of the solar system. They survive as chondrules and as calcium and aluminium-rich inclusions inside certain meteorites dated to about 4.567 billion years or 4567.30 million years with an uncertainty of 0.16 million.

Those specks predate the planet. They are the oldest objects that can be laid on a laboratory bench, and they are older than the oldest zircon in the Jack Hills by roughly one hundred and sixty million years. Deuterium ignited first. As the compressed center of the disk kept contracting, its core reached the temperature at which deuterium, the heavy form of hydrogen, will fuse, and that reaction began before ordinary hydrogen burning was possible. What followed depended entirely on mass.

Above roughly eighty times the mass of Jupiter, hydrogen fusion sets in behind the deuterium, and the object becomes a star. Below that line, contraction stalls, and the result is a brown dwarf, a body that shines on stored heat and slowly dims. The sun cleared the threshold. Hydrogen began fusing into helium, and after further contraction, a T Tauri star was burning at the center of the system. The birth of a new star of this kind occurs roughly one hundred thousand years after collapse begins.

Objects at that moment are called Class I protostars or young T Tauri stars or evolved protostars, depending on which part of the literature is speaking. By then, the star has already gathered nearly everything it will ever hold. The disk and whatever remains of the infalling envelope together come to no more than ten to twenty percent of the mass of the central object. The planets, the asteroids, the comets, and every rock that would ever be walked on were being assembled out of the leftovers.

The envelope did not last. Material kept draining onto the disk until the surrounding cocoon thinned and turned transparent. And at that point, the young stellar object became visible from outside for the first time, first in far infrared light and later in ordinary visible light. What emerged was a classical T Tauri star, a star still wearing its disk, still feeding. The disk at this stage carries something like one to three percent of the mass of the star itself, and it drains inward at a rate between ten to the minus seventh and ten to the minus ninth solar masses per year.

Put in plainer terms, the star was swallowing a mere sliver of a sun every million years, a slow and steady drip rather than a flood. That drip was spectacular where it landed. Accreting gas does not simply settle onto the star from all sides. It follows the magnetic field down and strikes the surface near the magnetic poles, and the impact lights up emission lines in the spectrum. Those lines can carry a flux as strong as the entire intrinsic luminosity of the star. Accretion accounts for the whole cluster of odd behaviors that define classical T Tauri stars.

The blazing emission lines, the magnetic activity, the way brightness flickers and varies from night to night, and the jets. A pair of bipolar jets is usually present, shooting out along the rotation axis in opposite directions. They are byproducts of the accretion, and they perform a specific job. Falling material has too much spin to fall the whole way, and the jets throw the excess angular momentum back out into space. Something has to carry away the rotation that infalling gas cannot keep, and in a young star, this is what does it.

The classical T Tauri stage runs for about ten million years. Around a million years into the described transition, such stars become weakly lined T Tauri stars, their emission fading as the supply thins, and from there the change is far slower. Over hundreds of millions of years, depending on initial mass, a weakly lined T Tauri star settles into an ordinary Sun-like star, unremarkable and steady. The disk did not survive that settling. Four processes emptied it. Some of the material fell onto the star.

Some was locked away into planets. Some was flung out by the jets. The rest was photoevaporated, boiled off by ultraviolet radiation from the central star and from neighboring stars in the same stellar nursery. This is the clock that governs everything that follows. Planet building had a deadline written into it because the raw material was being removed from every direction at once, and whatever had not been assembled into solid bodies by the end was simply gone. In the outer part of the nebula, the assembling had already begun.

Gravity gathered matter around density perturbations and around dust particles, and the disk started separating into rings, bands of concentrated material with cleared lanes between them. But gravity is not what starts the process at the smallest scale. It is far too weak to matter between two specks of dust. The first stage of building a terrestrial planet is almost entirely microphysical. Gas and dust grains collide and stick, held by van der Waals forces and by electromagnetic forces, the same kinds of attraction that make fine powder cling to a glass surface.

Out of that, micrometer-sized particles form. Nothing about this stage is gravitational in any meaningful sense. It is chemistry and static and the sheer number of encounters over long spans of time in a crowded disk. Then the process runs into a wall that has never been fully explained. Planetesimal formation in the centimeter to meter range is not well understood, and no convincing account exists of why grains at that size would accumulate rather than simply rebound off one another. The problem has a name, the meter size barrier.

As dust particles grow by coagulation, they acquire increasingly large relative velocities with respect to their neighbors. They also develop a systematic inward drift, sliding steadily toward the star as the gas around them holds them back. Faster encounters mean destructive collisions. Growth reaches some maximum size and then stops or reverses, and the object that should have become a boulder is shattered back into grains and spiraling inward toward the fire. Something evidently gets past this because the planets exist.

Ward offered a partial answer in 1996. When slow-moving grains collide, the very low yet genuinely non-zero gravity of the colliding grains impedes their escape. Two clumps that touch gently are held together by an attraction almost too faint to measure, and almost too faint is not the same as nothing. The suggestion covers part of the difficulty and not the whole of it. How objects grow to planetesimals of zero point one to one kilometer across, which is roughly a tenth of a mile to two-thirds of a mile, remains an open question in the physics of planet formation.

Past that barrier, whatever carries matter across it, the arithmetic changes sign. Self-accretion of cosmic dust accelerates growth into boulder-sized planetesimals, and once bodies are large enough for their own gravity to reach out, size becomes an advantage. The more massive planetesimals accrete the smaller ones. Others still shatter in collisions and return their material to the pool. Small perturbations from those impacts, together with the angular momentum of other large debris, nudged orbits into the configurations from which kilometer-sized protoplanets could form, circling the center of the nebula.

From there, the growth turns steep. In runaway accretion, successively larger fragments of dust and debris clump together, and the largest bodies sweep up material fastest because they present the biggest gravitational target. A body that pulls ahead keeps pulling ahead. The disk, which had been a fog of countless small things, thinned into a smaller number of large things on separate orbits. Underneath it all, dynamical friction and similar processes in the disk allowed orbiting gas to shed angular momentum and fall inward rather than circling forever.

The end of this phase came from the star. The solar wind of the newly formed T Tauri star swept out most of the material in the disk that had not already condensed into larger bodies. Everything loose went. What was left was solid, held together and orbiting, the raw inventory of a planetary system. The theory behind that account was assembled over three centuries, and much of the hard work happened within living memory. The eighteenth-century nebular idea described a system condensing out of a rotating cloud, but it said nothing about how a speck of dust becomes a world.

That gap stayed open for a very long time. Otto Schmidt closed part of it in 1944, proposing that Earth and the other terrestrial planets formed by accretion from meteoric material. The claim was specific and physical. Planets are made of the same stuff that still falls out of the sky, gathered slowly rather than spun off in one dramatic event. Others tried different routes. William McCrea proposed a protoplanet theory in 1960. Michael Woolfson proposed a capture theory in which material was drawn from elsewhere rather than condensed in place.

In 1978, Andrew Prentice went back to the original Laplacian ideas about planet formation and developed what became known as the modern Laplacian theory. None of these models proved completely successful. Many of them were descriptive rather than predictive, which is a polite way of saying that they told a story about what might have occurred without generating numbers that could be checked. A description can be elegant and still leave a researcher with nothing to test. The line that survived is Schmidt's, and it survived because someone made it quantitative.

In 1969, Victor Safronov took the accretion model and calculated in detail the different stages of terrestrial planet formation, working out how populations of small bodies grow, collide, and merge. That was the change in kind. Once the process had equations attached, it could be run, and disagreements could be settled by computation rather than by preference. Since then, the model has been developed further through intensive numerical simulations of planetesimal accumulation, which is how the subject is studied to this day.

The meter size barrier is a problem precisely because it is one of the places where those simulations still cannot show their work. Out of that inventory of swept up solids in one particular band of orbits, the Earth accumulated. It formed by accretion from the solar nebula around four point five four billion years ago, which places its origin at approximately one-third the age of the universe. The date carries its usual one percent of slack, and the sources are frank about that. How long the assembly took is a different question and a genuinely unsettled one.

One account holds that the process was largely completed within ten to twenty million years. Another, working from nebula theory, estimates that the primordial Earth likely took anywhere from seventy to one hundred million years to form. In June 2023, scientists reported evidence that Earth may have formed in just three million years, far faster than the ten to one hundred million years previously assumed. These figures have not been reconciled with one another. They sit in the literature side by side, differing by a factor of thirty, and the honest position is to hold all of them loosely.

There is one hard anchor at the top of the scale. The calcium- and aluminium-rich inclusions and chondrules preserved in certain meteorites, the oldest solid material anywhere in the solar system, were already dated in the account of the disk. Set that figure of about four thousand five hundred and sixty-eight million years against a finished planet at four thousand five hundred and forty million years, and the entire construction of a world occupies a gap of under thirty million years.

Whether that gap is comfortable or impossibly tight depends on which of the assembly estimates is correct, which is exactly the unresolved point. The building itself was violent in a way that has no modern parallel. During the course of its formation, Earth is thought to have experienced dozens of collisions with planet-sized bodies. Not passing strikes from stray rubble, but encounters with objects large enough to count as worlds in their own right, arriving one after another over millions of years.

Each delivered its mass and, more importantly, its kinetic energy, which had nowhere to go except into heat. Much of Earth was molten as a result, and the frequent collisions drove extreme volcanism across a surface that had no chance to settle into anything permanent. There were no continents to speak of, no oceans, and no crust worth the name, because nothing stayed solid long enough to become any of those things. Heat, once accumulated, did something to the inside of the planet that changed it permanently.

The proto-Earth kept growing by accretion until its interior became hot enough to melt the heavy siderophile metals. Siderophile means iron-loving, the class of elements that prefer to bond with metallic iron rather than with rock-forming silicates. Molten and considerably denser than the silicates around them, those metals began to sink. Droplets of liquid metal worked their way down through the hot interior toward the center, and the more that sank, the more gravitational energy was released as heat, and the more of the interior melted in turn.

This runaway sorting is known as the iron catastrophe, and it was over quickly. The separation of a primitive mantle from a metallic core was accomplished only about ten million years after Earth began to form. That single episode gave the planet its layered structure, metal in the middle, rock above it, everything arranged by density instead of mixed at random. It also set up the formation of Earth's magnetic field. The consequences of that are still shielding the surface. A liquid outer core of metal in motion generates a magnetosphere that deflects most of the destructive solar wind and cosmic radiation heading toward the planet.

Sorting the interior by density was not simply a matter of interior architecture. It equipped Earth with the field that continues to stand between the surface and a stream of charged particles from the sun. The sorting has never entirely finished. J.A. Jacobs was the first to suggest that Earth's inner core, the solid center distinct from the liquid outer core, is freezing and growing outward from within the liquid as the interior of the planet gradually cools. The rate of cooling is about one hundred degrees Celsius per billion years, which is roughly a tenth of a degree in a million years.

A change too slow to be noticed by any instrument over a human lifetime. Read that in the other direction and it becomes remarkable. Every second, somewhere beneath thousands of kilometers of rock, a little more liquid metal crystallizes onto a solid sphere at the center of the planet. The heat released by that freezing helps to stir the liquid layer above it, and the stirring maintains the field. The iron catastrophe of the first ten million years is not a closed chapter in Earth's history.

It is a process still underway, quietly, under every footstep taken on the surface. By the end of this stage, the planet had a core, a mantle, a magnetic field, and a molten and repeatedly battered surface. It did not yet have a moon. Somewhere in the same orbital neighborhood, on a track very like Earth's own, another world had been assembling out of the same disc. It was roughly the size of Mars. Under modern theories of planet formation, it belonged to a population of such bodies present in the solar system about four point five billion years ago.

Small planets that never grew large enough to dominate their own lane. Most of them were eventually swept up. One of them was not, at least not immediately, because of where it sat. The proposal is that it occupied one of the gravitational balance points of the Earth and Sun system, the positions labeled L4 and L5. A body placed at either of those points, sixty degrees ahead of or behind a planet along its orbit, feels the pull of the sun and the pull of the planet in a combination that tends to hold it there.

It is a shelf rather than a well, and objects parked on it can stay for a very long time. That is where this particular protoplanet is hypothesized to have orbited, sharing Earth's year, neither catching up nor falling behind. The word hypothesized carries real weight in that sentence. The location is inferred from the dynamics, not read off any surviving trace. The object has a name which is unusual for something no one has ever observed. The English geochemist Alex N. Halliday proposed calling it Theia in two thousand after the Greek Titan, who was the mother of Selene, goddess of the moon.

The name has been accepted across the scientific community since. It is a convenient label for an inferred body, and it comes with a certain quiet elegance, since the thing being named is understood only through what it is thought to have produced. The date of the collision comes with the same honest spread that attaches to everything in this period. It is given as approximately four point five billion years ago in the early part of the Hadean. It is also given as about twenty to one hundred million years after the solar system formed.

Astronomers working from other lines of evidence place it at about four point four to four point four five billion years ago, roughly zero point one billion years after the solar system began to assemble. These are not competing certainties. They are the range within which the event is bracketed, and the range is- itself is part of the finding. Computer models of what happens when a Mars-sized body strikes a growing Earth produce a consistent picture of where the pieces go. The impactor does not bounce, and it does not simply splash.

Its core, dense metal already sorted to its own center, would likely drive deep into Earth and fuse with Earth's core, adding its metal to the metal already there. The lighter material, the crustal and mantle rock from both bodies, is the part that escapes. Some of it is flung far enough to clear a particular threshold, and the threshold matters more than any other number in this account. That threshold is the Roche limit, the distance inside which a planet's tidal pull is strong enough to stop loose material from gathering itself together.

Debris orbiting nearer than that limit is pulled apart by the difference in gravity across its own width faster than its own weak attraction can knit it. Fragments thrown beyond the limit are free to coalesce. Those that stayed inside it were drawn back down and re-fused with Earth. Those that made it out formed a ring of hot rock, and out of that ring, according to the hypothesis, the Moon accumulated. This was not the only such encounter, and no account of the period should suggest that it was.

It was, however, the last significant one. After it, nothing of comparable size arrived again. The Earth that emerged from the debris had a companion, and the companion has been there ever since. The case for a collision rests on a set of measurements that are individually suggestive and collectively difficult to explain any other way. The first is geometric. The Moon's orbit shares a similar orientation with Earth's rotation, both tilted at a similar angle to the ecliptic plane of the Solar System.

A captured body snatched from elsewhere would not be expected to arrive so neatly aligned with the spin of its captor. The second comes from the laboratory. The stable isotope ratios of lunar rock and terrestrial rock are identical. Isotope ratios function as a chemical postcode in the Solar System, varying with where in the disk a body's material was gathered. Two objects that match that precisely were built from the same stock. Whatever the Moon is, it is not a stranger that wandered in.

The third is a bookkeeping problem. The Earth and Moon system carries anomalously high angular momentum, counting the spin of Earth, the spin of the Moon, and the Moon's motion around Earth. Measured against the other rocky planets, there is far too much rotation in the pair. Angular momentum does not appear from nothing. A large body arriving off-center at speed would deliver exactly that surplus, which is one of the strongest arguments in the hypothesis. The fourth comes from the samples themselves.

Lunar rock indicates that the Moon was once molten to a substantial depth, though how deep remains unknown. Melting a body that thoroughly takes a great deal of energy, plausibly much more than the ordinary accumulation of an object of the Moon's size and mass could ever supply. An extremely energetic process, such as a giant impact, could account for it. The fifth is a matter of density. The Moon has a relatively small iron core and is far less dense than Earth as a result. That is precisely what the models predict if the impact as metal was driven into Earth's interior while the ejected material was drawn from the lighter outer layers of both bodies.

A moon assembled from crust and mantle would be poor in metallic iron, and the Moon is. The sixth follows the same logic in the opposite direction. The Moon is depleted in volatile substances compared with Earth. Volatiles evaporate at comparatively low temperatures, and an event violent enough to vaporize them would have scattered them widely. Earth, with its stronger gravity, retained a share. The Moon, smaller and weaker, could not gather them back. The chemistry of the two bodies diverges exactly where a hot, high-energy origin predicts it should.

Beyond the solar system, the same kind of event appears to be common. Other star systems show debris disks interpreted as the aftermath of similar collisions, rings of pulverized rock circling stars where planets are still under construction. Giant collisions are entirely consistent with the leading account of how the solar system came together, so the mechanism does not need to be invented for this one case. What sits at the end of that argument is visible on any clear night. The Moon orbits at three hundred eighty- four thousand four hundred kilometers, which is two hundred and thirty-eight thousand eight hundred and fifty-five miles, or one point two eight light seconds.

It is roughly a quarter as wide as Earth, a proportion that makes the pair unusual among the rocky planets. Its gravity helps to stabilize the tilt of Earth's axis, which is the reason the seasons keep to a reliable pattern. It raises the tides. It is also very gradually slowing Earth's rotation. The exchange has already run to completion in the other direction, since Earth's pull has locked the Moon's rotation to its orbit, so the same near side faces the planet permanently. The details of the collision, though, remain genuinely open.

Cameron and Ward proposed a tangential impact, a grazing strike in which most of the outer silicates of the colliding body were vaporized while its metallic core was not, so that the material sent into orbit was overwhelmingly rocky and the coalescing moon came out deficient in iron. The more volatile substances released would probably have escaped the solar system altogether, while the silicates tended to gather. Against that, an analysis of lunar rocks published in a twenty sixteen report suggests the impact might have been a direct hit, fragmenting both worlds and mixing them thoroughly.

There is also a larger gap in the middle of the whole picture. No self-consistent model yet exists that begins with the giant impact and follows the evolution of the debris all the way into a single moon. The starting point is well argued and the outcome is overhead, and the passage between them is still being worked out. Reaching even that position took most of a century of argument. In eighteen ninety-eight, George Darwin suggested that Earth and the Moon had once been a single body, and that a molten moon had been spun off from a rapidly rotating Earth by centrifugal force.

It was a physical, testable idea, and it became the dominant academic explanation for decades. Darwin worked it with Newtonian mechanics, and along the way, he calculated something that turned out to be exactly right. The Moon had orbited much more closely in the past and was drifting away. His mechanism, however, would not close. The calculations could not resolve the physics required to trace the Moon back to the surface of Earth, and no amount of refinement made the separation work. A theory can be widely held and still have a hole in the center of it, and this one did.

The Canadian geologist Reginald Aldworth Daly of Harvard University challenged it in nineteen forty-six. His proposal was that the moon's creation was caused by an impact rather than by centrifugal forces. It was a substantial idea offered at a moment when almost no one was inclined to hear it, and very little attention was paid. It stayed quiet for nearly 30 years. The idea was reintroduced at a conference on satellites in 1974, then published and discussed in the journal Icarus in 1975 by William K.

Hartmann and Donald R. Davis. Their models suggested that at the close of the planet-forming period, several satellite-sized bodies existed that could collide with the planets or be captured by them. One of these might have struck Earth and ejected refractory, volatile-poor dust, and that dust could coalesce into the moon. The proposal had a specific virtue. It offered an explanation for the moon's peculiar geology and chemistry rather than treating those peculiarities as awkward extras.

Cameron and Ward were working the same ground with their tangential Mars-sized impactor. Two independent routes had arrived at broadly the same event, which is generally a good sign in physical science, and yet the field did not converge quickly. Three traditional theories still had partisans. A large body of researchers occupied what one participant described as an apathetic middle, unconvinced that the question would ever be settled at all. Eighteen months before a conference on lunar origins scheduled for October 1984, Bill Hartmann, Roger Phillips, and Jeff Taylor issued a challenge to their colleagues that has been quoted ever since.

"You have 18 months. Go back to your Apollo data, go back to your computer, and do whatever you have to, but make up your mind. Do not come to our conference unless you have something to say about the moon's birth." They came to Kona in Hawaii, and the giant impact hypothesis emerged as the most favored explanation. The change in the field was structural rather than merely numerical. Before the meeting, there were several camps and a large indifferent center. Afterwards, there were essentially two, those who accepted the giant impact and those who remained agnostic.

Sixteen years later, Halliday supplied the impactor with its name and a hypothesis that had spent decades as a minority position acquired a character in the story. Darwin's outward drift, the one piece of his work that survived the collapse of his mechanism, was eventually confirmed by direct measurement. American and Soviet experiments used laser-ranging targets placed on the lunar surface, bouncing light off them and timing the return. The moon is receding exactly as the Newtonian calculation implied, and the confirmation arrived long after the theory it was designed to support had been set aside.

The impact left Earth transformed. The energy of a collision on that scale is predicted to have heated the planet enough to produce a global magma ocean, a surface of molten rock with no shore anywhere on it. Evidence of the planetary differentiation that followed with heavier material sinking inward through the mantle has been documented in the rock record, which is one of the more solid observational anchors in this whole stretch of time. A fair fraction of the material involved would have been vaporized outright.

Rock behaves like any other substance given sufficient energy, and above a certain temperature, it becomes a gas. For a while, part of the planet's own mineral content was in the atmosphere, a sky carrying silicate vapor. That condition did not last. The rock vapor would have condensed and fallen within about two thousand years, which by the standards of this period is close to instantaneous. Two thousand years is a fraction of the time between the earliest cities and the present. And here it is the interval in which a planet's atmosphere rained back down as stone.

The magma ocean took longer, though not by much. It is thought to have solidified within five million years. A crust formed over molten rock broke and reformed, and the whole ocean of melt gradually froze from a global sea into something that could hold a shape. The impact did not, however, reset the planet entirely. Earth's present composition suggests that no complete remelting occurred because huge masses of rock are extremely difficult to melt and mix through. Heat does not distribute itself evenly through thousands of kilometers of silicate in a hurry.

The current reading is that the collision melted only one or two large regions of the planet rather than liquefying every part of it. Some of what Earth had already become survived the arrival of Theia intact, buried and unmixed, which is part of why the planet's chemistry can still be read at all. As the magma ocean froze, it gave up what it had been holding. Molten rock carries dissolved volatiles, and solidifying rock does not. So those substances were driven out into the air. The result was probably a heavy carbon dioxide atmosphere with hydrogen and water vapor in it.

The conditions beneath that atmosphere are among the strangest in the planet's history. The surface temperature stood at about two hundred and thirty degrees Celsius, which is four hundred and forty-six degrees Fahrenheit, and the atmospheric pressure exceeded twenty-seven standard atmospheres. Earlier in the Hadean, before the magma ocean froze, the air is thought to have been thicker still and rich in hydrides. Its composition resembling the solar nebula itself and the atmospheres of the gas giants.

Mostly water vapor with methane and ammonia. Nothing about it was breathable by any standard, and free oxygen was effectively absent. Oxygen was not entirely impossible to make, only impossible to keep. Shortwave ultraviolet light in sunlight could split surface water molecules into oxygen and hydrogen, a process called photodissociation. In an atmosphere that reducing the liberated oxygen reacted away almost at once, binding into compounds rather than accumulating as a gas. The hydrogen took the other route.

Being extremely light, it drifted upward and leaked away into space through atmospheric escape as it and helium continue to do to this day. Water was being taken apart, and the pieces were being disposed of by two entirely different mechanisms. The air itself is thought to have come from below. Volcanic outgassing probably created the primordial atmosphere and then, in time, the ocean. What is certain is that the exact composition of Earth's earliest air is not known, and the descriptions above are the likely bulk rather than a measured analysis.

There is no sample. There is a planet that behaved in a particular way afterwards and a set of chemical constraints that narrow the possibilities, and the account arrives somewhere just short of certainty, bounded by chemical constraints rather than a direct sample. Cooling turned the sky into weather, for as long as the surface stayed hot enough, the water in that heavy atmosphere could only exist as vapor, held aloft by the temperature of the rock beneath it. As the surface gave up its heat, the balance tipped.

Vapor condensed. Rain began to fall on a world that had never had rain, and much of it flashed back into steam on contact and fell again and again until the rock at last stayed cool enough to let water stand. What gathered was not a sea in any familiar sense. It was a super ocean covering very nearly the whole planet, and it turned Earth into an ocean world before there was any shore to describe. The physics of that ocean is genuinely strange to modern intuition. Water at ordinary sea-level pressure boils at one hundred degrees Celsius, so a surface hundreds of degrees hot should have no liquid on it at all.

Pressure changes the rule. Under the weight of an atmosphere exceeding twenty-seven standard atmospheres, the boiling point climbs far above its familiar value, and water stays liquid at temperatures that would flash it to steam today. The early ocean existed because the air above it was heavy enough to hold it down. A pot lid scaled to a planet. Where the water came from is a question with a clear favorite and a clean piece of counter-evidence. A sizable quantity of water was almost certainly present in the material that built Earth in the first place.

Mixed through the accreting solids rather than delivered afterwards, holding onto it was the difficulty. While the planet was still small, its gravity was weaker, and water molecules escaped more easily than they would from the finished world. The most likely source of the Hadean Ocean is outgassing from the mantle, the same volcanic exhalation that supplied the air. Water that had been locked in the interior worked its way out through eruption after eruption and accumulated on the surface.

The obvious alternative that comets and other icy bodies brought the ocean in from outside runs into a chemical objection. Water carries a signature in the proportions of its hydrogen isotopes, and that ratio varies between reservoirs in the solar system. Comets measured so far do not match Earth's ocean well enough. A bombardment origin for a substantial part of the water is therefore considered unlikely, though impacts certainly delivered some. The ocean appears to have come mostly from below rather than mostly from above.

Once standing water existed, it began to work on the air. Carbon dioxide dissolves readily into water, and a global ocean in contact with a carbon dioxide-rich atmosphere would have started pulling that gas out of the sky. Absorption alone, however, was not enough. The volume of gas involved was too large, and dissolved carbon dioxide reaches an equilibrium rather than disappearing indefinitely. Something else had to remove the carbon permanently, and taking it out of circulation meant taking it out of the ocean as well.

That is where the interior of the planet reenters the account. Mantle convection during the Eon named for the underworld was likely far more vigorous than anything happening beneath the surface now, and the reason is viscosity. Radiogenic heat levels were high since the short-lived and long- lived radioactive isotopes had barely begun to decay away. Water in the mantle had not yet fully outgassed, and water in silicate rock lowers its resistance to flow considerably. Hot, wet mantle is a runnier mantle.

Convection in it would have been faster and more turbulent than the slow overturn of the present day. Vigorous convection does not automatically mean plate tectonics, and this is one of the live disputes of the period. A mantle can stir energetically beneath a single unbroken shell of rock, a configuration described as a rigid lid in which heat escapes without the surface ever fracturing into separate moving pieces. Whether Hadean convection broke the surface into plates or was confined beneath such a lid remains debated.

Venus offers a nearby example of a planet where the lid apparently never gave way. One argument for plates points at the ocean itself. The presence of Hadean oceans is thought to have triggered plate tectonics since water alters rock at the surface and along fractures, weakening it and lubricating the places where one slab might begin sliding under another. A dry world resists that. A wet one is easier to break. If subduction did begin, it solved the carbon problem. Carbon dioxide dissolved in seawater combines into carbonate.

Carbonate settles into sediment on the sea floor, and a subducting plate carries that sediment down into the mantle. The gas is not merely relocated. It is buried. Subduction due to plate tectonics would have removed carbonate from the early oceans, and that removal contributed to stripping away the carbon dioxide-rich early atmosphere over time. The eventual disappearance of that heavy air is itself treated as evidence for Hadean plate tectonics, which makes the reasoning circular in a way that geologists acknowledge openly.

The atmosphere went somewhere. Subduction is the mechanism capable of taking it. Plates, if they existed, would have made something else as well. Subduction generates the melts that build continental crust, granitic and buoyant, and permanently afloat above the denser rock beneath. So the tectonic question becomes a question about land, and the models disagree sharply about how much of it there was. The work of Dieu and colleagues predicts that by the end of the Hadean, the continental crust amounted to only about twenty-five percent of today's area, a scattering of small provinces in an otherwise global ocean.

The models of Korenaga and colleagues predict something far bolder. Continental crust reaching present-day volume sometime between four point two and four point zero billion years ago, meaning the raw material of every later continent was already made even if it was distributed differently. Between a quarter and all of it is a very wide gap, and it exists because the direct evidence is so thin. Some geologists hold that the mineral signatures read as proof of early plate tectonics could have been produced by meteorite impacts instead, since impacts also generate heat, pressure, and melt.

The general picture that geophysical models paint is under-constrained enough to accommodate more than one story, and the samples that might decide between them are scarce. Impacts, in any case, were not a marginal detail of the period. They may have been the dominant process shaping the inner solar system for hundreds of millions of years after the planets were assembled, and the best record of that is not on Earth at all. The late heavy bombardment, also called the lunar cataclysm, is a hypothesized interval roughly four point one to three point eight billion years ago, during which a disproportionately large number of asteroids and comets struck the terrestrial planets and their satellites.

Mercury, Venus, Earth, the Moon, and Mars all took part. The impactors are held to have come from two sources, a leftover post-accretion population and a second group driven inward by instability among the planets themselves. On Earth, the interval corresponds to the Neo-Hadean and Eo-Archean eras, and the boundary between the Hadean and the Archean eons sits inside it. The evidence comes from moon rock. Earth erases its craters through erosion, sedimentation, and tectonics, so the terrestrial record of any such episode is gone.

The moon keeps everything. Apollo astronauts collected samples from lunar craters and isotopic dating of those rocks showed that they were last molten during impact events falling within a rather narrow interval of time, which suggested that a large proportion of the moon's craters formed during a single episode rather than accumulating steadily. The specific evidence is a category of sample called impact melt rock. A collision large enough leaves behind rock that was liquefied by the energy of the strike and then froze again, and the freezing resets its radiometric clock.

Such a rock records the date of the impact rather than the date the original material formed. The majority of the Apollo impact melts are thought to have been produced by collisions with asteroids or comets tens of kilometers across, forming craters hundreds of kilometers wide. The landing sites were not chosen at random. Apollo 15, 16, and 17 were sent for their proximity to the Imbrium, Nectaris, and Serenitatis basins respectively, three of the great multi-ring scars on the lunar near side.

When the melt ages from those missions were assembled, they clustered between about three point eight and four point one billion years. Investigators drew the obvious inference and postulated an intense bombardment of the moon with a dramatic increase in the impact rate around three point nine billion years ago. If those melts really did derive from the three named basins, then on stratigraphic grounds, many other prominent basins must also have formed within the same short window. Extrapolating the lunar cratering rate across to a larger, more massive planet gives numbers that are difficult to picture.

Earth would have received twenty-two thousand or more impact craters wider than twenty kilometers, which is twelve miles. It would have taken about forty impact basins measuring around one thousand kilometers across, roughly six hundred and twenty miles, each of them larger than any structure on the modern surface, and it would have taken several basins near five thousand kilometers wide, about three thousand one hundred miles, features on the scale of an ocean floor. None of them survives.

Every one has been recycled through the mantle or buried past recognition. Explaining why the flux should spike at all has proved harder than documenting it, and no consensus exists. The Nice model is the most popular proposal among planetary scientists. It holds that the giant planets underwent orbital migration, shifting their positions after formation, and that the migration scattered objects out of the asteroid belt and the Kuiper Belt into eccentric orbits that crossed the paths of the inner planets.

A rearrangement in the outer solar system on this account arrived as a rain of stone across the inner one. There is supporting chemistry from a different family of samples. Howardite, eucrite, and diogenite meteorites, along with H chondrites, all originate in the asteroid belt, and dating them reveals numerous ages between three point four and four point one billion years, with an earlier peak at four point five billion. The younger cluster has been interpreted as recording an increase in impact velocities rather than simply an increase in impact numbers.

Computer simulations using hydrocode show why that would matter. The volume of impact melt produced rises by a factor of one hundred to one thousand as impact velocity climbs from the current asteroid belt average of five kilometers per second to ten kilometers per second. Speeds above ten kilometers per second require very high orbital inclinations or the large eccentricities of asteroids on planet-crossing orbits, both rare in the belt today and both greatly increased by the sweeping of orbital resonances that giant planet migration would cause.

The same projectiles appear to have reached the innermost planet. Studies of highland crater size distributions suggest that one family of impactors struck both Mercury and the moon during the episode. If Mercury's bombardment decayed on the same schedule as the moon's, then Caloris, the youngest large basin found there, is comparable in age to Orientale and Imbrium, the youngest large lunar basins, and all of Mercury's plains units are older than three billion years. For all that, the cataclysm remains a hypothesis, and its foundations are more debatable than its popularity suggests.

It has gained widespread credence, particularly among dynamicists, and definitive evidence remains elusive. The first serious objection concerns the sample set. Apollo impact melts have commonly been attributed to whichever basin lies nearest the landing site, but Imbrium is the youngest and largest of the multi- ring basins on the central near side, and quantitative modeling shows that significant Imbrium ejecta should be present at all of the Apollo landing sites. If a large portion of the collected melt was thrown from that single event, then the tight cluster at three point nine billion years records one impact rather than an epoch of them.

The astronauts would have sampled the same catastrophe six times over and called it a pattern. A related criticism targets the dating method itself. The three point nine billion year spike identified in argon-argon measurements could be produced without any bombardment at all by episodic early crust formation followed by partial losses of argon forty as the impact rate declined. The apparent peak would then be an artifact of how the gas left the rocks, not of when the rocks were struck. The second objection concerns absence.

Impact melts older than about 4.1 billion years are missing from the collection, and missing is not the same as never made. Older melt rocks may well have existed and then had their radiometric ages reset by four billion years of continued cratering, each later impact partially rewinding the clock of the material it disturbed. Or they may have been pulverized, ground down by repeated small impacts into fragments too fine for standard radiometric dating. A record can be erased by the very process it is supposed to record.

Lunar meteorites were meant to settle this, and they only partly did. Rocks blasted off the moon and eventually falling to Earth sample the surface far more randomly than any landing site can, and many feldspathic lunar meteorites probably came from the far side, terrain no mission has visited. Impact melts inside them have been dated. Consistent with the cataclysm, none proved older than about 3.9 billion years. Inconsistent with it, their ages do not cluster at that date, but spread across the interval from 2.5 to 3.9 billion years, which looks less like a single violent episode than like a long tail of ordinary impacts thinning out over time.

A range of evidence points that way. Rather than a sharp spike, there may have been an extended period of lunar bombardment lasting from approximately 4.2 billion years ago to 3.5 billion years ago, heavy at the start and gradually easing. The lunar cataclysm was considered controversial when it was first proposed, became more popular over the following fifty years, and is still controversial now, resting on assumptions that can reasonably be questioned. Scientists continue to study the moon's bombardment history precisely because it is the clock by which the early history of the whole inner solar system is read.

The framing matters for Earth as well, because the hypothesis changed what geologists expected of the Hadean surface. Before it was formulated, the general assumption was that Earth remained molten until about 3.8 billion years ago, a figure that could be read off many of the oldest known rocks. The idea of a late spike offered an alternative, a planet that cooled early and was then repeatedly disturbed rather than one that simply stayed liquid for three-quarters of a billion years. Disturbed is the right word for what impacts did to the ocean.

Asteroid strikes during the Hadean and on into the Archean would have periodically wrecked it, and the geological record from 3.2 billion years ago preserves evidence of multiple impacts by objects up to one hundred kilometers in diameter, which is sixty-two miles. An object of that size arriving at planetary speeds delivers enough energy to change the state of a significant fraction of the world's water. Each such impact could boil off up to one hundred meters of a global ocean, three hundred and thirty feet of seawater lifted into the atmosphere as steam.

The air temperature would rise temporarily to about five hundred degrees Celsius, which is nine hundred and thirty-two degrees Fahrenheit. A sky of superheated vapor standing over a scolded and shallowed sea. Then the energy would radiate away. The vapor would condense, and the water would come back down. The ocean refilled. Whatever chemistry had been running in the shallows had to begin again or survive somewhere deep enough and sheltered enough to ride the episode out. How often this happened is still under study.

The frequency of meteorite impacts through the period is not well constrained, and the possibility that matters most is a gap. Earth may have gone through long stretches of quiet in which liquid oceans stood undisturbed and conditions suitable for life persisted, punctuated at intervals by an event that sterilized the surface layers and reset the experiment. A history of long calm and rare violence looks very different from a history of continuous chaos, and the rock record from that time is too sparse to distinguish them confidently.

One more complication sits underneath all of this, and it concerns the sun. Four billion years ago, the sun shone at only about seventy percent of its present brightness because a young star of that class produces less energy than a middle-aged one. Less sunlight should mean a colder planet. Applying modern atmospheric physics to that dimmer sun gives a frozen Earth, an ocean locked under ice long before life could have begun in it. The rocks say otherwise. Liquid water was present, and the ocean persisted.

The mismatch has a name, the faint young sun paradox, and it is not resolved. The available answers all involve the atmosphere rather than the star. Either carbon dioxide levels were far higher than the present, trapping enough energy to compensate for the weaker sunlight, or other greenhouse gases were doing the work. Methane, in particular, produced later by microbial metabolism. Scientists continue to research how the Earth stayed warm enough for water before life existed to help. A dim sun, a warm sea, and no agreed mechanism between them.

Into that unresolved warmth, chemistry began doing something new. Abiogenesis is the name for it, sometimes called biopoesis, and the definition is deliberately modest. The natural process by which life arises from non-living matter, from simple organic compounds. The prevailing scientific view is that no single moment can be pointed at. What happened was a rise in complexity, stage upon stage. A habitable planet formed, supplied with minerals and liquid water, and steady inflows of free energy from sunlight above and geothermal heat below.

Prebiotic synthesis then produced a range of simple organic compounds. Those were assembled into polymers such as proteins and RNA. Somewhere along that ladder came molecular self-replication, self-assembly, autocatalysis, and finally, the emergence of cell membranes, which closed a patch of chemistry off from the rest of the world and gave it an inside. Everything alive since has run on carbon and water and on four families of molecules. Lipids build the membranes. Carbohydrates, the sugars, supply structure and fuel.

Amino acids assemble into proteins, which do most of the work of metabolism. The nucleic acids, DNA and RNA, carry heredity. Any successful theory of the origin of life has to account for where all four came from and how they began interacting, which is a considerably harder demand than explaining any one of them alone. The most favored route through that difficulty is called the RNA world. RNA has an unusual double talent. It can store information in a sequence as DNA does, and it can also fold into a shape that catalyzes chemical reactions as a protein does.

A molecule of RNA acting as a catalyst is called a ribozyme. If early RNA could both hold a recipe and speed up its own copying, then heredity and metabolism did not need to be invented separately. Researchers think life descends from such a world while allowing that other self-replicating and self-catalyzing molecules may have come before RNA and left no trace. A different family of ideas starts from the opposite end. The metabolism-first hypotheses set the question of information aside and ask how catalysis on the early Earth could have supplied the precursor molecules in the first place.

Reactions on mineral surfaces driven by heat and chemical gradients might have built up a self- sustaining network of chemistry long before anything resembling a gene existed. Replication in that account arrives late into an environment already busy with organic reactions. The setting is just as open as the mechanism. One proposal places the earliest cells at a deep sea white smoker hydrothermal vent, where a natural difference in acidity across a thin mineral wall provides a proton gradient, effectively a chemical voltage supplied free by the planet.

A cell with a leaky membrane could tap that gradient instead of generating one for itself. Another proposal puts the origin inside the continental crust in warm water threading through rock. A third puts it in surface water exposed to sunlight and to the slow evaporation and rewetting of shallow pools. Geochemical and fossil evidence informs most of these studies, and none of the three settings has been ruled out. One experiment made the question tractable. In 1952, the Miller-Urey experiment demonstrated that amino acids can be synthesized from inorganic compounds under conditions like those of the early Earth.

Gases, water, and a source of energy sealed in glassware produced the building blocks of proteins without any living thing present. The result did not create life, and it was never claimed to. What it did was move the first step of the problem out of speculation and into the laboratory. The universe then supplied a second confirmation from an unexpected direction. Amino acids have since been found in meteorites, in comets, in asteroids, and in star-forming regions of space. The molecules that build proteins are not rare, and they are not restricted to planets.

They form in cold clouds and in the interiors of small bodies, which suggests that the raw ingredients were delivered to the early Earth from several directions at once and were probably being made on the surface as well. NASA defines life as a self- sustaining chemical system capable of Darwinian evolution, a definition that says nothing about cells or carbon and everything about behavior. Reproduction with heritable variation is the whole requirement. The twenty fifteen NASA strategy on the origin of life framed the research program accordingly.

Identify the interactions, the intermediary structures and functions, the energy sources, and the environmental factors that contributed to evolvable macromolecular systems, and map the chemical landscape of potential primordial informational polymers. The advent of such polymers was most likely the critical step, and those polymers derived in turn from simple compounds, from nuclear bases and amino acids and sugars formed by ordinary reactions in the environment. Between that chemistry and the first organism whose descendants are still alive, there is a gap that has never been closed.

The organism at the far side of the gap has a name and an acronym. LUCA, the last universal common ancestor, is the most recent organism from which every living thing today descends. It was not the first life. LUCA existed millions of years after life began, which means an unknown depth of earlier biology lies behind it, all of it now extinct without issue. LUCA was presumably a single-celled organism, and it lived some four billion years ago. Its portrait has been assembled without a single fossil by comparing genomes.

Life today divides into major branches, and two of them, the archaea and the bacteria, are the deepest. A gene present in both branches is likely to have been present in their shared ancestor. Applying that logic across sequenced genomes turns up about sixty proteins common to all life and three hundred fifty-five prokaryotic genes that trace back to LUCA. Their functions add up to a description of an organism. The description is not simple. LUCA already had hundreds of genes encoded in the DNA genetic code that remains universal today.

That implies a full suite of cellular machinery, messenger RNA to carry instructions out of the genome, transfer RNA to bring in the correct amino acids, and ribosomes to translate the code into proteins. It implies enzymes running anaerobic respiration by the Wood-Ljungdahl pathway, a route that builds organic carbon without oxygen. It implies energy derived by chemiosmosis drawn from a gradient across a membrane, and it implies a DNA polymerase, the enzyme that copies genetic material. That last item contains the sharpest version of the whole problem.

A cell copies its DNA using DNA polymerase. DNA polymerase is itself produced by translating the DNA polymerase gene, which is written in the DNA. Neither the enzyme nor the molecule that encodes it can be produced without the other already present and working. The system is tightly interlinked, and at first glance, every part of it is necessary for any part of it to function, which is precisely the arrangement that gradual evolutionary steps are supposed to be unable to produce. The RNA world is one proposed way out.

Since a molecule that both stores and catalyzes removes the need for two separate systems to appear simultaneously, self-assembly of membranes and autocatalysis in a ribozyme environment fill in other steps. Some scientists take the view that life and the origin of life are aspects of the same process rather than two different subjects with a boundary between them. What can be said cleanly is where the difficulty sits. Everything after LUCA is readily understood. Biological evolution produced the whole range of species, forms, and biochemical capabilities that followed, and the mechanism is not in dispute.

Deriving LUCA from simple components is far from understood. The transition from non-life to life has never been observed experimentally, and many competing proposals cover the different stages. Earlier centuries had answers that felt easier. Spontaneous generation, held from Aristotle until the nineteenth century, proposed that lower animals such as insects were generated directly from decaying organic matter and that life arose by chance out of rot. It was questioned from the seventeenth century onward in works such as Thomas Browne's Pseudodoxia Epidemica.

In sixteen sixty-five, Robert Hooke published the first drawings of a microorganism, and the world acquired a scale of living things it had not known was there. In 1668, Francesco Redi showed that no maggots appeared in meat when flies were prevented from laying eggs on it, which is as clean a controlled experiment as the period produced. In 1676, Antonie van Leeuwenhoek drew and described microorganisms, probably protozoa and bacteria, and he disagreed with spontaneous generation from the start.

By the 1680s, he had convinced himself it was wrong, using experiments that ranged from sealed and open incubations of meat to patient close study of how insects actually reproduce. By the middle of the 19th century, the idea was considered disproven. The other old answer is still with us in modified form. Panspermia, traceable to Anaxagoras in the fifth century before the common era, holds that life originated elsewhere in the universe and arrived here. The modern version proposes delivery by meteoroids, asteroids, comets, or planetoids.

The observation that undercuts it is simple enough. Shifting the origin to another heavenly body relocates the question rather than answering it. Something, somewhere, still had to make the first self-sustaining chemistry out of ordinary matter. Earth remains the only place known to harbor life, and astrobiologists proceed on the assumption that similar processes operate on other worlds while looking for the evidence that would confirm it. The rock record picks the story up well after LUCA, and it picks it up cautiously.

The earliest undisputed evidence of life dates from at least three point five billion years ago in the Eoarchean era, by which time a geological crust had solidified and was capable of preserving something. In Western Australia, sandstone three point four eight billion years old contains microbial mat fossils, stromatolites, the layered structures built by communities of microbes trapping and binding sediment grain by grain. They are not fossils of individual organisms. They are fossils of a habit of many generations of cells living in sheets on a shallow sea floor and leaving behind a laminated mound that hardened into stone.

Older evidence exists in a subtler form. Metasedimentary rocks three point seven billion years old in southwestern Greenland contain graphite of probable biogenic origin. Graphite is simply carbon, and carbon by itself proves nothing. But living chemistry sorts carbon isotopes in a distinctive way, and that signature can survive long after every trace of structure has been cooked out of the rock by heat and pressure. The Greenland material carries no shape to look at. It carries a proportion.

Further back, the claims become genuinely contentious, which is the correct condition for material at that depth. In twenty fifteen, traces of carbon minerals interpreted as remains of biotic life were found in four point one billion- year-old rocks in Western Australia. If that reading holds, life was already present in the Hadean remarkably soon after the ocean condensed. Separately, possible fossil microorganisms from Quebec may have lived in hydrothermal vent precipitates, mineral deposits laid down around vents also soon after the ocean formed.

The earliest evidence is also cited at three point eight billion years, again from Western Australia. Three different figures, three different standards of proof, and all three figures belong in the account, each carried at the weight of the evidence that supports it. The implication drawn from the early dates has reached well beyond geology. If life took hold on this planet quickly, rather than requiring hundreds of millions of years of chance, then the process may not be difficult and may not be rare.

As one of the researchers associated with the earliest evidence put it, "If life arose relatively quickly on Earth, then it could be common in the universe." That sentence is an inference resting on a contested date, resting on carbon minerals in a very old rock, and it remains one of the most consequential inferences in the field. Cyanobacteria are thought to have evolved as early as three point five billion years ago. Photosynthetic organisms appeared somewhere between three point two and two point four billion years ago and began enriching the atmosphere with oxygen.

The span in those dates is wide, and the consequence was not. Chlorophyll-based photosynthesis splits water using sunlight, takes the hydrogen for building sugars, and discards the oxygen. For a cell, dioxygen is waste. For a planet with an atmosphere practically devoid of it, waste on that scale is a geochemical event. The Great Oxidation Event, also called the Great Oxygenation Event, the Oxygen Catastrophe, the Oxygen Revolution, the Oxygen Crisis, and the Oxygen Holocaust, was the interval during the Paleoproterozoic era when the atmosphere and the shallow seas first experienced a rise in free oxygen.

It began approximately two forty-six to two forty-two six billion years ago during the Siderian period and ended around two point o six billion years ago in the Rhyacian. Before it, oxygen in its O two form stood at about zero point zero zero one percent of recent atmospheric levels. By the end of the event, levels had reached as high as ten percent of the modern value. The modern atmosphere is nearly twenty-one percent oxygen, which makes it an oxidizing atmosphere and makes almost every rock at the surface a chemical target.

The change from weakly reducing to oxidizing is the single largest alteration life has ever made to this planet. Working out that it happened at all was the achievement of Preston Cloud in the 1970s. He noticed that detrital sediments older than about two billion years contain grains of pyrite, uraninite, and siderite. These are minerals holding reduced forms of iron or uranium, and they are simply absent from younger sediments because in an oxidizing atmosphere, they oxidize rapidly and cannot survive being tumbled down a river.

Their presence in old fluvial and deltaic sands is widely interpreted as direct evidence of air with no oxygen in it. Cloud also observed that continental red beds, red-colored sandstones coated with hematite, an oxidized ferric mineral, begin to appear in the record at about the same period. One set of minerals disappears. Another set requiring the opposite conditions arrives. Heinrich Holland elaborated these ideas through the 1980s and placed the main interval of oxygenation between two point two and one point nine billion years ago.

Fossil soils tell the same story. Paleosols older than two point four billion years have low iron concentrations consistent with weathering under an anoxic sky, where iron dissolves and washes away rather than rusting in place and staying put. The largest signal is banded iron formation, and its ending is what matters. These are sedimentary rocks of alternating iron-rich and silica-rich layers, and they can only form when abundant dissolved ferrous iron is carried into a depositional basin.

An oxygenated ocean blocks that transport because oxygen converts the iron into insoluble ferric compounds that drop out long before they reach the basin. Banded iron formation peaks around two point five billion years ago and largely vanishes from the record at one point eight five billion years ago. And that disappearance is read as the moment the deep ocean itself became oxygenated. For the biosphere already in place, the gas was poison. Free oxygen is highly reactive and readily oxidizes organic compounds, genetic material most damagingly of all, and the biosphere of the time was largely anaerobic.

It had no defense against it. Oxygen also had an enormous chemical debt to work through first. Ferrous iron, sulfur, hydrogen sulfide, and atmospheric methane all consumed it, and depleting that surface reducing capacity took nearly a billion years. Only after those sinks were exhausted could the gas accumulate freely. What it then did to living communities was severe. The oxidative change, compounded by a global glaciation, devastated the microbial mats that had blanketed the Earth's surface.

Among the casualties may have been archaeal colonies that used retinal, a pigment harvesting green spectrum light, to power a form of anoxygenic photosynthesis. That possibility is the basis of the Purple Earth hypothesis, which imagines a shallow world tinted by pigment rather than by chlorophyll. Isotope geochemistry from sulfate minerals has been interpreted to indicate a decrease in the size of the biosphere of more than eighty percent associated with changes in nutrient supply at the end of the event.

It is inferred to have constituted a mass extinction, though it is typically left off the conventional lists, which are implicitly limited to the Phanerozoic and to organisms large enough to count. Out of that came the arrangement that every complex organism now depends on. Surviving archaea took in aerobic proteobacteria, which stayed, becoming endosymbionts and eventually mitochondria, the compartments that use oxygen to extract energy from food. That symbiogenesis may have opened the way to eukaryotic organisms and in time to multicellular life.

The gas that poisoned the old biosphere became the power supply of the new one. The timing remains genuinely disputed. There is widespread consensus that initial oxygenation occurred during the first half of the Paleoproterozoic and no agreement on when within it. Scientific publications between twenty sixteen and twenty twenty-two differ by approximately five hundred million years, with estimates including two point seven, two point five zero one to two point four three four, two point five zero one to two point two two five, two point four six zero to two point four two six, two point four three zero, two point three three, and two point three billion years ago.

Depositional ages for many ancient units are uncertain. Different geochemical proxies are open to different interpretations. Paleontologists have discussed and quantified the effects of an incomplete record for decades under the name of the Signor-Lipps effect. But that correction is rarely applied to geochemical records, which may be adding uncertainty nobody has measured. Against all of that, the finished planet is the standard by which the whole sequence is judged. It is an ellipsoid roughly forty thousand kilometers around, which is twenty-four thousand nine hundred miles, the densest planet in the Solar System and the largest and most massive of the four rocky ones.

It sits about eight light- minutes from its star, one astronomical unit, completing a revolution in a year of about three hundred and sixty-five point two five days and turning on its axis in slightly less than a day, about twenty-three hours and fifty-six minutes. The axis is tilted relative to the perpendicular of the orbital plane, which is why the seasons turn. Ocean covers seventy point eight percent of the crust, and the remaining twenty-nine point two percent is land, most of it gathered in one hemisphere, most of it humid and covered by vegetation.

The polar ice sheets hold more water than all the groundwater, lakes, rivers, and atmospheric moisture combined. The greenhouse effect of water vapor and carbon dioxide maintains an average surface temperature of fourteen point seven six degrees Celsius, which is fifty-eight point five seven degrees Fahrenheit, and at that temperature, water is liquid under ordinary pressure. The equator receives more sunlight than the poles, and that difference alone drives the winds and the ocean currents, the climate belts, the rainfall, and the cycling of carbon and nitrogen.

Against that quietly balanced arrangement, life took its time. It remained small and microscopic until about five hundred eighty million years ago, when complex multicellular forms arose. The diversification culminated in the Cambrian Explosion about five hundred and thirty- eight point eight million years ago, which produced most of the major phyla known today and divides the Proterozoic from the Paleozoic. The Phanerozoic that followed splits into three eras. The Paleozoic of arthropods, fishes, and the first life on land.

The Mesozoic, spanning the rise, reign, and extinction of the non-avian dinosaurs. And the Cenozoic, in which mammals rose. Humans emerged three hundred thousand years ago in Africa. Around ninety-nine percent of all species that ever lived, over five billion of them, are already gone, and of the many millions of species estimated to share the planet now, only a small fraction have ever been formally named and described. Almost none of the story just told was watched. It was inferred from isotope ratios, from the minerals that are missing where they ought to be, from the ages of melted rock, from computer models run and rerun against samples that fit in a hand.

An entire eon of Earth's history rests substantially on granular crystals from one dry locality in Western Australia, small enough to be lost on a fingertip. The planet kept almost nothing from its own beginning, and what it kept, it kept by accident. Nearly all of the physical evidence behind this account would fit on a single laboratory bench. A few meteorites holding pale inclusions of calcium and aluminium, the oldest solids anywhere in the system at four thousand five hundred and sixty-eight million years.

A sachet of zircon grains from the Jack Hills, each one a speck. Several boxes of gray lunar impact melt returned by the Apollo missions. Slabs of banded iron, red and silver in alternating bands, cut and polished. A dark smear of graphite in metasedimentary rock from southwestern Greenland. Laid out together, the whole primary record of the planet's origin weighs less than a suitcase. That disproportion is the strangest feature of the subject. A claim about the assembly of an entire world rests on objects that can be lifted with one hand, and each object answers only one narrow question.

The meteorite inclusions fix the start of the clock. The zircons report on a crust and a temperature. The lunar melts date collisions. The banded iron and the old fossil soils track a gas. The Greenland graphite reports a proportion of carbon isotopes and nothing else at all. None of them was made to be read. They are accidental survivors, and the account is built from the places where their separate testimonies happen to overlap. There is no continuous record anywhere. The rock that formed the earliest crust is gone, recycled through a mantle that has been turning over for four and a half billion years.

What remains from the first eon is scattered, out of sequence, and mostly held inside younger material that gathered it up by chance. Geologists have learned to work with that, which is why the dates in this field so often arrive with a plus and a minus attached, and why so many of the conclusions are stated as ranges rather than as figures. Time in geochronology is measured in Ma, millions of years before the present, and the international convention that fixes the boundaries between eons is a working agreement, revised when the samples require it.

So the Hadean can be set back where it belongs, at the far end of a scale that no single mind can hold. A planet not long finished, still shedding the heat of its own construction. A sun that had not yet reached its present brightness, shining more dimly on a sea that stood warm regardless. Weather of a kind, rain and evaporation and cloud on a world with almost no shore for it to fall against. No sound anywhere except water, wind, thunder, and the long percussion of falling stone. The rest of the story took four billion years and left the planet still working on it.

The working has never stopped, and that is the part most easily missed when the early chapters end. Earth's crust consists of slowly moving tectonic plates, and the interactions between them are still producing mountain ranges, volcanoes, and earthquakes. Every range now standing is temporary. The Himalaya is a collision in progress, and the ocean floors are conveyor belts feeding older crust back down into the interior at their far edges. Plate tectonics continues to shape the continents and the oceans and the life they harbor, which means the geography that seems permanent within a human lifetime is a snapshot of an arrangement that has never once held still.

The atmosphere is doing the same steady work overhead. It sustains the surface conditions that keep water liquid, and it also shields the ground, burning up most meteoroids at entry and absorbing most of the ultraviolet light that arrives with the sunlight. Nitrogen and oxygen, with water vapor widely distributed through it and cloud covering most of the planet at any moment. The composition is a biological product as much as a geological one, and the protection it offers is a side effect of a waste gas released by microbes that never had a use for it.

Life continues to move as well. Species take on new forms, split into daughter species, or fail in the face of physical environments that will not stay the same, and the sorting has been running without interruption since the first cells. Estimates of the number of species alive now range from ten million to fourteen million. About one point two million have been formally documented, which leaves over eighty- six percent undescribed, unnamed, and unexamined, most of them small and most of them in places nobody has surveyed.

The catalog of the present is less complete than the calendar of the past. Into that arrangement, very late, came a species that could read the rocks. Humans depend entirely on the biosphere and on the planet's natural resources, and their effect on both has grown steadily heavier. Humanity's current impact on Earth's climate and biosphere is described as unsustainable, threatening the livelihoods of humans and of many other forms of life, and driving widespread extinctions. That sentence sits oddly beside the rest because it is the only part of the account in which the outcome is not yet settled.

Every other event described here has already happened and can only be measured. This one is still being decided by the participants. Even the name is a piece of evidence about how recently the whole picture came together. The modern English word Earth descends through Middle English from an Old English noun most often spelled eorthe. Cognates appear in every Germanic language, and from that spread of forms, linguists have reconstructed an ancestor written as ertho. In its earliest attested uses, eorthe was pressed into service to translate all the senses carried by the Latin terra and the Greek ge at once.

The ground underfoot, the soil itself, dry land as opposed to sea, the human world, the whole surface including the sea, and finally, the globe. That last meaning was not the original one. The word began as a word for dirt. For most of its history, it named the stuff a farmer turned over and only gradually stretched to cover the entire object on which farming happens. As with the Roman terra, also called tellus, and the Greek Gaia, Earth may have been a personified goddess in Germanic paganism, and late Norse mythology included Jörð, whose name simply means Earth and who is often given as the mother of Thor.

The planet was named after the ground long before anyone understood that the ground was a planet. Writing followed the same slow path. Historically, the word appeared in lowercase. During the early Middle English period, the definite sense, meaning the globe rather than the soil, began to be expressed as the Earth. By early modern English, the capitalization of nouns had spread, and the Earth was also written The Earth, particularly when it appeared alongside other heavenly bodies. More recently, the name is sometimes given simply as Earth by analogy with Mars and Venus and the rest.

Though Earth and The Earth remain perfectly common. House styles still differ. One convention capitalizes Earth as a proper name, as in Earth's atmosphere, but drops to lowercase after the definite article, as in the atmosphere of the Earth. In colloquial expressions about what on earth someone is doing, it is almost always lowercase because in those, the old meaning is still faintly present. Other names survive in the technical and poetic registers. Terra is used occasionally in scientific writing and in science fiction, and it is the ordinary name of the planet in Italian and Portuguese, altered to Tierra in Spanish and Terre in French.

Tellus turns up in poetry as the personified form. The Latinate Gaia from the Greek Gaia is rare. Though the spelling Gaia has become widespread because of the Gaia hypothesis and is now usually pronounced with a hard opening syllable rather than in the older way. The adjectives fan out from the same three roots. Earthly comes from the English word. Terran, terrestrial, and by way of French, terrain, come from terra. Tellurian and telluric come from tellus. A single object and a small crowd of words for it, most of them older than the knowledge that it moves.

The solar system around it has not finished changing either. Collisions between bodies have continued without interruption up to the present day and have been central to the system's evolution. Though the objects involved now are small and the events are rare. Beyond Neptune, many sub-planet-sized bodies formed and remain, and several thousand trans-Neptunian objects have been observed so far. Unlike the planets, most of them travel on eccentric orbits inclined to the plane in which the planets move, which is a preserved record of disturbance rather than a tidy leftover.

The positions of the planets themselves might have shifted through gravitational interaction, and planetary migration accounts for parts of the system's present structure. Further out in time, the arrangement ends. In roughly five billion years, the fusion of hydrogen into helium will stop in the Sun's core, and its surface will begin to cool as it becomes a subgiant. Over the following two and a half to three billion years, it will expand outward to many times its current diameter, becoming a red giant not once but twice, with an intermediate phase between the principal red giant branch and the asymptotic giant branch.

Then it will cast off its outer layers as a planetary nebula, a slowly widening shell of gas, and leave behind a stellar remnant known as a white dwarf, the exposed core, small and dense and cooling. What happens to the planets after that is decided from outside. In the distant future, the gravity of stars passing nearby will gradually reduce the Sun's retinue. Some worlds will be destroyed. Others will be flung out of the system entirely into interstellar space on trajectories that will not return.

Over the course of tens of billions of years, it is likely that the Sun will be left with none of the original bodies still in orbit around it. The same accretion that gathered a disk into planets runs in reverse at the end, patiently, one encounter at a time, until the collection is dispersed. Which returns the question this account began with to something like an answer. A fragment of a cold molecular cloud became a rocky world with a large moon, a global ocean, and oxygen in its air through a chain of ordinary physical processes, each of which is observed elsewhere in the universe, and none of which required anything unrepeatable.

Gravity concentrated matter. Rotation flattened it. Collisions built it up, and one collision very nearly took it apart. Density sorted its interior. Heat escaped. Water condensed. Chemistry became complicated enough to copy itself, and one branch of that chemistry rebuilt the atmosphere. The unusual thing about Earth is not the mechanism. It is that the entire sequence ran to completion in one place, and that something on the surface eventually became capable of reconstructing it from grains of sand.

The reconstruction is unfinished, and the gaps in it are genuine. The assembly time is disputed by a factor of thirty. The geometry of the impact that made the moon is unsettled. Whether the Hadean surface broke into plates is argued. The bombardment record may be one catastrophe misread as an epoch. The origin of life has never been observed. The date of the first free oxygen shifts by five hundred million years, depending on which proxy is trusted. These are not failures of the account.

They are the places where the samples run out, and every one of them is a question that further rock, further modeling, or further sampling of the moon may eventually close. Meanwhile, the planet keeps its own record, indifferent to whether anyone reads it. The inner core adds another shell of frozen metal. A plate slides a few centimeters beneath another and lifts a ridge nobody will see finished. Rain falls on stone and takes a little of it into the sea, where it will settle, harden, and in some far future be lifted, cooked, or carried down into the mantle.

Somewhere in an old conglomerate, a zircon that has already outlasted the rock it grew in waits for weathering to release it again. The whole of it is far too long to hold at once, so it helps to shrink it. Set the age of the Earth against a single candle burning down through one evening, and the Hadean is the moment of lighting, the flare, and the settling of the wick. The first ocean and the first cells follow almost immediately. Oxygen arrives around the middle. Everything with a skeleton, everything with a leaf, every forest and every ocean of fish belongs to the last portion of the wax.

Recognizable humans appear in the final thread of smoke above the flame, and the part of that thread in which anyone knew what a zircon was is thinner still. There is no urgency in any of this. The planet took its first billion years to make an ocean and a living thing, and it has spent the three and a half billion since on everything else without hurry and without a plan. Deep time asks nothing of anyone. It only continues at a rate of about one hundred degrees of cooling per billion years in the dark beneath everything.

So the young Earth can be left where it lies now, molten and battered and still four billion years from the present, with the record of it folded into a few grains of crystal in a dry range of hills in Western Australia. The evidence is safe in its drawers. The moon is where the collision put it, drifting outward by about three point eight centimeters this year, as it has every year since. The work will still be there in the morning, and the rocks are extremely patient. Let the room go quiet.

Let the long, slow sequence of cloud and disc and dust and stone settle back into the deep time it came from, where it will keep without attention. There is nothing left to do here except to let the last of the light go soft and blow that candle out. Rest well, sleep deeply, and let the ancient Earth fade softly into dreams.