Insight Editions · No. 27
The Planet in a Pebble
A Journey into Earth's Deep History
Jan Zalasiewicz
Jan Zalasiewicz traces the geological history of a Welsh slate pebble, linking its minerals and fossils to Earth's formation, tectonic change, burial, erosion, and human encounters.
A study guide to The Planet in a Pebble by Jan Zalasiewicz. Shared free for personal reading; not for sale and not endorsed by the authors or publishers. Please read and buy the original.
Prologue: A Capsule of Stories
A grey Welsh slate pebble is a capsule of stories held by evidence, not invention. Thirteen paths through one object echo the Geological Society's founding on Friday 13 November 1807. The method is to enlarge the pebble with instruments until vanished worlds come into focus, then to keep only what the measurements allow.
Zalasiewicz starts on the beach between Aberystwyth, Clarach and Borth on the coast of west Wales, where rivers such as the Ystwyth, Rheidol and Claerwen bring down the local rock. Almost every pebble is the same: grey Welsh slate, flattened into a disc, with grey and blue-grey stripes, crazed white lines (veins) and tiny gold and red flecks. Pick one at random, he says, and it will do.
His claim is that a pebble is a capsule of stories, and that geology is a way of building stories that are constrained by evidence. One can, through "a combination of narrative and technical wizardry," enlarge a pebble to the size of a mountain and recreate the distant past. The book is organised as thirteen journeys through the same object, a nod to the Geological Society of London, founded at the Freemasons' Tavern in Great Queen Street with 13 original members.
The human mind loves invented stories of elves and vanished empires. The book's wager is different: stories that cannot contain anything shown to be untrue at the time of telling. They are not guaranteed true, but they are the most reasonable interpretations the evidence will bear, better or worse in focus depending on what can be measured. That discipline, not the romance of deep time alone, is what he wants the reader to take away.
The visible features already sketch a timeline. The stripes were laid down first, on a Silurian sea floor; the gold flecks grew during shallow burial; the cleavage that shapes the disc formed in mountain-building; and the white veins cut across everything, so they came last. Reading those relationships before reaching for any instrument is the first lesson of the book.
:::key Evidence-constrained storytelling A pebble becomes an archive only when every claim stays revisable under new measurement. The prologue's rule is the method of the whole book. :::
What "thirteen journeys" asks of the reader
Each journey re-enters the same 50-gram disc with a different instrument and a different question. The beach between Aberystwyth, Clarach and Borth is not scenery; it is the sample tray. Rivers (Ystwyth, Rheidol, Claerwen) deliver the cliff's product; waves finish the rounding. Grey and blue-grey stripes, white veins and gold flecks are already a relative-age diagram before any laboratory opens.
The Geological Society's Friday-the-13th founding is more than a joke: it licenses a structure in which thirteen partial views are allowed to disagree until the clocks converge. The prologue's discipline, stories that cannot contain what is known to be untrue at the time of telling, is the standard against which every later chapter should be judged.
Read any object as five layers: Composition (what it is made of), Structure (how it was assembled), Inclusions (what got trapped), Alteration (what happened later), Clocks (when each event occurred). The stories must agree across layers.
Take a worn coin, a library book, or a scratched phone case. List one observation under each of C-S-I-A-C. Which layer gives you the strongest date? Which is most ambiguous? That is the geologist's daily trade-off.
A fluent narrative about a pebble (or a product, or a career) can feel complete while resting on one untested analogy. The prologue's safeguard is that every claim must stay revisable when new measurements arrive. If a story cannot say what would falsify it, it has left geology and entered myth.
"These stories cannot contain anything that can be shown to be untrue, at the time of telling." - Jan Zalasiewicz
| Item | Detail |
|---|---|
| Founding date | Friday 13 November 1807 |
| Place | Freemasons' Tavern, Great Queen Street, London |
| Original members | 13 |
| Journeys in this book | 13 chapters after the prologue |
1. Stardust
The pebble holds about 10^24 atoms, half of them oxygen. This chapter inventories those atoms, traces them from the Big Bang through stellar forges and supernovae, and gathers them into the young Earth just as Theia is about to hit.
An inventory of atoms
How many atoms are in the pebble? Zalasiewicz borrows from Amedeo Avogadro (Count of Quaregna and Cerreto, a man with revolutionary leanings), whose constant says a mole of any element holds about 6 × 10^23 atoms. The pebble weighs about 50 grams; about half is oxygen, and much of the rest is silicon (atomic weight 28) and aluminium (27). With an average atomic weight of about 25, that gives roughly two moles, or about 10^24 atoms: a "million billion billion." If each atom were a wrapped sweet, the sack would be about the size of the Moon.
To find out which atoms, you sacrifice a neighbouring pebble ("slate is slate is slate") and use X-ray fluorescence (XRF): X-rays knock electrons out of atoms, the atoms re-emit photons with energies that fingerprint each element, and the result is good to a couple of per cent.
| Level | Elements |
|---|---|
| About half by mass | oxygen, locked in silicates (a greater share than in the atmosphere) |
| Major | silicon, aluminium |
| A few per cent | iron, magnesium, potassium, sodium, titanium |
| Under 1 per cent | calcium, manganese, phosphorus; water and carbon (loss on ignition) |
| Parts per million | vanadium, chromium, copper, zinc, lead, barium, rubidium, strontium, yttrium, cerium, lanthanum, niobium, arsenic |
| Parts per billion | gold, silver, platinum; iridium at about 1 ppb, still about 10^15 atoms |
Some atoms are radioactive: all 16 isotopes of uranium (92 protons), plus isotopes of thorium, potassium, samarium and rubidium. These are the clocks the rest of the book depends on. Short-lived elements such as promethium (half-life under 17 years) wink in and out, and the pebble now also carries a sprinkling of plutonium, made by the kilogram in reactors and spread by bomb tests.
The book mentions copernicium "if the name is ratified." IUPAC approved the name and symbol Cn for element 112 on 19 February 2010. The seventh row of the periodic table was completed in 2016, when the names nihonium (113), moscovium (115), tennessine (117) and oganesson (118) were approved. None of these superheavy atoms will be in the pebble.
From the Big Bang to the first stars
Almost all the Universe is hydrogen and helium; the elements of a rocky planet (oxygen, silicon, aluminium, iron, magnesium) are about a thousandth part. Zalasiewicz dates the Big Bang to about 13.7 billion years ago. In the first second, protons and neutrons formed; within about three minutes, about a quarter of the matter had become helium, with a trace of lithium. Matter and antimatter mostly annihilated; no antimatter galaxies have been found.
After about 250,000 years the Universe, now at about 3000°C, became transparent; that light survives as the cosmic microwave background, cooled to about 3 K. The hydrogen in the pebble's water is its oldest component. Gravity pulled gas into clumps; when a core reached about 100 million degrees, fusion began, turning four protons into a helium nucleus and converting the lost mass to energy by E = mc².
The final Planck 2018 analysis of the cosmic microwave background gives an age of 13.797 ± 0.023 billion years, usually rounded to 13.8 billion, slightly older than the book's 13.7. The same data give a Hubble constant of 67.4 km/s/Mpc, lower than distance-ladder measurements: the unresolved "Hubble tension."
Forges and explosions
Our Sun is "an ordinary slow burner": it has burned for a little over 4.5 billion years and has about 5 billion to go. Stars ten times the Sun's mass or more forge the heavier elements in stages, each needing more heat, until the chain stops at iron-56, beyond which fusion absorbs energy instead of releasing it.
Stellar nucleosynthesis as Zalasiewicz describes it. Each stage needs a hotter core; iron is the end of the line, and the heaviest atoms in the pebble come from the neutron blizzard of a supernova. The Crab Nebula is his example of a supernova: the star exploded about 7300 years ago, and its light arrived about 6300 years later, in 1054, when it was the brightest object in the sky after the Sun for some three weeks, recorded in Arabia, China and Japan, perhaps in North America and Irish monasteries. In the explosion, a blizzard of neutrons builds elements all the way to uranium. Dying stars also shed stardust: olivine, pyroxene, "graphite whiskers" and microdiamonds. The Stardust mission caught a few interstellar specks in a gel-coated collector, and meteorites contain presolar grains with wild isotope ratios.
The atoms' earlier paths are, in detail, unknowable: many probably passed through earlier solar systems, were swallowed by growing stars and expelled again. Telescopes, the Hubble above all, show how it began, catching light that has travelled over 12 billion years. The first stars, formed from hotter primordial gas, could only be giants a few hundred times the size of the Sun; they "lived fast, and died young," and their deaths started the factory of elements. The most distant quasar then known lit up dust containing carbon, oxygen and iron less than a billion years after the Big Bang.
The James Webb Space Telescope, launched in December 2021, has pushed the frontier far beyond Hubble. In 2025 the ALMA array detected oxygen in JADES-GS-z14-0, a JWST-discovered galaxy seen about 300 million years after the Big Bang (redshift 14.18), one of the most distant galaxies confirmed. Heavy elements were being made far earlier and faster than expected, which strengthens the book's picture of a rapid early element factory.
Gathering the Earth
A nearby supernova probably seeded the Sun's parent cloud with short-lived radioactive aluminium and may have triggered its collapse. In the spinning disc, chondrules formed: melt droplets flash-heated to about 1500°C, "fossil caviar," that lost water, potassium, magnesium and iron. Dust clumped into planetesimals tens of kilometres across within a few million years. A little over 4.5 billion years ago, the pebble's atoms were spread through a ball of rock about 10,000 km across, a few per cent smaller than today's Earth.
Then came the iron catastrophe: metal droplets sank thousands of kilometres to form the nickel-iron core, and the heat released created a magma ocean. Most of the Earth's nickel, iron, iridium and gold went with them; the pebble's few atoms of these are survivors. Five to twenty million years after accretion, the chapter ends: "Theia is coming."
Zalasiewicz is careful to mark uncertainty ("probably," "perhaps"), but the chapter's sequence (supernova trigger, chondrule heating by shock waves) compresses active debates into a smooth story. The trigger hypothesis in particular remains contested: short-lived isotopes could also come from stellar winds. Read the chapter as a best current model, not a settled chronology.
Why the inventory matters
The inventory is not a shopping list for its own sake. Once you know the pebble holds about 10^24 atoms, half oxygen, with uranium clocks and even 10^15 atoms of iridium at one part per billion, every later chapter has something to measure. William Blake saw a world in a grain of sand; Zalasiewicz notes that Blake was only about an order of magnitude out.
XRF on a neighbour pebble assumes "slate is slate is slate." That local assumption is fair on one beach from one cliff, and it is the first of many sampling choices the book will make. Mass spectrometers then sift the parts-per-billion rarities: gold, silver, platinum, iridium. In an almost infinite jungle, even the rarest orchid is abundant. Promethium, with a half-life under 17 years, winks in and out by the thousands as uranium decays. Artificial transuranics, made by the kilogram as plutonium, now cling to every beach pebble on Earth.
Nucleosynthesis in fuller detail
Hydrogen fusion at about 100 million °C makes helium and converts lost mass by E = mc², where c² is 9 × 10^16. Helium burning near 200 million °C makes carbon; carbon at about 800 million °C yields oxygen, neon, sodium and magnesium; silicon burning near 3000 million °C pushes toward iron. The Sun cannot do this heavy lifting: it is an ordinary slow burner with about 5 billion years of hydrogen left. Only giants ten times the solar mass race through the chain until fusion dies at iron-56.
The Crab Nebula records a supernova whose light arrived in 1054. A neutron blizzard jerry-builds atoms up to uranium; unstable ones fall apart on timescales from fractions of a second to aeons. Stardust minerals (olivine, pyroxene, graphite whiskers, microdiamonds) are seen in nebulae, caught by the Stardust mission's gel collector, and found as presolar grains in meteorites with isotope ratios alien to this Solar System. First stars were giants hundreds of times the Sun; quasar dust already held carbon, oxygen and iron within a billion years of the Big Bang.
From cloud to core
A nearby supernova likely seeded the solar cloud with short-lived aluminium-26 and may have triggered collapse. The Oort Cloud reaches about a light-year out. In the disc, chondrules flash-heated to about 1500°C ("fossil caviar"; heat perhaps from shock waves) built planetesimals tens of kilometres across in a few million years. About 4.5 billion years ago the young Earth was a ball about 10,000 km across, a few per cent smaller than today. The iron catastrophe sank nickel-iron (and most iridium and gold) to the core, melting a magma ocean and leaving a mantle about 3000 km thick. Five to twenty million years later: "Theia is coming."
Avogadro's sack and the cosmic otherness of rock
Amedeo Avogadro, Count of Quaregna and Cerreto in Piedmont, glimpsed the link between particles, volume and mass; his teaching paused briefly for revolutionary leanings when a king lived nearby. A mole of oxygen is 16 grams; the kitchen scales put the pebble at about 50 grams; an averaged atomic weight near 25 yields on the order of 10^24 atoms. That is the enormity of the submicroscopic world. The pebble is a microcosm of the Earth's upper crust and cosmically unrepresentative: the Universe is mostly hydrogen and helium, while rocky planets concentrate oxygen, silicon, iron and their kin (about a thousandth of cosmic matter).
The sacrificial neighbour is crushed, pressed into a pellet or melted into a rock-glass bead. High-energy X-rays knock electrons; photons fingerprint elements to within a couple of percent. Loss on ignition drives off water and a soupçon of carbon. Among the ppm rarities sit familiar metals and a few sinister ones such as arsenic. Beyond uranium lie short-lived artificial elements; copernicium is now ratified, and the seventh row of the table is complete to 118, yet none of those superheavy atoms live in the pebble.
Antimatter galaxies would announce themselves with telltale flashes where matter meets antimatter; none have been found. Either matter won in the first minutes, or anti-pebbles lie beyond any telescope. After three minutes the ingredients were a plasma of nuclei and electrons in a Universe opaque as pea soup. A quarter of a million years later electrons settled into orbits, photons streamed free, and the glow that astronomers still measure as the cosmic microwave background began its 13.8-billion-year stretch into microwaves at about 3 K.
Had the Big Bang been perfectly smooth, history would have been an ever-thinner dispersal of lone atoms. Fluctuations, still visible as regional variations in the microwave background, let gravity gather the first stars. Small stars like the Sun cannot forge past helium in quantity; giants race through carbon, neon, oxygen and silicon until iron kills the furnace. A Sun-as-supernova thought experiment (impossible for our star) would make the sky briefly 10 billion times brighter and sand-blast the Earth in days. The productive truth is narrower: only that maelstrom forges the pebble's copper, zinc, gold and uranium.
Presolar grains and the iron catastrophe's leftovers
Presolar grains in meteorites carry isotope ratios wild enough to mark them as aliens to this Solar System. Graphite whiskers and microdiamonds give spectral signals in dusty nebulae; the Stardust mission's gel collector caught a few interstellar specks. Chondrules flash-heated to about 1500°C lost water and volatiles and became fossil caviar. Planetesimals tens of kilometres across grew in a few million years. When the iron catastrophe drained nickel, iron, iridium and gold into the core, the pebble's remaining atoms of those metals became survivors of a planetary sorting event. The mantle, about 3000 km thick, inherited the rest. Five to twenty million years after accretion, Theia was already on a collision course.
Hydrogen (Big Bang), Helium burning onward in stars, stop at Fe-56, then SuperNova neutron blizzard for everything heavier.
Assume a wrapped sweet is a 1 cm cube. The Moon's volume is about 2 × 10^19 m³. How many sweets fill it? Compare with 10^24. The order-of-magnitude match is the point.
Counting atoms without asking where they came from and what reset the clock produces false confidence. Always name the sampling assumption (here: "slate is slate").
In AI, dataset cards play the role of XRF.
2. From the Depths of the Earth
Theia smashes into the Earth; the Moon forms from the splash. Mantle convection then stores the pebble's atoms for billions of years until an island-arc melt, dated by Sm-Nd, builds the seed of Avalonia.
Impact
Theia, a planet about the size of Mars, shared an orbit close to the Earth's. They met at something like 40,000 km/h; Theia lost its identity in a few minutes and the Earth was smashed "like a grapefruit hit hard with a hammer." Theia's core sank into the Earth's; some of both planets' outer layers splashed out into a cloud that became the Moon, which then looked more than twice as large in the sky as it does now.
| Evidence | Capture idea | Giant impact |
|---|---|---|
| Physics of the orbit | Capturing a body this size is implausible | Holding ejecta in orbit is plausible |
| Isotopes (e.g. oxygen) | Moon should differ, as Mars meteorites differ | Earth and Moon are near identical, mixed in the impact |
| Water and volatiles | No reason for a dry Moon | Volatiles boiled off: a remarkably dry Moon |
:::update Theia: where it came from, and where it may still be Two results since the book, plus a 2025 follow-up. In 2023 Qian Yuan and colleagues (Nature 623) proposed that the two continent-sized large low-velocity provinces at the base of the mantle, under Africa and the Pacific, are buried relics of Theia's mantle, 2.0 to 3.5 percent denser than the proto-Earth's and able to survive 4.5 billion years. A 2025 modelling paper shows asteroid-sized, rheologically strong Theia debris can persist as LLVP-like piles through Earth history. In 2025 Timo Hopp and colleagues (Science 390) used iron isotopes plus other elements to infer that Theia formed in the inner Solar System, possibly closer to the Sun than Earth. All remain hypotheses under test. :::
The engine below
After the magma ocean solidified, the mantle began its long convection. Early on, currents ran perhaps twice as fast, because the deep Earth was hotter: its natural radioactivity was higher and heat from the Theia impact lingered. The evidence is komatiite, a lava denser and richer in iron and magnesium than any erupted today, which betrays higher melting temperatures. As the surface cooled, rain began, and water from volcanic degassing and from ice-rich comets collected into the first oceans. Plate tectonics, as far as the oldest crustal remnants show, was already working in its basic pattern, something no other planet in the Solar System is known to do. Today mantle rock flows, solid, like glacier ice, at a few centimetres a year, in cycles of hundreds of millions of years. Mantle plumes are faster routes: fountains several hundred kilometres high and one or two hundred across. Iceland's plume, active for about 60 million years, lifts the island about 3 km; Hawaii's builds Mauna Loa, the largest volcano on Earth.
How do we know what is down there, at 1000 to 4000°C? The diamond anvil cell squeezes tiny grains between diamond tips while X-rays reveal their structure. As pressure rises, atoms collapse into more compact frameworks, often with a palpable "pop." Minerals such as ringwoodite, perovskite, post-perovskite and ferropericlase appear at depths of a thousand kilometres and more; the pebble's atoms spent billions of years bound in such structures, slowly carried by currents that, like glacier ice, flow while remaining solid. The core-mantle boundary, at about 4000°C, blocks some earthquake waves because the outer core is liquid iron; the mantle above acts like a thermos flask.
A clock that remembers the mantle
Between about 1.5 and 1 billion years ago, the pebble's atoms gathered in a wedge of mantle hundreds of kilometres long and tens thick, the smallest volume they had ever occupied. The clock is samarium-neodymium: 147Sm decays to 143Nd by losing an alpha particle, with a half-life of 106 billion years. Melting puts more neodymium than samarium into the magma, so the ratio records when material left the mantle. This neodymium model age survives almost everything that happens later, so it can be read even in a slate.
| Clock | Parent to daughter | Half-life | What it dates |
|---|---|---|---|
| Sm-Nd model age | 147Sm to 143Nd | 106 billion years | When crust separated from the mantle |
| U-Pb (Chapter 3) | 238U to 206Pb; 235U to 207Pb | 4.46 billion; 704 million years | When a zircon crystallised |
A new continent
About a billion years ago most land formed the supercontinent Rodinia (Africa, Canada, Greenland, Siberia, Australia), ringed by an ocean sometimes called Mirovia. A time traveller would have found oxygen about as thin as on the summit of Everest and carbon dioxide several to many times higher: like working all day in a brewery, out of breath, with blood turning more acid. Life was microbial, in mats; a modern gooseberry-sized gromiid rolls along the sea floor leaving trails like billion-year-old "worm" traces from Australia and India.
Far south, the pebble's magma rose at subduction zones. The result was an island arc like the Marianas or Lesser Antilles, with explosive volcanoes, pyroclastic currents and tsunamis: the birth of Avalonia, the land of the Arthurian Avalon, of Shakespeare's country and of the Industrial Revolution, which today runs to Newfoundland but not Scotland. Where ocean crust sinks under a continent, as beneath the Andes (Cotopaxi, Popocatepetl and their neighbours), the rising magma is contaminated by much older continental crust; Avalonia's neodymium ages show pure, new crust, so the islands grew over ocean floor.
How Avalonia began: a wet slab sinks, releases water that melts the mantle above, and an island arc of new continental crust grows over ocean floor.
Theia and the Moon, step by step
Theia, roughly Mars-sized, struck at about 40,000 km/h. Zalasiewicz likens it to hitting a grapefruit with a hammer. Capture of a ready-made Moon is implausible; oxygen isotopes match Earth while Mars meteorites differ ("orange from apples"); the dry Moon fits a hot splash. After the impact the Moon looked more than twice as large in the sky, and a magma ocean thousands of kilometres deep reset the surface.
Mantle as glacier ice
Early mantle currents ran about twice as fast; komatiites record hotter melts. Mantle rock flows like glacier ice at a few centimetres a year on cycles of hundreds of millions of years. Plumes still rise: Iceland has ridden one for about 60 million years and stands about 3 km higher than neighbouring crust; Mauna Loa is the largest volcano. Laboratory diamond-anvil cells, with X-rays watching minerals "pop," map perovskite, post-perovskite, ringwoodite and ferropericlase. Whether the mantle is one shell or two remains debated. At the core-mantle boundary, near 4000°C, liquid iron blocks some earthquake waves; the mantle acts as a thermos flask.
Rodinia, Mirovia, Avalonia
About a billion years ago the supercontinent Rodinia faced the Mirovia ocean. Oxygen was thin, like the summit of Everest; carbon dioxide was many times today's, "brewery" thick. Microbial mats had been around for over 2 billion years; gooseberry-sized gromiids left rolling trails in Australia and India that look, misleadingly, like worm tracks.
Avalonia is the microcontinent that would one day host King Arthur stories, Shakespeare, and the Industrial Revolution, and that stretches to Newfoundland rather than Scotland. Continents float because their crust is light; ocean crust lasts only about 100 million years (at most 200). Avalonia began about 1 billion years ago as an island arc like the Marianas (Guam) or Lesser Antilles, with a trench about 100 km away, explosive volcanoes, pyroclastic density currents and tsunamis. Andean volcanoes such as Cotopaxi and Popocatépetl are contaminated by old continental crust; Avalonia's neodymium ages show pure new crust over ocean floor.
Why capture fails, and why wet slabs win
A ready-made Moon captured by Earth is dynamically implausible. The oxygen-isotope match between Earth and Moon, set against Mars meteorites that differ like oranges from apples, plus a dry lunar interior, favour a hot splash from Theia. After impact the Moon's disc looked more than twice as large; a magma ocean thousands of kilometres deep reset the crust.
Oceans came from degassing and from comets. Plate tectonics, as the book frames it, has no clear twin among the planets we know; the enabling conditions are water, size and cooling rate. Mantle temperatures of 1000-4000°C and a core-mantle boundary near 4000°C keep the thermos flask sealed by liquid iron that blocks some seismic waves. Partial melts rise because they are buoyant.
The Sm-Nd clock works even after the rock becomes slate: samarium-147 alpha-decays to neodymium-143 with a half-life of 106 billion years; melt takes more neodymium, so the model age dates escape from the mantle. The magma rose at about the latitude of Tierra del Fuego; longitude is unknowable. Gathering of the pebble's atoms in a mantle wedge about 1.5-1.0 billion years ago sets the stage for Avalonia's birth as an island arc: explosive, tsunami-prone, chemically juvenile.
Komatiites, plumes, and a thermos flask
Early mantle currents about twice as fast left komatiites as witnesses. Iceland has ridden a plume for about 60 million years and stands about 3 km higher than neighbouring crust; Mauna Loa remains the largest volcano. Diamond-anvil cells watch perovskite, post-perovskite, ringwoodite and ferropericlase "pop" under X-rays. Whether the mantle is one convective shell or two is still argued. Liquid iron at the core-mantle boundary blocks some earthquake waves and keeps the mantle as a thermos flask. Partial melts rise because they are buoyant relative to their source. Avalonia's juvenile Nd signature is the chemical receipt for an oceanic island-arc birth, not an Andean contamination story.
Impact (Theia), Flow (mantle like glacier ice), Clock (Sm-Nd model age), Arc (wet slab melts into an island arc).
Draw a trench, a descending ocean slab, the 50-100 km water-release zone, and an island-arc volcano. Label why Avalonia's Nd ages imply oceanic basement.
Treat uniqueness as a prompt to specify enabling conditions (water, size, cooling rate), not as a stop sign for comparison with Venus, Europa or Mars.
Yuan and colleagues (2023, Nature) proposed that the LLVPs under Africa and the Pacific are dense Theia mantle relics. A 2025 follow-up showed asteroid-sized, rheologically strong Theia debris can survive 4.5 billion years of convection. Hopp and colleagues (2025, Science) used iron isotopes to argue Theia formed in the inner Solar System. All three remain under test.
In AI, pretrained representations that survive fine-tuning are a shallow cousin of compositionally distinct deep stores.
3. Distant Lands
The pebble's grains are Avalonia in ruins. Thin sections, heavy minerals and zircon's twin uranium clocks turn a handful of sand into a provenance map, with concordance as the test of trust.
The present as a (partial) key
The grains record the dismantling of Avalonia, much as the layered ruins of Troy's seven cities record successive towns. Interpreting them rests on Charles Lyell's uniformitarianism, the idea that slow everyday processes, given enough time, explain the past. Lyell owed his sense of deep time to James Hutton, the 18th-century savant who, seeing tilted Scottish strata resting on the eroded roots of older mountains, wrote of "no vestige of a beginning, no prospect of an end." (Lyell reportedly danced when he heard Darwin's theory of coral atolls.) Zalasiewicz adds a caveat: the present is not always the key to the past, as the alien Silurian sea in Part III will show.
Looking closer
Under a hand lens, the pale stripes resolve into thousands of silt and sand grains. About a century and a half ago Henry Clifton Sorby of Sheffield invented the thin section: rock ground to about a thousandth of an inch on a glass slide, so light passes through. He was mocked for "studying mountains with a microscope." In polarised light, quartz shows as shades of grey, and its quirks reveal its origins.
| Feature in a grain | What it records |
|---|---|
| Wavy extinction under polarised light | Strain from mountain-building stress |
| Mosaic of tiny crystals | Deep shearing |
| Hair-like rutile (titanium oxide) inside | A fellow-traveller mineral that fingerprints the parent rock |
| Several rounded crystals cemented together | A recycled older sediment |
| Moulded around mica and feldspar | Late crystallisation in a granite |
Quartz is durable and comes from igneous, metamorphic and sedimentary rocks alike. It vibrates at a precise frequency under voltage, the basis of the quartz watch (about 32,000 vibrations a second). A thin section is "a room of explorers," each grain with a different past. One grain may have crystallised late in a granite, another been strained in a mountain chain, another sheared in the roots of a mountain belt, another recycled from a still older sandstone. Taken together they describe not one source but the whole eroding continent.
Heavy minerals and the zircon clock
The rarest grains say most. Geologists crush a pebble and drop the sand into a heavy liquid such as bromoform (CHBr3): quartz floats, and the dense minerals sink. Zircon, tourmaline, topaz, rutile, garnet, staurolite, monazite and apatite each record conditions: garnet needs about 500°C at 10 to 15 km; apatite forms in granites up to about 900°C. Scanning electron microscopes and microprobes then map them.
Zircon is the star. Its lattice accepts hafnium, yttrium, thorium and uranium ("atomic wallflowers" other minerals reject), and backscatter images show concentric zones like a tape recorder of the magma chamber. Above all, zircon excludes lead when it grows, so all lead inside is radiogenic.
Uranium has two decay chains: 238U (146 neutrons) to lead with a half-life of 4.46 billion years, and 235U (143 neutrons) to lead with a half-life of 704 million years. If both clocks give the same age, the grain is concordant and the age trustworthy, often to better than 1 per cent (about plus or minus a million years). If they disagree, the crystal has lost lead and the age must be treated with care. A built-in cross-check is the most valuable feature any measuring system can have.
A pebble holds a few dozen to a few hundred zircons. Avalonian sediments carry a characteristic signature of ages, which also turns up as a "mineral calling card" on fragments of Avalonia scattered round the North Atlantic.
The Avalonian zircon signature as the book describes it: many grains a little over 600 million years old, some over 700 million, a cluster near 1 billion and a few near 2 billion.
The same U-Pb method now underpins the oldest dates for Earth. A Jack Hills (Western Australia) zircon was dated to 4,404 ± 8 million years (Wilde and colleagues, Nature, 2001), and in 2014 John Valley's team used atom-probe tomography to show that its radiogenic lead sits in nanoscale clusters that cannot bias the age, confirming crust within about 160 million years of the Solar System's formation. Oxygen isotopes in these grains point to liquid water before 4.3 billion years ago (re-confirmed in 2023), and a 2024 Nature Geoscience study of very light oxygen in Jack Hills zircons argues that a freshwater hydrological cycle had begun by 4.0 billion years ago or earlier.
Ancient scenery, though, is lost: one cannot tell from grains whether Avalonia's volcanoes were perfect cones like Fuji or blown apart like Mount St Helens.
Mud and dissolved matter
The dark stripes were mud. Clay is now mica blades, colourful in polarised light. On Avalonia, rainwater made acid by the higher carbon dioxide of the time rotted rock into clay; land plants were only beginning (algae, fungi, liverwort-like forms), soils were thin with no humus, and dust blew as Saharan dust blows to Europe today. One gram of clay has hundreds of square metres of surface. Sand rolls along as ripples and dunes; clay travels in suspension.
Some matter travels dissolved, and its journey depends on its residence time in the sea.
| Ion | Average time in the sea | Consequence |
|---|---|---|
| Sodium | nearly 70 million years | The ocean holds about 10,000 million million tons and can hold more |
| Aluminium | about 200 years | Drops out almost at once |
| Magnesium | hitches a ride | Attaches to clay flakes, reaching the pebble that way |
Uniformitarianism and its limits
The pebble's grains are Avalonia dismantled over half a billion years, like Troy's seven cities ground to powder. Charles Lyell danced on hearing Darwin's coral-atoll theory; James Hutton saw "no vestige of a beginning, no prospect of an end" in a Scottish unconformity. The present is a key to the past, but not always: higher carbon dioxide, thinner soils and missing land plants change the chemistry of weathering.
Quartz under the microscope
Henry Clifton Sorby of Sheffield cut thin sections about 1/1000 inch thick and was ridiculed for "studying mountains with a microscope." Under polarised light quartz goes grey. A quartz watch crystal vibrates 32,768 times a second. Grains remember stress: wavy extinction from mountain squeeze, mosaic texture from deep shear, rutile needles, polycrystalline recycled sediment. Zalasiewicz likens a thin section to a room of explorers, each with a different passport stamp.
Heavies, clocks, residence times
Heavy minerals (zircon, tourmaline, topaz, rutile, garnet, staurolite, monazite, apatite) sink in bromoform. Garnet remembers about 500°C at 10-15 km; apatite can record granites to 900°C. Zircon hosts hafnium, yttrium, thorium and uranium ("atomic wallflowers"); backscatter images of hafnium zoning are a tape of the magma chamber. Becquerel found radioactivity in 1896. Uranium-238's half-life is 4.46 billion years (146 neutrons); uranium-235's is 704 million years (143 neutrons). Concordant ages agree within better than 1 percent (about ±1 million years). The pebble holds a few dozen to a few hundred zircons. The Avalonian age signature: many about 600+ Ma, some 700+, a gap, a cluster near 1 Ga, a few near 2 Ga from old Africa-like crust. Mountain shapes (Fuji versus St Helens) are lost; the ages remain.
Mud is now colourful mica blades under polarised light. Early land plants were algae, fungi and liverwort-like forms under acid rain from higher carbon dioxide. One gram of clay can cover hundreds of square metres. Sodium's ocean residence time is about 70 million years (oceans hold about 10^16 tons of sodium); aluminium's is about 200 years. Magnesium sticks to clay. Residence time, not rarity alone, organises which ions stay in the sea.
A room of explorers, and ions that linger
Under the hand lens the paler stripes resolve into thousands of silt and sand grains. Each quartz grain is an explorer with a passport: igneous, metamorphic or sedimentary origin; amethyst colour from irradiation; late granite quartz moulding around mica and feldspar; durability that outlasts almost everything else on the journey to the sea.
Heavy-liquid separation is still the workhorse: crush, float quartz on bromoform (CHBr₃), sink the denser suite, then SEM and microprobe. Zircon's twin clocks are the prize because zircon excludes lead while it grows, so both uranium-lead systems start clean. Discordance is not always failure: it can record later lead loss. Concordance within about 1 percent is the trust threshold.
Dissolved ions sort by residence time. Sodium lingers for about 70 million years in an ocean that holds about 10^16 tons of it; aluminium clears in about 200 years. Magnesium attaches to clay. Early land plants (algae, fungi, liverwort-like forms) and higher carbon dioxide meant acid rain and thin soils without humus, so dust travelled like today's Saharan plumes to Europe. Sand rolls as ripples and dunes; clay stays in suspension until the deep rendezvous of Chapter 4.
Avalonian calling cards around the North Atlantic
The Avalonian zircon age signature (many about 600+ Ma, some 700+, a gap, a cluster near 1 Ga, a few near 2 Ga) is a calling card scattered around the North Atlantic. Mountain shapes are lost; ages remain. Sorby's thin sections, once mocked, still turn each quartz into a witness. Concordance within about 1 percent (±1 million years at these ages) is the trust threshold. Residence time, not rarity alone, organises which ions stay in the sea: sodium for about 70 million years, aluminium for about 200 years. That distinction will matter again when Chapter 10 shows that "immobile" rare earths can move after all.
Float quartz on bromoform, Sink the heavies, read zircon Zones like a tape, trust the age only when the two U-Pb clocks Agree.
Find any claim that rests on a single metric. Name a second independent metric that should agree. If you cannot, you are running a discordant zircon.
Rare earths later move centimetres during oil migration (Chapter 10). A beautiful zircon age still needs a check that later fluids did not reset the signal.
4. To the Rendezvous
Sorted on a Silurian shore with higher tides and familiar winds, the grains ride a turbidity current about 100 km into deep water. Each grey stripe is one event; each dark lamina is the quiet rain between storms.
A Silurian shore
About 420 million years ago, in the Silurian, the shoreline of Avalonia lay not much more than 50 miles to the south, near today's Pembrokeshire, but facing north across open sea. Winds worked then much as now; tides did not.
Tides puzzled the ancients: Pliny the Elder recorded their link with the Moon's phases, Seleucus thought the Moon pressed down on the atmosphere, and Bede recognised that high tide was a progressive wave moving along England's east coast. The Marquis de Laplace, who solved what he called the "spiniest problem in all of celestial mechanics," showed the tide is two bulges raised by the Moon (and Sun) on the rotating Earth, with spring and neap tides as the bodies align and diverge. Tidal friction slowly brakes the Earth and pushes the Moon away. So 430 million years ago the Moon was nearer, tides higher, and a year had a little over 400 days.
Winds had gods everywhere: Njord, Vayu, Stribog, Fujin, and Aeolus with the bag of winds he gave Odysseus. Xenophanes thought winds came from the great sea. Evangelista Torricelli, a poor boy with a Jesuit education who invented the barometer and made a vacuum, saw the true cause in the 17th century: winds are driven by differences in air density and temperature, ultimately by the Sun. Waves carry energy, not water; each water particle moves in an orbit.
On the shelf, trilobites and other deposit feeders ate the mud and changed the clay chemistry in their guts, while filter feeders were smothered. Storms such as Hurricane Katrina raise a surge; its retreating ebb surge carries sediment out and leaves a storm layer, or tempestite.
Into deep water
Sediment piles up on the shelf edge until the slope fails. A slump of a single cubic metre is about 25 builder's wheelbarrow loads; big events remove billions of cubic metres. The slurry mixes with water to become a turbidity current, a dense, turbulent underwater flow that can run over 1000 km. The pebble's current probably travelled about 100 km in a couple of hours.
From shelf to basin: a slump becomes a turbidity current that lays down one graded layer. Flute casts on the base of such beds point back towards the book, here the south. Each grey stripe in the pebble, a couple of centimetres thick, is one such turbidite: "a teaspoon or two of what-once-was-mud," part of an event that spread millions of tons of sediment over hundreds or thousands of square kilometres. Cliffs nearby are striped with thousands of them (one in Scotland is 40 m thick). Scoop marks called flute casts on the bases of beds show the currents came from the south.
Between turbidites, which arrived every few decades, the dark laminae, each under a millimetre, built up slowly: carbon-rich hemipelagic mud from drifting fine plumes and a constant rain of dead plankton, "a procession of death," with carbon from volcanoes worldwide passed through air and ocean.
The reading of grey stripes as turbidites and dark laminae as slow background rain is well founded, but the time between them is not directly measured in the pebble: "every few decades" is an estimate by analogy with modern basins. A pebble records events well and durations poorly, a limitation that returns in every chapter.
Tides, winds, and a Silurian shore
The shoreline that sorted the grains lay about 50 miles south, near today's Pembrokeshire, and faced north. Pliny, Seleucus and Bede watched tides; Laplace left Paris during the Revolution and called the two-bulge problem "spiniest." Spring and neap tides, tidal friction, and a Moon that was nearer mean a Silurian year had a little over 400 days.
Wind gods (Njord, Vayu, Stribog, Fujin, Aeolus with his bag of winds) give way to Torricelli: the barometer, the vacuum, and winds driven by temperature and density under the Sun. Xenophanes already linked winds to the great sea and noticed fossils in hills. Waves carry energy, not mass; water moves in orbitals.
Storms, trilobites, turbidites
About 420 million years ago grains were sorted on that shore. Deposit-feeding trilobites changed clay chemistry in their guts; filter feeders smothered under mud. Storms and storm-ebb surges left tempestites (Katrina is a modern image of the energy). On slopes of a few degrees, a cubic metre of slump is about 25 wheelbarrows. Turbidity currents can run more than 1000 km; the pebble's current ran about 100 km in a couple of hours and settled over hours to days. Each grey stripe a couple of centimetres thick is one turbidite, part of millions of tons spread over hundreds to thousands of square kilometres. One Scottish bed is 40 m thick. Flute casts point back to the southern source.
Dark laminae thinner than a millimetre are carbon-rich hemipelagite: nepheloid plumes and a "procession of death" of plankton. Carbon had travelled from volcanoes worldwide via air and sea. Turbidites arrived every few decades by modern analogy; the pebble does not stamp the waiting time.
Laplace's spiniest problem, and a current of millions of tons
Bede watched the progressive tidal wave along the east coast of England; Seleucus and Pliny recorded the pull of the Moon; Laplace left revolutionary Paris still wrestling with two bulges, springs and neaps. Tidal friction has been lengthening the day ever since. A Silurian shore facing north near Pembrokeshire felt higher tides because the Moon was nearer.
Torricelli's barometer replaced wind gods with density and temperature. Waves are energy on orbital water, not mass transport. Storms leave tempestites; Katrina is a modern image of surge and ebb. On slopes of a few degrees, failure becomes a turbidity current that can outrun a ship and travel more than 1000 km. The pebble's current covered about 100 km in a couple of hours. Grey stripes a couple of centimetres thick are the graded beds; flute casts are the arrows pointing home to the southern source. Between them, dark laminae thinner than a millimetre hold the hemipelagic rain and the carbon that had already circled the planet through volcanoes, air and sea.
From wind gods to flute casts
Njord, Vayu, Stribog, Fujin and Aeolus (with his bag of winds for Odysseus) stand at the mythic end of the wind story. Xenophanes already linked winds to the great sea, mocked ox-shaped gods, and noticed fossils in hills. Torricelli's barometer and vacuum made winds a problem of temperature and density under the Sun. On the Silurian shore, deposit-feeding trilobites altered clay chemistry in their guts while filter feeders smothered under mud. A cubic metre of slump is about 25 wheelbarrows; once fluidised, the current can outrun ships. One Scottish turbidite bed reaches 40 m. The pebble's couple-of-centimetre grey stripes are fragments of events that spread millions of tons over hundreds to thousands of square kilometres. Dark laminae hold nepheloid plumes and the procession of death; carbon in them had already travelled from volcanoes worldwide.
Shore sorts, Slope fails, Slurry becomes a turbidity current, Stripe is the graded bed you can still see.
Fill a jar with mixed sand, silt and clay, shake, and let it settle. Sketch the graded bed. Ask what information the jar cannot give you (the time until the next shake).
Turbidites record events and omit waiting times. Incident logs and transaction ledgers fail the same way when they store spikes and invent the gaps.
In distributed systems, cascading retries after a partial outage are the engineered cousin.
5. The Sea
The floor beneath the pebble was deep and dead: no wave marks, no burrows, pyrite in dark mud. Modern Black Sea anoxia and an ice-linked sea-level switch explain how a whole ocean can lose its breath.
Deep, and dead
There are no wave or tide marks in the pebble, so the sea floor lay below about 200 m; beyond that, 300 m and 3000 m look the same. The fine dark laminae are undisturbed: no animal burrowed through them. The sea floor was dead. John Marr noticed this nearly a century ago, saying the sea had been "poisoned below the 100 fathom line."
What could poison it? Not excess salt, and not volcanic gases: the same laminated black muds appear at the same time in the Lake District, Scotland, Nova Scotia, North Africa and Poland. The answer is anoxia, water without oxygen.
The Black Sea is over 2 km deep and has no oxygen below about 300 m; its deep water holds hydrogen sulphide, the rotten-eggs gas, as poisonous as hydrogen cyanide. Ancient Greek and Roman shipwrecks lie preserved on its floor. The Santa Barbara basin off California collects thick turbidites from storms that strike about once a century, and between them, annual varves: a pale flood layer each winter and a dark plankton layer each summer, countable like tree rings.
The pebble's laminae are too wispy to count, a tangle the author calls "spifflication." Some wisps may be single flakes of Silurian marine snow, the constant fall of dead plankton and faecal matter that Rachel Carson called an "eternal snowfall." Others may be the remains of microbial mats, which coordinated by chemical signalling (quorum sensing) and dominated sea floors for some 3 billion years until grazing animals arrived.
Breathing and suffocating
Some pale layers in nearby rocks are burrowed: oxygen came and went. The alternation ran over millions or thousands of years. In Wales the first 5 million years or so of the Silurian were mostly anoxic, and the alternation of anoxic and oxygenated mudrocks is the basis of how the Welsh rocks are mapped.
| Mechanism | How it works | Example |
|---|---|---|
| Too much food | Excess organic matter rots and uses up oxygen (eutrophication) | Modern coastal dead zones |
| Supply cut off | A density lid stops oxygenated water reaching the deep | The Black Sea today |
Anoxia coincided with high sea level (changes of a few tens of metres), oxygenation with low. The link may be ice. The end of the Ordovician had a brief, intense glaciation, with ice on South America and southern Africa, then joined and near the South Pole; ice waxed and waned for some 10 million years either side. Ice grows, sea level falls, and the cold drives the polar motor: dense, salty water sinks near the poles and carries oxygen to the deep. In a warm world the motor weakens and the deep stagnates.
The carbon-burial thermostat hypothesis: anoxic seas bury billions of tons of carbon, cooling the climate until ice restarts deep circulation, which ends the burial and lets the world warm again. Zalasiewicz calls this a "nice hypothesis," a planetary thermostat guarding against both runaway greenhouse and runaway icehouse, and plausible in a world with little land vegetation. It still needs "testing to destruction."
The glaciation the book describes is now usually called the Hirnantian glaciation, about 445 to 443 million years ago, and it coincided with the Late Ordovician mass extinction, the first of the "Big Five," which struck in two pulses, the second linked to expanding anoxia as ice retreated. The book's modern dead zones (Chesapeake Bay, the Gulf of Mexico off the Mississippi, the Baltic) persist: NOAA measured the 2025 Gulf zone at about 4,402 square miles, below average but still more than double the target of under 1,900 square miles by 2035.
How deep, and how dead
Three hundred versus three thousand metres leave the same silence. Undisturbed laminae and pyrite say the floor was dead. John Marr spoke of water "poisoned below the 100 fathom line." The pattern was simultaneous from the Lake District and Scotland to Nova Scotia, North Africa and Poland, so the cause was ocean-wide anoxia, not local volcanic poisoning or salt.
Living analogues
The Black Sea, more than 2 km deep, is anoxic below about 300 m, with hydrogen sulphide as poisonous as hydrogen cyanide, and preserves Greek and Roman wrecks. The Santa Barbara basin records turbidites from century-scale storms and annual varves (pale flood winters, dark plankton summers) countable like tree rings. The pebble's wispy laminae refuse a clean count; Zalasiewicz calls the mess "spifflication." Rachel Carson's "eternal snowfall" of marine snow feeds microbial mats that used quorum sensing and dominated for about 3 billion years until grazers arrived.
Two routes to anoxia, and a thermostat
Pale burrowed layers mark oxygenated states. Alternation lasted millions or thousands of years; the first about 5 million years of the Silurian in Wales were mostly anoxic and still structure the Welsh geological map. Two mechanisms compete: too much organic rain (eutrophication) versus a lid that cuts oxygen supply (the Black Sea pattern). Anoxia tracks high sea level; changes of a few tens of metres flip the switch. The Hirnantian glaciation at the end of the Ordovician (South America and South Africa at the pole) waxed and waned for about 10 million years either side. Polar brine rejection, the motor of deep circulation, weakens in a warm world.
The author's thermostat: bury billions of tons of carbon, cool the world, oxygenate the floor, then release carbon again. He invites the reader to test the idea to destruction. Modern dead zones (Chesapeake fed by the Susquehanna, Gulf of Mexico by the Mississippi, the Baltic) cover thousands of square kilometres seasonally.
Spifflication, marine snow, and testing the thermostat
John Marr's "poisoned below the 100 fathom line" was an early reading of a dead floor. Salt and local volcanoes fail as explanations once the same anoxic signature appears from the Lake District to Nova Scotia and North Africa. The Black Sea's hydrogen sulphide and preserved wrecks, and the Santa Barbara basin's countable varves, are the living textbooks. The pebble's laminae are too wispy for a clean annual count; Zalasiewicz's word for the mess is "spifflication."
Rachel Carson's "eternal snowfall" of marine snow feeds mats that once covered the seas for about 3 billion years. Quorum sensing organised those mats before grazers arrived. Pale burrowed layers mark the oxygenated interruptions that still structure Welsh map units. The thermostat hypothesis (bury carbon, cool, oxygenate, release) is offered as a model to be tested to destruction, not as dogma. Today's Chesapeake, Gulf of Mexico and Baltic dead zones show that anoxia is not only a Silurian story.
High stands, low stands, and polar motors
The first about 5 million years of the Silurian in Wales were mostly anoxic and still structure the Welsh geological map. Sea-level changes of a few tens of metres flip the floor between dark unburrowed mud and pale burrowed layers. The Late Ordovician ice, with South America and South Africa at the pole, waxed and waned for about 10 million years either side of the Hirnantian pulse. Polar brine rejection, the motor of deep circulation, weakens in a warm world, favouring stagnant deep water. Two mechanisms still compete in modern oceans: eutrophication from excess organic rain, and lids that cut oxygen supply. The book's invitation stands: take the carbon-burial thermostat and test it to destruction against the pebble's own alternations.
No wave marks (deep), no burrows (dead), pyrite and dark mud (no oxygen), ice and sea level as the switch.
Write: (1) what presence would look like, (2) whether it would survive, (3) what positive signal should accompany the absence. Run it on "no burrows implies anoxia."
Negative feedbacks stabilise until they saturate. Always name the conditions under which your favourite thermostat becomes an accelerator.
NOAA's mid-summer 2025 Gulf forecast footprint was about 4,402 square miles of low oxygen. The end-Ordovician glaciation is now routinely named Hirnantian (about 445-443 Ma).
6. Ghosts Observed
Acid frees the durable dead: acritarchs, chitinozoans and graptolites. Charles Lapworth's zones turn one colony into a date finer than a million years, the shared standard that ties Wales to the world.
Dissolving the rock
The pebble still holds over 1 per cent carbon; originally it was nearer 10 per cent. To find what living things it came from, palynologists dissolve the rock in hydrofluoric acid, leaving organic residues. Much is amorphous, but some is shaped.
Acritarchs ("things of confused origin") are mostly the resting cysts of algae, numbering thousands in the pebble; their modern relatives include the dinoflagellates behind toxic red tides such as Pfiesteria. The author bets on finding Moyeria, shaped like a tiny deflated rugby ball, a freshwater form related to Euglena and washed out to sea, whose spiral pattern suggests it swam towards light. Living Euglena with that spiral marking do move towards light, a behaviour studied, the author notes with envy, at the "Laboratory of Marine Serendipity and Spectaculars" in British Columbia. Euglena is neither quite plant nor quite animal: it photosynthesises, but also hunts and eats other microbes, and forms cysts in bad times.
Chitinozoans are flask-shaped, pinhead-sized, and the pebble may hold dozens to hundreds. They were discovered in the 1930s by Alfred Eisenack, an East Prussian whose research survived two long spells as a prisoner of war; of one, in Siberia after the First World War, he wrote that it was the best spell of his life. Their walls are "pseudochitin," and their describers use words like scabrate, vermiculate, foveolate and verrucate. The leading hypothesis is that they are the egg cases of an unknown soft-bodied planktonic animal a few millimetres to centimetres long, since they are found in chains and spiral clusters. A "Rosetta Stone" slab with the parent attached has never been found.
A graptolite under the needle
The gold fleck on the pebble's surface, with a black rim and pale fibrous mineral, is a graptolite. Freeing it is brain surgery on rock: an old steel gramophone needle in a pin-vice, under a binocular microscope. Graptolites were colonies of tiny zooids, each housed in a cup (theca) about 1 mm apart along a tube of collagen, built ring by ring with visible growth bands. The colonies drifted in the plankton.
Their living relatives are the pterobranchs ("winged arm"), shy animals of Norwegian fjords, Plymouth Sound and the Bahamas, which build their tubes ring by ring with a cephalic disc, much as a graptolite did. Graptolites also added cortical bandages over the outside, plastering the colony as if with a trowel. Beside the pterobranchs' untidy, spaghetti-like tubes, graptolite colonies were, in the author's words, a Fabergé egg against a child's clay model. Whether they floated passively or swam is still debated.
The book links graptolites to pterobranchs. A 2013 phylogenetic analysis (Mitchell and colleagues, Lethaia) went further: the living pterobranch Rhabdopleura nests within the graptolites, sharing the features that define the group. In current classification, graptolites (Graptolithina) are pterobranch hemichordates, and Rhabdopleura is an extant graptolite.
Charles Lapworth's clock
Graptolites lived from about 500 to 400 million years ago and evolved fast, so they tell time. Charles Lapworth, a bookish child from Faringdon and admirer of Walter Scott, became a schoolteacher in Galashiels to live in Scott's country. Walking the Southern Uplands, where 99 per cent of the rock is barren sandstone and the strata were thought monstrously thick, he found thin graptolite-bearing mudrocks and mapped them across some 300 square miles.
| Stage | Graptolite time resolution |
|---|---|
| Lapworth, 19th century | Ten successive assemblages, each in a metre or two of mudrock; repeats meant folding, not monstrous thickness |
| Rival view | Joachim Barrande in Bohemia explained out-of-place faunas as "colonies" |
| Today | Over 60 zones, each under a million years on average |
| Species | Several thousand worldwide; 697 in Britain at the last count |
| The pebble | One identifiable graptolite dates it to within a single zone |
A single graptolite can date the pebble to within less than a million years, which is better than many radiometric dates of the same age. Acritarchs are less reliable, because they can be reworked from older rocks.
Lapworth's method works because everyone agrees on the reference sequence and its definitions. The value lies less in any one specimen than in a controlled vocabulary that lets rocks in Wales, Bohemia and Arabia be compared. Any organisation reconciling records across systems needs the same thing.
Acid, cysts, and flasks
Carbon in the rock is now a little over 1 percent; it was about 10 percent when the mud was soft. HF maceration frees amorphous organic matter for palynologists. Acritarchs ("things of confused origin") are algal cysts by the thousand; their kin include dinoflagellates such as Pfiesteria of red-tide fame. Moyeria, a deflated rugby ball from freshwater, is a relative of Euglena with a spiral pattern that hints at phototaxis; the Laboratory of Marine Serendipity and Spectaculars in British Columbia studies such oddities.
Chitinozoans are flask-shaped, pinhead-sized, found in dozens to hundreds. Alfred Eisenack, an East Prussian working in the 1930s, named them between two spells as a prisoner of war (he called Siberia "the best spell of his life"). The wall is "pseudochitin"; adjectives such as scabrate, vermiculate, foveolate and verrucate describe the ornament. They are likely egg cases (chains, spiral clusters) of a soft-bodied planktonic animal still awaiting its Rosetta Stone.
Graptolites and Lapworth
The gold fleck is a pyritised graptolite: collagen tubes, thecae every millimetre, colonies that floated or swam. Excavate with a gramophone needle in a pin-vice; draw with a camera lucida. Living pterobranchs (Rhabdopleura) in fjords and Plymouth Sound still build tubes ring by ring. Time control belongs to Charles Lapworth, the Galashiels schoolteacher who walked more than 300 square miles of the Southern Uplands, found ten assemblages repeating every metre or two, and proved folding against Barrande's "colonies." Today more than 60 zones, each under a million years on average, turn one specimen into a date. Cosmopolitan index fossils beat local thickness every time.
Eisenack's flasks and Lapworth's walk
Alfred Eisenack named chitinozoans in the 1930s between two spells as a prisoner of war; he called Siberia the best spell of his life. Pseudochitin walls take adjectives such as scabrate, vermiculate, foveolate and verrucate. Chains and spiral clusters favour an egg-case interpretation for an unknown soft parent.
Charles Lapworth, teaching in Galashiels, walked more than 300 square miles of the Southern Uplands and found the same ten assemblages repeating every metre or two. Barrande's "colonies" could not survive that map: the repetition was folding. More than 60 zones now slice the Ordovician and Silurian into averages under a million years. One graptolite from the pebble, excavated with a gramophone needle and drawn with a camera lucida, can lock the bed to a single zone and correlate Wales with Bohemia. Living Rhabdopleura still builds tubes ring by ring in fjords and in Plymouth Sound.
Moyeria, Pfiesteria, and the shared standard
Acritarchs number in the thousands of kinds; some kin cause red tides (Pfiesteria). Moyeria, a deflated rugby ball from freshwater, carries a spiral hint of phototaxis and sits near Euglena. The Laboratory of Marine Serendipity and Spectaculars in British Columbia studies such oddities. Carbon in the rock is a little over 1 percent now and was about 10 percent when soft. HF maceration is the gateway. The payoff of Lapworth's zones is geopolitical as well as geological: a Welsh cliff and a Bohemian quarry share a clock without sharing a language.
Acid frees the organics, flask (chitinozoan) and graptolite under the needle, date by Lapworth's zone.
Take any layered log. Define assemblages that recur. Ask whether repetition means cycles in one place or folding of the same stack.
Acritarchs can be reworked into younger mud. If your "index" can migrate without context, concordance with a second clock is mandatory.
Mitchell and colleagues (2013) reinforced the link between living Rhabdopleura pterobranchs and fossil graptolites. The planktonic graptolites themselves remain extinct since the Devonian.
7. Ghosts in Absentia
Most of the Silurian plankton left nothing. Reading absence, carbon isotopes and the promise of Re-Os dating, the chapter asks what a fossil record is allowed to forget.
What did not survive
Sir Alister Hardy's The Open Sea, Part 1, a whole book on plankton, changed the author's view of the world. Most of what Hardy describes left nothing in the pebble. Diatoms, coccolithophores and dinoflagellates had not yet evolved; radiolaria lived but cannot be extracted from slate. Krill, copepods, comb jellies, arrow worms, salps, jellyfish and pteropods (now threatened by ocean acidification) leave no trace.
So how big a share of the plankton were graptolites? Perhaps 10 per cent, perhaps under 1 per cent; there is no "graptolite coal." Some colonies show signs of being caught, folded like paperclips. Modern plankton make a daily vertical migration of about 100 m; graptolites may have too.
| Clue | What it shows |
|---|---|
| Isorenieratane | Pigment of a green, sulphur-metabolising photosynthetic bacterium that cannot tolerate oxygen; found in early Silurian rocks of North Africa, so anoxic water reached nearly to the surface (not yet found in Wales) |
| Parasite blisters | Graptolite zooids walled over attached parasites |
| Carbon isotopes (12C vs 13C) | Plankton prefer light 12C; mass burial leaves the ocean enriched in 13C, recorded worldwide |
| Gene sequencing (modern analogue) | Millions of bacterial species in a few litres of sea water; about 10 viruses per microbe |
Graptolites may have been low-oxygen specialists, booming and crashing as conditions changed; the planktonic forms died out in the mid-Devonian, about 50 million years after the pebble's time. Whether they hosted symbiotic algae, as corals do (and lose when they bleach), is unknown, but the author insists it is not an unknowable question.
Carbon isotopes and a new clock
A large carbon isotope excursion just before the Silurian, enriched in 13C, coincides with the glaciation; Zalasiewicz notes the alternative that eroding limestones could cause it. "You are what you eat, plus one part per mil": predators are slightly enriched in 13C over their prey, so isotopes might one day show what graptolites ate.
For dating black shales directly, he points to rhenium-osmium: rhenium, rarer than gold, decays to osmium, the densest metal (22.6 g/cm³), and both concentrate with organic matter and sulphur. Several samples give an isochron. In 2010 it was "promising, not routine," and the Welsh Silurian has few volcanic ash layers for conventional dating.
Since the book, Re-Os geochronology of organic-rich sedimentary rocks has become a standard tool for dating black shales and petroleum systems, and is used to calibrate parts of the geological time scale where ash beds are lacking. The best shale ages now reach about 1 per cent precision, though results still depend on the samples sharing the same initial osmium composition.
Chapters 5 to 7 rely heavily on negative evidence: no burrows means no oxygen; no fossils means no hard parts. Zalasiewicz handles this carefully, using independent positive evidence (isorenieratane, pyrite, modern analogues) where possible. The general rule: an absence is only evidence when you know what presence would have looked like and that it would have been preserved.
Hardy's open sea, rewritten for the Silurian
Alister Hardy's The Open Sea is the lens. Diatoms, coccolithophores and modern dinoflagellates had not yet arrived; radiolaria were present but hard to extract from these rocks; soft swimmers (krill-analogues, comb jellies, arrow worms, salps, jellyfish) left almost nothing. Graptolites may have been under 1 percent of the plankton yet dominate the durable archive. Soft majorities leave hard minorities.
Biomarkers and chemical ghosts
Isorenieratane from green sulphur bacteria in North African early Silurian rocks shows anoxia reaching nearly into the sunlit zone; Wales still awaits that biomarker. Carbon isotopes add a ghost clock: plankton prefer ¹²C, so mass burial leaves the ocean heavy in ¹³C, a shift readable worldwide. Rhenium-osmium dating of organic mud (rhenium rarer than gold; osmium the densest metal) is promising and not yet routine in these Welsh rocks. Absence is evidence only when presence would have been preserved and recognised.
Soft majorities, hard minorities
Hardy's plankton census, rewritten for the Silurian, is mostly ghosts. Diatoms and coccolithophores had not evolved; radiolaria hide in these lithologies; jelly and salp leave almost no imprint. Graptolites may have been under 1 percent of the living community and still dominate the durable record. That asymmetry is the chapter's warning for every archive that keeps what is hard and loses what is common.
Isorenieratane in North African early Silurian rocks pushes anoxia into the photic zone; Welsh rocks still await that biomarker. Carbon-isotope excursions record mass burial of light carbon. Re-Os dating of organic mud offers a chemical clock when ash beds are missing, now often near 1 percent precision where conditions allow, though not yet routine for these Welsh samples.
What the open sea refuses to leave behind
Krill-analogues, comb jellies, arrow worms, salps and jellyfish almost certainly filled Silurian water columns and left almost no body fossils here. Radiolaria were present but hard to free from these lithologies. The durable minority (graptolites, acritarchs, chitinozoans) therefore misrepresents abundance. Carbon-isotope shifts still record the soft majority's burial as a chemical ghost. When ash beds are missing, Re-Os on organic mud is the candidate clock; Welsh Silurian application remains sparse, which is itself a finding about where the method has been pushed.
Soft majorities and the archive's bias
The Silurian water column was almost certainly full of soft-bodied swimmers that leave almost no imprint in Welsh slate. Comb jellies, salps, arrow worms and jelly-like forms fill Hardy's modern census; their Silurian cousins are inferred, not seen. Radiolaria were present in the period but hard to free from these lithologies. Graptolites, acritarchs and chitinozoans therefore dominate the durable record even if they were a minority of the living community. That bias is not a failure of geology; it is a property of every archive that keeps what is hard and loses what is common.
Carbon-isotope excursions still record the soft majority's burial as a chemical ghost: light carbon locked in mud leaves the ocean heavy. When ash beds are missing, Re-Os on organic mud is the candidate clock. Welsh Silurian application remains sparse, which is itself a finding about where the method has been pushed. The chapter's discipline is to name the gap, back absence with positive signals where possible, and refuse to invent soft bodies the rock cannot support.
Read Hardy to see what is missing, respect the gap, use carbon isotopes as a chemical ghost, try Re-Os when ash beds will not help.
List ten actors in a system you know. Mark which leave durable records. Write one decision you would not make from the recorded set alone.
Skip the backups (pyrite, analogues, biomarkers) and "ghosts in absentia" becomes storytelling.
Rhenium-osmium dating of organic-rich mudrocks now often reaches about 1 percent precision in favourable sections.
8. Where on Earth?
A frozen magnetic compass puts the sea floor at about 35°S. The pebble then rides more than 10,000 km north, while field mapping shows a half-graben, not a textbook fan, trapped the sand.
A compass frozen in rock
Tiny magnetic grains (magnetite, goethite) align with the Earth's field as sediment settles, recording it. The story of magnetism is old: the Chinese knew lodestone over 2000 years ago and made a compass from a lodestone spoon on a smooth board; Europe followed about a thousand years later. In 1600 William Gilbert moved a lodestone needle over a lodestone sphere (a terrella) and realised the Earth itself is a magnet: a needle dips not at all at the equator and straight down at the poles.
| Measure | Value from the book |
|---|---|
| Pebble sea floor, Silurian | about 35°S (like the north tip of New Zealand or the south tip of Africa today) |
| Devonian to early Carboniferous | southern subtropics, about 5 km underground |
| Equator crossed | about 310 million years ago, under a tropical coal swamp |
| About 250 million years ago | beneath a desert like the Sahara |
| Total journey | over 10,000 km in a little over 400 million years |
| Average speed | more than 2 cm a year, almost all underground |
The geodynamo in the liquid outer core makes the field. The magnetic pole wanders (now over 10 km a year) and the field reverses, lately every few hundred thousand years; in the Silurian, flips were frequent, with one quiet phase. The field shields life, as Venus and Mars, which lack one, show.
Oceans and a freezing core
In the Silurian, Avalonia had already closed most of the Iapetus Ocean to its north (towards Scotland and North America). To the south, Gondwana, centred on Africa, lay over 1000 miles away across the Rheic Ocean, which was beginning to close and would vanish in another 100 million years.
Deep below, the core was freezing. The liquid outer core is about 7000 km across; the solid inner core about 2400 km. At today's rate of about 5000 tons a second, the century it took to lay down a pebble's worth of sediment would freeze some 1.5 thousand billion tons of metal. The inner core began to form roughly a billion years ago (estimates ranged, he says, from over 3 billion to half a billion), and a full freeze in perhaps a billion years would end the magnetic field.
Since the book, palaeomagnetists have pushed the inner core's birth toward the young end of his range. A 2019 Nature Geoscience study found an ultra-weak field about 565 million years ago (Ediacaran) and inferred the inner core began to nucleate soon after; a 2022 Nature Communications study put renewal of the geodynamo at about 550 million years ago. But 2025 results show the field stayed weak through much of the Cambrian (about 525 to 505 million years ago), so the exact nucleation age remains unresolved.
A basin like the North Sea
The Welsh basin sat on stretched Avalonian crust, with no ocean-floor basalt: more like the subsiding North Sea than the Atlantic. Fossils help too. Time fossils are cosmopolitan; place fossils are endemic, like Galapagos finches or Komodo dragons. The pebble's graptolite is found around Avalonia, into central Europe, and across the narrowing ocean towards Scotland and North America, but probably not in the tropical seas of the time (ironically, today's Arctic Canada), which held a different plankton. Silurian graptolite faunas also show a drop in diversity from low to high latitudes, like the change today where cold polar currents meet temperate seas: a "polar front" effect.
Fossil species are morphospecies, defined by shape. Splitters and lumpers disagree, and many modern species are cryptic, identical in form but genetically distinct. Zalasiewicz suspects that a single graptolite "species" ranging from pole to tropics is probably several cryptic species (it comes in broad and narrow varieties). The same tension affects any classification built from observable features alone.
Walking across a sea floor
Field geology needs boots, bramble-proof trousers, a cagoule and a sunhat; most Welsh rock is hidden under soil, vegetation and ice-age debris, so the geologist hunts it in streams, gorges, crags and farmyard quarries. The strata have been crumpled and dislocated, leaving a four-dimensional jigsaw. Zalasiewicz calls field geology "the ultimate forensic science, the art of the possible," whose result is "an idea of a sea floor that we are building in our heads," a working model to be tested and revised. The striped strata of the Aberystwyth cliffs were among the first rocks interpreted as turbidity current deposits, "quite a breakthrough in geology." Early researchers, working along the cliffs, saw beds thinning and fining to the north and deduced a turbidite fan of steadily weakening currents. It became a standard model for strata worldwide, and it is "right, but wrong": the sea lay to the west and the crumpled hills to the east, so nobody had looked across the basin. Mapping east-west with graptolite zones (units of about half a million years) revealed the surprise.
The Aberystwyth half-graben: strata thicken from a few hundred metres at the coast to over two kilometres a few miles inland, then thin abruptly against a fault that trapped the turbidity currents. Strata thicken from a few hundred metres at the coast to over two kilometres a few miles inland, with individual beds over 2 m thick, then abruptly give way to thin mudrocks. The cause was a half-graben: a crustal trapdoor hinged in the west and faulted in the east, which sagged, captured the currents (positive feedback: more sediment, more sag), filled and seized up in a little over a million years.
Closing oceans and a freezing core
Place fossils and morphospecies
Time fossils are cosmopolitan; place fossils are endemic (Galápagos finches, Komodo dragons). The pebble's graptolite turns up around Avalonia and Europe across a narrow ocean, less in the tropics (Arctic Canada). Fossil "species" are shapes: splitters versus lumpers, cryptic species, polar-front diversity drops. A graptolite drawn from pole to tropics is probably several cryptic species.
Ten thousand kilometres underground
Over 10,000 km at more than 2 cm/year, almost all underground. The Iapetus closed northward; the Rheic closed southward. The inner core still freezes at about 5000 tons a second; a century adds about 1.5 thousand billion tons and about 5 cm of diameter.
Biogeography splits time fossils from place fossils. Morphospecies inflate ranges; cryptic species hide inside one drawing. Field mapping with graptolite time-slices exposed the half-graben that the textbook fan had missed: thickness jumping from a few hundred metres to over 2 km, then snapping to thin mudrocks, filled and seized in little over a million years.
North Sea geometry under Wales
The Welsh basin sat in stretched Avalonian crust without ocean basalt: North Sea geometry, not Atlantic. Positive feedback filled the half-graben and then seized it in little over a million years. Place fossils endemic to Avalonia and Europe across a narrow ocean are less common in the tropics. A single morphospecies drawn from pole to tropics is a warning label: broad and narrow varieties may be cryptic species. The compass reading of about 35°S is the fixed point; everything else in the chapter is the path and the trap that caught the sand.
Compass (palaeolatitude), Core (freezing geodynamo), Basin (stretched crust), Fault (half-graben that traps the sand).
Take a reference architecture you use. Shade where local constraints capture the traffic. Label the shaded zone your half-graben.
Customer segments and alert types built only on visible features fall into the same trap as morphospecies.
Ultra-weak field about 565 Ma (2019) and dynamo renewal about 550 Ma (2022) pushed birth young; 2025 results showing a persistently weak Cambrian field keep the age open.
9. Gold!
Shallow burial turns mud into a microbial metal shop: pyrite framboids, phosphate at the redox boundary, then methane when sulphate runs out. Fool's gold here is a clock of early diagenesis.
Fool's gold
The pebble's mud layer was then about 25 cm thick. With no oxygen, microbes breathed sulphate, producing sulphide that combined with dissolved iron to make pyrite (FeS2), fool's gold: lighter than gold, brittle, and tarnishing. Pyrite often grows as framboids (French framboise, raspberry), balls of hundreds of microcrystals, about 9 thousandths of a millimetre across if formed in the water column and up to 0.1 mm in the mud. They can grow in sterile laboratory conditions, but microbes help. Sulphur travels globally (as volcanic gas); iron travels only millimetres to metres.
Pyrite also filled the graptolite chambers, 2 to 3 mm long, quickly, before compaction flattened them. The dissolved ingredients arrived by compaction-driven upward flow and diffusion, the random jostling of molecules.
Random molecular motion is visible as Brownian motion, named for the botanist Robert Brown, a Scot celebrated for pursuing science "with constancy and a cold mind" (level-headedness, not coldness: he once teased Darwin with "it's my little secret"). His death freed the slot at the Linnean Society on 1 July 1858 at which the Darwin and Wallace papers were read. Brown saw the jiggling in century-old pollen and in mineral dust, showing it was physical, not living. Later workers, including Albert Einstein, explained it as the bombardment of invisible water molecules.
Electric microbes and phosphate
The dark stripes are cemented by apatite (calcium phosphate), which weathers white. It formed at a redox boundary a few centimetres below the sea floor, where oxygen-free and oxidised chemistry meet, the same place uranium precipitates today. Changes at such boundaries propagate faster than diffusion allows, and the best candidate is bacteria acting in concert, relaying electrons over distances of about 10,000 body lengths.
These cable bacteria are now classed in the genera Electrothrix (marine) and Electronema (freshwater).
Methane and deeper burial
Methane clathrate fizzes and burns at the surface; it is a promised fuel and a threat if a warming climate destabilises it. The pebble may have passed through a clathrate zone, but the evidence is ambiguous. At 100,000 years the pebble-to-be was still about 50 per cent taller than now, and its fossils were yellowing.
Framboids, Fantastic Voyage, apatite
The mud layer was then about 25 cm thick. Microbes reduce sulphate to sulphide; with dissolved iron it becomes pyrite (FeS₂), lighter and more brittle than gold, prone to tarnish. Framboids (from framboise, raspberry) pack hundreds of microcrystals: about 9 thousandths of a millimetre in the water column, up to 0.1 mm in mud. They grow in sterile labs (self-organising) though microbes help. Sulphur travels globally as volcanic sulphur dioxide; iron may move only millimetres. Hundreds to thousands of framboids sit in the pebble.
Pyrite fills graptolite chambers 2-3 mm across by diffusion and compaction-driven upward flow, a "Fantastic Voyage" that finishes before flattening. Partially flattened fills mark the race. X-rays of the pebble reveal graptolites the eye missed. Robert Brown saw the jitter now explained by Albert Einstein (the book slips and says "Alfred"); he had watched century-old pollen and mineral dust with "constancy and a cold mind."
Apatite (calcium phosphate) cemented the dark stripes at the redox boundary a few centimetres down, where uranium still precipitates today and where cable-like bacteria ferry electrons across about 10,000 body lengths (centimetres) faster than diffusion.
Methane, clathrates, rotting feldspars
By about 10,000 years and 10 m burial, sulphate is gone; fermentation yields heavy carbon dioxide and light methane. Carbonate nodules are rare in this calcium-poor mud. Methane clathrate (ice cages that burn) may have hosted the pebble for a time; soft-sediment slides are ambiguous evidence. By about 100,000 years, more than 100 m down and about 3°C warmer, the rock is plastic, microbes may divide once a year, and the pebble is about 50 percent taller and fatter. Feldspars rot to clays; potassium, sodium and calcium leave; potassium drives the smectite-to-illite conversion that will matter in the oil window.
Brownian secrets and electrical relays
Robert Brown won a Linnean Society slot that opened on 1 July 1858, the day Darwin and Wallace were read into the record. He watched jitter in century-old pollen and in mineral dust and kept "constancy and a cold mind." Einstein's explanation made the motion a thermometer of molecular scale. Pyrite's depletion conveyor into a graptolite chamber is the same physics plus chemistry: dissolve nearby, precipitate in the sink, finish before compaction flattens the tube.
Apatite seams mark the redox boundary a few centimetres down, where uranium still drops out of seawater. Cable bacteria of the Desulfobulbaceae, named in genera Electrothrix and Electronema, ferry electrons across centimetres, about 10,000 body lengths, faster than diffusion. Pfeffer and colleagues confirmed them in 2012, two years after the book's electrical speculation. By 10,000 years sulphate is gone and methane appears; by 100,000 years feldspars rot and the rock turns plastic.
From centimetres to a hundred metres
In the first days to centuries, microbes make framboids centimetres down. Sulphur may have circled the planet as volcanic sulphur dioxide before landing in the mud; iron may have travelled only millimetres. Hundreds to thousands of framboids sit in one pebble. Partially flattened pyrite fills show the race against compaction. X-ray images reveal graptolites the eye missed. Carbonate nodules stay rare because the mud is calcium-poor. Soft-sediment slides hint that the pebble may have spent time in the methane-clathrate stability zone, though the evidence is ambiguous. At about 100,000 years the pebble is about 50 percent taller and fatter; fossils are yellowing; potassium leaves as smectite begins its long conversion to illite.
Sulphate feed, framboid pyrite, cable electrons across the redox line, then methane when sulphate is exhausted.
In any layered process, mark where the "oxidising" control stops and the "reducing" exception path begins. What precipitates there?
Framboids grow in sterile labs. Demand a test that could show the agent is unnecessary before you fund the agent story.
Pfeffer and colleagues (2012, Nature) identified centimetre-long Desulfobulbaceae filaments; genera Electrothrix and Electronema now name them. The book's electrical-relay speculation was confirmed within two years.
Cache invalidation, liquidity buffers and attention queues fail in the same geometry when the sink is faster than the refill.
10. The Oil Window
At about 2 km and 100°C the mud yields a teaspoon of oil and grows monazite. Discordant clocks become clues: inherited neodymium misleads; uranium-lead dates the fluid pulse.
Rapid burial and rare earths
Three to five million years later the mud lay about two kilometres down, buried quickly as the Iapetus closed in a "soft collision" that uplifted the book land to the south. At about 100°C, life fades (vent extremophiles survive to 113°C).
Tiny dimples about 1 mm across in the dark layers are monazite, a phosphate of rare earth elements (lanthanum, cerium, neodymium, samarium and others). These grew in the mud (they enclose clay and silt), with cores rich in heavy rare earths such as gadolinium and rims rich in light ones (cerium, lanthanum). X-rays show them swarming in the dark layers, which hold about ten times the average rare earth content while the pale layers are depleted; the whole rock is normal. The rare earths, long assumed insoluble and used to trace sediment sources, had moved over centimetres, an awkward finding for provenance studies.
Oil
The pebble has produced its share of oil: perhaps a fraction of a teaspoonful. Oil generation began at one to two kilometres; the oil window is about 80 to 150°C, between roughly 2 and 5 km, yielding alkanes, cycloalkanes and aromatics, with a gas window (methane, ethane, propane, butane) another kilometre or two deeper. The pebble's carbon fell from about 10 per cent to 1 to 2 per cent.
How oil escapes from mud is still uncertain. It forms tiny elongated droplets, which oilmen call slugs, that must squeeze through infinitesimal gaps between packed clay flakes, "an escape trick worthy of Houdini"; gas pressure may help by fracturing the rock. Once in roomier sandstones, it travels upward more easily; seeps feed sea-floor communities, as in the Gulf of Mexico today. Wales may once have been an oil province to rival the Middle East: the main oil source rock in Saudi Arabia is a Silurian organic-rich mudrock laid down in an anoxic sea, "the exact equivalent" of central Wales, with beautiful graptolites. Here the reservoirs were eroded long ago, and speculative boreholes found nothing.
The Saudi source rock is the Qusaiba "hot shale," the organic-rich lower member of the Early Silurian (Llandovery) Qalibah Formation, the principal source for the Arabian Paleozoic petroleum system, including major gas fields in central Saudi Arabia and the Qatar-Iran area. Because it lacks woody material, its thermal maturity is measured partly from the reflectance of graptolites and chitinozoans: the pebble's fossils, used as industrial thermometers.
The author's "factor X" for the rare earths: organic compounds raise their solubility, and migrating oil may have carried them into the organic-rich layers.
Two clocks in one monazite
| Clock | Result | Why |
|---|---|---|
| Samarium-neodymium | Over 100 million years too old | The crystal inherited extra neodymium, winding the clock forward |
| Uranium-lead (tiny amounts) | 415 to 420 million years | Several million years after deposition: dates oil generation and fluid flow |
Meanwhile the mud became rock: feldspars vanished, quartz dissolved under pressure and re-precipitated as cement, and smectite (the shrink-swell clay that cracks houses) turned to illite at about 2 km, releasing a pulse of water. A rubidium-strontium isochron from several pebbles was reset at the same time as the monazite, about 20 million years before mountain-building, consistent with that fluid pulse. The rock now broke along its bedding. Then came some 20 million years of quiet.
Each radiometric clock dates a specific event: crystallisation, closure to diffusion, a fluid pulse. The monazite shows that the same crystal can give a misleading answer from one system and a meaningful one from another. The question is always "what reset this clock?", not just "what number does it give?"
Burial rate and the monazite swarm
Three to five million years later the mud sat more than 2 km down near 100°C. Burial was rapid: the Iapetus had closed in a "soft collision," uplifting the South Wales source while mountains waited. Extremophiles live to about 113°C at vents; here life was fading.
Monazite dimples about 1 mm across grew in the dark layers: rare-earth phosphates with inclusions of clay and silt, not detrital grains. Rare earths were long treated as insoluble provenance tracers; that habit misleads. Zoning shows heavy rare earths (gadolinium) in cores and light ones (cerium, lanthanum) on rims. X-rays show monazites swarming in dark layers that hold about ten times the rare-earth content of average mud, while pale layers are depleted and the whole rock looks normal: redistribution over centimetres.
Oil, gas, and Factor X
Carbon fell from about 10 percent to 1-2 percent. Oil migrates as buoyant "slugs," Houdini-like, sometimes fracturing rock with gas pressure (as in the Gulf of Mexico today). Seeps feed communities; traps and biodegradation act within about 1 km of the surface. Wales may once have rivalled the Middle East; Arabian Silurian anoxic shale still holds beautiful graptolites. Welsh reservoirs are gone; speculative boreholes failed.
Factor X: organic complexing raises rare-earth solubility, so oil migration carried them into organic layers. (Rare earths also appear in high-temperature superconductors.) Sm-Nd in monazite ran more than 100 million years too old because of inherited neodymium; U-Pb gave 415-420 Ma, dating the oil pulse. Feldspars were gone; quartz pressure-solved into silica cement; smectite dehydrated to illite at about 2 km, releasing a water pulse that reset Rb-Sr on the same schedule. Then about 20 million years of quiet.
Houdini oil and a misleading tracer
Oil moves as buoyant slugs, sometimes fracturing rock with gas pressure, feeding seep communities, biodegrading within about 1 km of the surface. Wales may once have rivalled Middle Eastern provinces; the Arabian Qusaiba hot shale still shows what Silurian anoxic mud can do, with maturity read from graptolite and chitinozoan reflectance. Welsh reservoirs are eroded; speculative boreholes failed.
Rare earths, long treated as immobile provenance tracers, rode the oil as organic complexes into dark layers, leaving pale layers depleted and whole-rock averages looking normal. Sm-Nd ages wound more than 100 million years too old from inherited neodymium; U-Pb on tiny uranium in the same monazite dated 415-420 Ma, the oil pulse. Rb-Sr reset with the smectite-to-illite water release. Discordance was the clue, not the error.
Twenty million years of quiet
After the fluid pulse, about 20 million years passed before the Acadian squeeze. Rock then broke along bedding. Feldspars were gone; quartz had pressure-solved into silica cement; clays had grown. Extremophiles at vents live to about 113°C; near 100°C and more than 2 km down, life in the mud was fading. Gas (methane, ethane, propane, butane) belongs to a window another 1-2 km deeper than oil. Factor X, organic complexing, is why rare earths moved with the oil into dark layers that hold about ten times average mud concentrations while pale layers look depleted. Whole-rock averages stay normal and hide the centimetre-scale redistribution.
Enter the oil window, grow monazite, read discord between Sm-Nd and U-Pb as a clue, tie both to the smectite-illite fluid pulse.
Find two systems that should date the same event. If they disagree, write three hypotheses: lost daughter, inherited parent, wrong event.
Schedule a review of which features you still treat as immovable, and what observation would show they have migrated.
Saudi Arabia's principal Lower Silurian source is the Qusaiba hot shale. Maturity is read partly from graptolite and chitinozoan reflectance.
11. Making Mountains
Docking against Iberia and Armorica squeezes mud into slate about 396 million years ago. Cleavage, barrel micas and an argon "heartbeat" record slow mountain-building, not a single snap.
A squeeze from the south
A little under 400 million years ago, in the Devonian, the pebble lay 3 to 4 km deep at over 100°C, sterile. Above, graptolites had boomed and gone, fish thrived, and land was greening. The sea had filled to a coastal plain. The Iapetus Ocean, over 1000 km wide 50 million years earlier, had closed about 20 million years before, but its oblique docking caused little deformation.
The decisive force came from the south. The Rheic Ocean was disappearing, and Avalonia docked against Gondwana's promontory of Iberia (now Spain) and Armorica (largely France); as evidence accumulates, the timing matches the Welsh mountain-building.
| Process | Value in the book |
|---|---|
| Direction of squeeze | north-west to south-east |
| Crustal shortening | about 10 per cent at first, ending perhaps a third narrower |
| Peak temperature | climbing towards 250°C (compare 400 to 600°C in the Alps, Andes, Himalayas) |
| Mass lost from mudrock | up to a fifth, mostly silica |
| Fluid left in the rock | a few per cent |
| Barrel-shaped micas | grew to ten or more times their original thickness |
Slow force makes rock flow (like Silly Putty or glacier ice); fast force makes earthquakes. Under pressure, platy micas rotate and regrow at right angles to the squeeze, like paper pressed by a hand, creating slaty cleavage: the planes along which craftsmen split roof slates less than a quarter-inch thick. The pebble's disc shape follows the cleavage, not the bedding, which is why its stripes run askew across its faces. Tiny microfolds, crenulations, wrinkle the cleavage.
The silica squeezed out travelled by paths still mysterious, perhaps emerging in mineral springs that fertilised the first land plants (the author once bought mineral water whose label listed mercury).
Barrels, bandages and a bull's-eye
Some micas stand on the bedding surfaces like barrels, grown ten or more times thicker by fluids adding layer after layer, "converting a slim paperback into the heftiest of tomes." Micas engulfed by the monazites 20 million years earlier stayed flat, protected "like steel ball-bearings embedded within plasticine."
Graptolites were broken into segments, and mica coatings grew on them in pressure shadows. Tellingly, micas also coat graptolites that are only carbon films, while other objects attract quartz: the organic matter catalysed mica growth. That matters for the Burgess Shale, whose celebrated soft-bodied fossils also carry shiny mica films, once read as clues to early preservation; in the pebble's light, they likely grew millions of years later.
Potassium in the mica coats decays to argon. Ordinary slate gives meaningless mixtures of old and new mica, but the coats on graptolites are pure tectonic mica. The first attempt "scored a bull's-eye": 396.1 ± 1.4 million years, unheard-of precision for a slate. Later laser spot profiles found zones of about 395 million years alternating with zones of excess argon that give ages too old: rhythmic fluid pulses, the heartbeat of a mountain belt.
| Event | Age (million years ago) | Clock |
|---|---|---|
| Mud deposited | a little over 420 | Graptolite zone |
| Oil generation, fluid pulse | 415 to 420 | Monazite U-Pb; Rb-Sr |
| Slaty cleavage | about 396 | Argon in mica coats |
| Gap between burial and squeeze | about 20 million years | Comparison of the clocks |
Soft docking, then a southern squeeze
Less than 400 million years ago, in the Devonian, graptolites had gone through boom and bust; fish thrived; land was greening; the sea had filled to a coastal plain. The Iapetus, more than 1000 km wide 50 million years earlier, had closed about 20 million years before with oblique docking and little deformation. Force then came from the south: the Rheic closing, Avalonia docking against the Gondwana promontory of Iberia and Armorica, the Acadian event that built Welsh mountains.
Flow, cleavage, barrels, heartbeat
Directed pressure ran north-west to south-east; folds grew as large as cliffs. Slow force makes rock flow (Silly Putty, glacier ice); fast force makes earthquakes. Shortening of about 10 percent up to a third, heat to about 250°C, and loss of up to a fifth of the rock's mass as silica produced slate. The disc follows cleavage, not bedding; craftsmen still split roof tiles under a quarter of an inch. Micas rotate and regrow perpendicular to stress; crenulation microfolds appear; a few percent of fluid remains. Pathways for the lost silica are mysterious; mineral springs may have fertilised new land plants.
Paperback to tome, and a catalytic halo
Mica stacks fattened about tenfold via fluid, from paperback to tome. Micas trapped inside monazite stayed flat, ball-bearings in plasticine, proving later growth outside. Graptolites shattered into segments; mica pressure shadows grew even on carbon films, so organic matter catalysed mica while other objects grew quartz. Burgess Shale mica jackets are likely late, not early seals.
Acadian timing and lost silica
Graptolites had already gone through boom and bust; fish thrived; land was greening; the sea had filled to a coastal plain. The Iapetus, more than 1000 km wide 50 million years earlier, had closed about 20 million years before with oblique docking and little deformation. The southern Rheic closure against Iberia and Armorica coincides with Welsh mountain building. Slate lost up to a fifth of its mass as silica; pathways remain mysterious; mineral springs may have fertilised new land plants. Roofing craftsmen still exploit cleavage thinner than a quarter of an inch. The 20 million year gulf between oil-window dates (415-420 Ma) and cleavage argon (about 396 Ma) is the quiet interval the heartbeat later interrupts.
Dock (Rheic closure), flow (slow stress), cleave (slate fabric), heartbeat (rhythmic argon zones).
Name one organisational change that needs slow squeeze (strangler migration) and one that is an earthquake (big-bang cutover).
A mineral jacket around a fossil can be late decoration. In audits and ML interpretability, a neat halo is not proof of original cause.
12. Breaking the Surface
Quartz veins, ice and waves return the pebble to light. Ten clocks, from mantle escape to nuclear fallout, close the forensic case, and the fallout marker frames the Anthropocene debate.
The metal factory
Wales is a country-sized metal factory, about 100 km long, 60 km across and 6 km deep, that produced copper, silver, zinc, lead and gold, mined since at least the Bronze Age, more than 3000 years ago, with antler picks and cobbles. (Hayao Miyazaki honoured Welsh miners in Castle in the Sky.)
The pebble's white vein cuts both the strata and the cleavage, so it is younger than both: the principle of cross-cutting relationships. It is quartz: glassy, does not fizz in acid, not scratched by steel. As erosion unloaded the crust, it rose (isostasy, like an iceberg) and cooled below about 100°C, becoming brittle enough to crack. Hot fluids carried metals as chlorides and dropped them as sulphides where they met hydrogen sulphide: pyrite, chalcopyrite, galena, sphalerite, sometimes gold.
Tiny bubbles of fluid trapped in vein quartz are sealed samples of the water that flowed through the crust. Warm them on a microscope heating stage until the vapour bubble disappears, and you have the trapping temperature; ablate them with a laser into a mass spectrometer and you have the brine's composition. Some quartz crystals are double-terminated, grown while suspended in fluid rising at up to a metre a second: the plumbing was valve-like and explosive, like a geyser such as Old Faithful.
Unrecorded history
Earthquakes passed through the pebble with almost no record, "very low fidelity," unlike the 78 rpm record, the tape and the seismograph, which are new kinds of fossil. About 50 million years after the squeeze, another tectonic vice gripped Wales; later the Alps rose and the Atlantic opened, and the Irish Sea basin sank over 3 km. The impact in Mexico about 65 million years ago, suspected of killing the dinosaurs, would have shaken the Earth like a bell, equivalent to about 13 on the Richter scale even 1000 km away.
The book's "65 million years" has been refined. In 2013 Paul Renne and colleagues (Science) used recalibrated argon-argon dating to put the Chicxulub impact at 66.04 million years ago, with the impact and the end-Cretaceous extinction synchronous within about 32,000 years.
Above the buried pebble, the surface world went by unseen: armoured fish, Carboniferous coal swamps, Permian and Triassic deserts, the end-Permian extinction about 250 million years ago, Jurassic seas.
Ice and the sea
Unloading opened joints; rain-derived water and then microbes returned. Ice came late: to Antarctica over 30 million years ago, to the north less than 3 million years ago, when the Laurentide, Greenland and Scandinavian ice sheets grew (the Laurentide perhaps by the "snowgun" effect of moist Pacific air over cold North America). Hundreds of metres of ice pressed down the land. Where little ice grew but it was bitterly cold, the ground froze iron-hard: permafrost penetrated tens and then hundreds of metres, as in Siberia and Canada today, with only the top metre or so (the active layer) thawing each summer. The British countryside still shows its traces in disturbed soils, circular hollows where ice blisters grew, and frost-crack patterns. The pebble stratum was almost certainly frozen many times in the last million years or so. Because permafrost forms and thaws over thousands of years, the underground temperature lagged far behind the climate above, a "slow-moving weather underground" that geologists now exploit: temperature profiles measured in boreholes reconstruct past climate from cold and warm waves still passing through the ground. Water expands as it freezes (which is why ice floats and oceans stay habitable), and frost pre-fractured the slab that would become the pebble.
About 20,000 years ago sea level was over 100 m lower; it stabilised about 5000 years ago, and waves reached the cliffs. The excavation, perhaps only a few centuries ago, was "brutally efficient." Breaking waves drove air into fractures already opened by long-vanished ice; compressed air with tons of storm-driven water behind it acted like a chisel, aided by boulders hurled at the cliff. A slab broke free after a little over 420 million years; storm waves smashed it against the cliff into sharp shards, its face a cleavage plane from about 396 million years ago, and longshore drift, with the swash and backwash of the shingle, rounded it through thousands of collisions into the disc that waits to be picked up.
The ten chronometers Zalasiewicz finds in one pebble, from the mantle to the Trinity test.
Ten clocks
The final tally adds four surface clocks to six deep ones. Cosmogenic dating: cosmic rays break silicon into radioactive aluminium and oxygen into radioactive beryllium, measurable after a few millennia of exposure (a "baby technique," then not much more than two decades old). Optically stimulated luminescence (OSL): grains on the pebble's underside, kept dark, store a signal that light releases. Lichenometry: lichen grows about a millimetre a year. And the newest clock, "one of our very own making," started at 5:29:45 am on 16 July 1945 at Alamogordo, New Mexico, with the first atomic bomb. With Hiroshima and Nagasaki (some 220,000 killed), atmospheric tests and later Chernobyl, artificial radionuclides including plutonium and a long-lived caesium isotope coat everything exposed since, including the pebble.
The book never uses the word, but its tenth clock is exactly the marker proposed for a new epoch. Zalasiewicz was the founding chair of the Anthropocene Working Group, convened by the ICS in 2009, which proposed an Anthropocene Epoch with a Crawfordian Age beginning in 1952, marked by the rise in plutonium in varved sediments of Crawford Lake, Ontario. In March 2024 the ICS Subcommission on Quaternary Stratigraphy rejected it (reported as 12 against, 4 for, 2 abstentions, short of the 60 percent needed); the IUGS and ICS approved the decision and the working group was dissolved. Officially we remain in the Meghalayan Age of the Holocene. A September 2026 comment in Nature Reviews Earth & Environment restates the live distinction: human-driven Earth-system change is not in doubt, but the mid-20th-century boundary failed the correlatability tests required for a formal epoch. The IUGS-ICS statement still calls the term "an invaluable descriptor of human impact on the Earth system."
Wales as a metal factory
Wales ran as a metal factory about 100 × 60 × 6 km: copper, silver, zinc, lead, gold, mined since the Bronze Age for over 3000 years (antler and cobble tools; Miyazaki's Castle in the Sky is a cousin image). White quartz veins cut both bedding and cleavage, so they are younger by cross-cutting logic. Quartz is glassy, does not fizz in acid, and steel will not scratch it. Cooling toward 100°C made the rock brittle; isostasy (the iceberg principle) lifted the pile. Fluids dissolved metals as chlorides and precipitated sulphides with hydrogen sulphide, sulphur perhaps from organic laminae. Reflected light shows pyrite, chalcopyrite, galena, sphalerite and gold. Fluid inclusions, heated until the bubble vanishes, give homogenisation temperature; laser ablation plus mass spectrometry reads brine chemistry. Double-terminated quartz grew suspended in upward flow as fast as 1 m/s, released in valve-like pulses like Old Faithful.
What the pebble missed
Earthquakes leave almost no rock memory (a stadium roar that fades). About 50 million years later came another squeeze (Variscan); about 200 million years later the Alps rose and the Atlantic opened; the Irish Sea basin subsided more than 3 km. Chicxulub at 66 Ma would have shaken like about magnitude 13 at 1000 km. Surface history passed unseen: armoured fish, Carboniferous coal swamps about 50 million years after burial, Permian and Triassic deserts, the end-Permian extinction about 250 Ma, Jurassic seas, the red dragon of later myth.
Ice, waves, ten clocks
Unloading joints opened; rain-derived fluids returned; microbes came back. Antarctic ice is over 30 million years old; northern sheets under 3 million. The Laurentide "snowgun," ice hundreds of metres thick, and permafrost to hundreds of metres (active layer about 1 m) pressed and chilled the crust. Boreholes still read that climate. Ice expands on freezing (the anomaly that lets ice float and oceans stay habitable). Frost pre-fractured the cleavage. 20,000 years ago sea level was more than 100 m lower and hills were pressed down tens of metres; about 5000 years ago sea level stabilised and waves reached the cliffs. Wave-compressed air chiselled slabs free; longshore drift rounded the disc. Humans had been present for over 100,000 years before that freeing.
Surface clocks join the deep ones: cosmogenic ²⁶Al from silicon and ¹⁰Be from oxygen (techniques about two decades old when the book was written), optically stimulated luminescence on the underside, lichenometry (about 1 mm/year), and the nuclear marker from Trinity at 5:29:45 a.m. on 16 July 1945 at Alamogordo, followed by Hiroshima, Nagasaki (about 220,000 dead), atmospheric tests and Chernobyl. Plutonium and caesium cling to the pebble; lichens scavenge them. That is ten clocks.
Castle in the Sky, Old Faithful, Trinity
Ice expanded on freezing, floated, and kept oceans habitable. Frost pre-fractured cleavage; waves chiselled the disc free only centuries ago. Ten clocks close the case: Nd model age, zircon U-Pb, Re-Os, monazite U-Pb, Rb-Sr, mica argon, cosmogenic nuclides, OSL, lichenometry, and the 1945 nuclear marker. Detectability is not governance: the Anthropocene proposal built on that tenth clock failed the March 2024 vote and remains a descriptor, not an epoch.
Surface clocks in brief
Cosmogenic ²⁶Al (from silicon) and ¹⁰Be (from oxygen) date exposure. Optically stimulated luminescence dates the last time sand grains under the pebble saw light. Lichenometry advances at about 1 mm/year on suitable surfaces. The nuclear marker begins at Trinity, 5:29:45 a.m. on 16 July 1945 at Alamogordo, then Hiroshima and Nagasaki (about 220,000 dead), atmospheric tests and Chernobyl. Plutonium and caesium cling to pebbles; lichens scavenge them. Together with the six deep clocks, that is ten. Cross-cutting relationships still date the veins without any of those instruments: whatever cuts cleavage is younger than about 396 Ma.
Whatever cuts is younger. Then tally the ten clocks so no single date carries the story alone.
In a codebase or process map, draw three layers and two shortcuts that bypass a layer. Date the shortcuts as younger.
Do not confuse detectability with governance agreement. The Anthropocene vote shows how a clear signal can still fail formal ratification.
Zalasiewicz chaired the Anthropocene Working Group (from 2009), which proposed a Crawfordian start in 1952 at Crawford Lake, Ontario, marked by plutonium. In March 2024 the SQS rejected the proposal (reported 12-4-2, short of 60 percent); IUGS/ICS approved the rejection and dissolved the group. We remain in the Meghalayan Age of the Holocene. A September 2026 Nature Reviews Earth & Environment comment restates: planetary transformation is real; the mid-20th-century boundary failed chronostratigraphic correlatability tests. The term stays an invaluable descriptor, not a formal epoch.
Renne and colleagues (2013, Science) put the impact at 66.04 Ma, synchronous with the end-Cretaceous extinction within about 32,000 years.
13. Futures
The beach abrades the pebble toward a diaspora of grains, while the far future belongs to a brighter Sun and a frozen core. The lasting lesson is how to weigh converging evidence against single fragile links.
The diaspora
On an exposed coast a pebble can lose between a third and a half of its mass in a season. A sea-level rise of a few metres over coming centuries would drown the beach; a later retreat might come in some 100,000 years. Then the pebble's components disperse:
| Component | Fate |
|---|---|
| Quartz | Clusters break to grains, to be recycled again |
| Zircon, rutile, tourmaline | Recycled a second, third or fourth time; concentrated by currents like Klondike gold |
| Micas | Back to mud |
| Carbon (now graphite) | Mostly indigestible; some re-enters the carbon cycle |
| Pyrite | Oxidises to rust; sulphur to gypsum (selenite), the ocean, or as dimethyl sulphide aerosols that seed clouds (the book says "dimethyl sulphide," a slip) |
| Monazite | Into Welsh stream sands, a reworking never studied |
New strata will hold, in Darwin's words, "endless forms most beautiful and most wonderful."
The far future
The cycles will continue for perhaps another billion years. The Sun brightens; the oceans are lost; plate tectonics, which needs water, stops; the core freezes and the magnetic field dies. In about five billion years the Sun becomes a red giant, and it is "touch and go" whether the Earth is engulfed; the Sun then collapses to a white dwarf not much bigger than the Earth. There will be no supernova, only perhaps a planetary nebula like the Cat's Eye, and the pebble's atoms may join a new star system.
Zalasiewicz shows the power of multiple independent lines of evidence that converge. But many links in the narrative rest on a single study (the 396.1 million-year argon date, the monazite ages), on analogy (Black Sea, Santa Barbara), or on hypotheses he flags as untested (the thermostat, the clathrate episode). The book is honest about this; readers should keep his hedges in view when retelling the story.
Diaspora on the beach
On the beach the pebble loses under 0.1 percent of its weight per tide and may shed a third to half its mass in a season on an exposed coast. Sea-level rise of a few metres over centuries would drown the beach; in about 100,000 years ice cycles would pull the sea back. Grains diaspora: quartz as clusters; zircons, rutile and tourmaline on their second to fourth recycling; heavy-mineral concentrates like a tiny Klondike; micas returning to mud. Carbon as graphite is mostly indigestible; some may ride into mantle diamond or into space. Pyrite rusts; sulphate may become selenite or gypsum and, via marine microbes, dimethyl sulphide aerosols that seed clouds (the book prints "dimethyl sulphate"; the cloud-seeding molecule is dimethyl sulphide). Monazite sands in Welsh streams remain little studied. Darwin's "endless forms most beautiful" still applies to the afterlife of atoms.
Deep future and the method's edge
Further out, the book sketches 1-2 billion years more of familiar cycles; then a brighter Sun, lost oceans, a frozen core, stalled plates without water, perpetual night; a red-giant Sun about 5 billion years from now that may engulf Earth; a white dwarf about Earth-sized; no solar supernova; planetary nebulae such as the Cat's Eye; atoms that may yet join another star system. The sober close is methodological: converging clocks are strong; single studies and analogies are weaker links. The book was published by Oxford University Press in 2010. Jan Zalasiewicz, at Leicester, later chaired the Anthropocene Working Group whose proposal Chapter 12's tenth clock foreshadows.
Endless forms, after the beach
Quartz travels as clusters; zircons may be on their second to fourth cycle; heavy minerals concentrate like a tiny Klondike; micas return to mud; monazite sands in Welsh streams await study. Graphitic carbon is mostly indigestible to the next biosphere; some may become mantle diamond or leave for space. Pyrite's sulphate, via marine microbes, can become dimethyl sulphide and seed clouds. Darwin's phrase about endless forms fits the atomic diaspora as well as living lineages.
The far-future sketch (brighter Sun, lost oceans, frozen core, stalled plates, red giant, white dwarf, possible new star system) is scaffolding for values, not a schedule. The book's lasting claim is narrower and stronger: stories constrained by converging evidence outrank stories that cannot say what would falsify them. Oxford University Press published the book in 2010; its method outlives any single date inside it.
Abrasion rates and the white dwarf
Under 0.1 percent mass loss per tide scales to a third or half of the pebble in a season on an exposed coast. Sea-level rise of a few metres would drown the beach for a while; ice cycles on about 100,000-year periods would pull the sea back. After oceans are lost under a brighter Sun, plate tectonics stalls without water; the core freezes; the field dies; night becomes perpetual at the surface. A red-giant Sun about 5 billion years from now may engulf Earth; the remnant white dwarf is about Earth-sized and will not go supernova. Planetary nebulae such as the Cat's Eye show how expelled shells look from outside. Pebble atoms may yet join another star system. None of that is scheduled fact; all of it is scaffolding around the forensic habit the book trains.
What the method allows the future to claim
Abrasion under 0.1 percent per tide, seasonal mass loss of a third to half on exposed coasts, and ice-driven sea-level cycles on about 100,000-year periods are near-term claims with analogues. The brighter Sun, lost oceans, frozen core and stalled plates are longer sketches. A red-giant Sun about 5 billion years from now, a white dwarf remnant, and atoms that may join another star system are scaffolding for values, not schedules with error bars.
The book's lasting claim is narrower and stronger than any far-future sentence: stories constrained by converging evidence outrank stories that cannot say what would falsify them. Oxford University Press published the book in 2010; Jan Zalasiewicz's later work on the Anthropocene Working Group shows the same forensic habit applied to a contested present. Keep the hedges when you retell the pebble's end.
Tide abrades, grains recycle, sulphur may seed cloud, the story cools into testable claims about the far future.
List claims that rest on (a) converging clocks, (b) a single dated study, (c) modern analogy. Circle the weakest link you still repeat. Replace it with the hedged sentence Zalasiewicz would allow.
Use deep-future sketches to stress-test values, not as schedules. The book's strength is the forensic present, not the prophetic end.
The pebble's stories hold because independent lines constrain each other. Enlarge that habit and ordinary objects become readable without becoming mythical.
Synthesis: Reading Systems Like a Pebble
Zalasiewicz's pebble is a lesson in forensic reasoning under uncertainty: reconstructing a long, partly erased history from fragments, with independent clocks that check each other. This synthesis translates the book's methods into those three settings.
- Inventory before interpretation: count the atoms before telling the story (Chapter 1).
- Independent clocks: trust an age only when two systems agree, as with zircon's two uranium chains (Chapter 3).
- Know what reset the clock: each date records a specific event (Chapters 10 and 11).
- Absence is evidence only if presence would have been preserved (Chapters 5 to 7).
- Cross-cutting relationships: whatever cuts something is younger than it (Chapter 12).
Why these lessons transfer
The pebble problem has three features that recur in modern institutions. First, the record is incomplete and biased: most plankton left no trace, and most of the turbidity currents' timing is unrecorded. Second, later processes overprint earlier ones: diagenesis moved the rare earths, mountain-building reset some clocks and not others. Third, independent methods disagree, and the disagreement is informative: the monazite's two clocks differed by over 100 million years, and that difference revealed inherited neodymium. A good analyst treats discordance as a clue rather than a nuisance. The book's deepest lesson is that the disagreement between two measurements often tells you more than either one.
For AI
Provenance is the zircon signature of data. Just as Avalonian zircon ages fingerprint where grains came from, a model's behaviour carries the signature of its training data. Data lineage, dataset cards and content provenance are the AI equivalent of heavy-mineral separation: the rare components often tell you most.
Concordance as validation. A zircon age is trusted when the 238U and 235U clocks agree. AI evaluation should demand the same: offline benchmarks, human review and production telemetry should converge before a claim is accepted. A model that scores well on one measure and poorly on another is "discordant": something has leaked or been lost.
Diagenesis, or model drift. Monazite showed that rare earth elements, assumed immobile, had moved over centimetres, misleading provenance studies for decades. Features assumed stable (customer behaviour, document formats, vocabulary) also migrate after deployment. Monitor the assumptions, not just the outputs. The half-graben story adds a second warning: a model fitted along one easily observed transect (the cliffs) can be "right, but wrong" when the data across the basin are finally collected. Evaluation sets drawn from the most accessible data reproduce exactly that bias.
Ghosts in absentia. Most Silurian plankton left no trace. Training data likewise over-represents what is easy to record and under-represents the soft-bodied majority. Ask what your data could never have captured before claiming a model "knows" a domain.
The pebble has ten clocks because each event left its own record. For a high-stakes AI decision, keep the equivalent: data snapshot version, feature pipeline version, model version, prompt or configuration version, policy rules version, human-override log, and timestamps for each. When an outcome is disputed, you need to know which clock reset when.
For banking
Residence time and liquidity. Sodium stays in the sea for about 70 million years; aluminium for about 200. Bank balance sheets have the same spread: core deposits that behave like sodium, wholesale funding that drops out like aluminium. Liquidity risk management is, in effect, estimating residence times and knowing which can change suddenly.
The carbon-burial thermostat and credit cycles. The book's thermostat is a negative feedback that may also fail. Credit cycles have similar loops: rising defaults tighten lending, which slows growth, which raises defaults. Stress testing should look for the conditions under which a stabilising feedback flips into a destabilising one, as the Silurian ocean flipped between anoxic and oxygenated states.
Dead zones as concentration risk. Modern dead zones come from too much nutrient in one place. Portfolio concentration works the same way: an excess of exposure to one sector or counterparty can strip a balance sheet of resilience even when each loan looks sound.
Independent clocks for reconciliation. Payment and ledger reconciliation is concordia in practice: two independent records of the same event must agree, and a discordant pair points to lost or added value, like lead lost from a zircon. The book's insistence on a built-in cross-check is the core of control design.
The 2024 Anthropocene decision is a model of formal governance: a working group proposal, a defined 60 per cent supermajority, a subcommission vote, ratification by the parent bodies, and a clear statement of what the decision does and does not mean (no formal epoch; the concept stays useful). Both have value; confusing them causes trouble.
Applying the pebble's discipline without stealing its metaphors
The honest transfers are mechanistic. Concordant twin clocks map to independent model metrics and reconciliations. Cross-cutting maps to dependency graphs and override logs. Event-versus-duration maps to incident systems that must store waits, not only spikes. Immobile-until-fluid maps to feature stores and risk factors that migrate when the chemistry of the business changes. Slow squeeze versus earthquake maps to strangler migrations versus big-bang cutovers. Descriptor versus standard maps to useful internal language that has not earned regulatory ratification.
Dishonest transfers are prophetic or flattering: calling a product roadmap a "deep time strategy," or treating pretrained weights as sacred LLVPs beyond audit. Keep the hedges. Single-lab ages, single papers and living analogues are weaker links than converging clocks. People, unlike rock, adapt to the measures you apply; geological discipline transfers better than geological metaphors about behaviour.
Main Points
- Inventory before interpretation: count the atoms, fields or positions before telling the story (Ch 1).
- Independent clocks beat single dates: trust zircon, monazite or models only when two systems agree (Ch 3, 10).
- Know what reset the clock: each date records a specific event, not "age" in the abstract (Ch 10, 11).
- Absence is evidence only if presence would have been preserved (Ch 5-7).
- Cross-cutting relationships date structure: whatever cuts something is younger (Ch 12).
- Discordance is a clue, not only an error: inherited Nd and excess argon both taught something (Ch 10, 11).
- Event records need waiting-time records: turbidite stripes without gaps invent false tempos (Ch 4).
- Local faults beat textbook fans: half-grabens make "right but wrong" models (Ch 8).
- Tracers can move: rare earths left their "immobile" reputation in the oil window (Ch 10).
- Rate selects rheology: slow squeeze flows; fast stress breaks (Ch 11).
- Provenance is a signature, from Avalonian zircons to training data (Ch 3).
- Negative feedbacks can flip: anoxia-ice-carbon thermostat and credit cycles alike (Ch 5).
- Soft majorities leave hard minorities: graptolites may be under 1 percent of plankton yet dominate the archive (Ch 7).
- Descriptors are not standards: Anthropocene remains useful after formal rejection (Ch 12).
- Ten clocks beat one narrative: audit trails should match the pebble's redundancy (Ch 12).
- Analogy is a tool, not a proof: Black Sea and Santa Barbara illuminate; they do not certify (Ch 5, 13).
- Catalysis can mimic early sealing: late micas on fossils warn interpretability teams (Ch 11).
- Residence times organise reservoirs: sodium versus aluminium; core deposits versus wholesale funds (Ch 3).
- Field mapping still wins: walking the sea floor found the trapdoor the model missed (Ch 8).
- Far futures are scaffolds: use them to test values, not to schedule destiny (Ch 13).
- Stories must stay revisable: the prologue's rule for geology is the rule for models and balance sheets.
- Compositionally distinct deep stores (LLVPs, ledgers, pretrained weights) shape the surface long after their origin event (Ch 2).
- Index fossils are shared protocols; reworked ones are poisoned packets (Ch 6).
- Depletion conveyors fill sinks until the neighbourhood is empty (Ch 9).
- Governance needs supermajorities for irreversible labels; descriptors can stay useful at lower consensus (Ch 12).
Frameworks and mental models
| Name | What it says | Use it when |
|---|---|---|
| C-S-I-A-C archive | Composition, Structure, Inclusions, Alteration, Clocks | Reading any layered artefact |
| Concordant twin clocks | Trust a date only if two independent systems agree | Validation, reconciliation, model risk |
| Right-but-wrong map | A general model fails under a local fault | Reference architectures, basin or market structure |
| Event vs duration | Spikes are recorded; waits are not | Incident logs, turbidites, transaction journals |
| Immobile-until-fluid | Conservative tracers migrate when chemistry changes | Feature stores, REE provenance, risk factors |
| Slow squeeze / earthquake | Rate selects whether a system flows or fractures | Migrations, restructures, stress events |
| Descriptor vs standard | Useful language need not be ratified law | Anthropocene, internal "gold" policies |
| Weakest-link weighing | Converging lines strong; single studies fragile | Any long causal narrative |
The forensic method is the achievement. Individual links still vary: some ages are single-lab results, some mechanisms are analogies, some far-future sketches are scaffolding. Keep Zalasiewicz's hedges when you retell the story. Rock also lacks incentives; people adapt to the measures you apply, so geological discipline transfers better than geological metaphors about behaviour.
Rock does not have incentives, and markets and organisations do. The pebble's processes are indifferent; people adapt to the measures applied to them. Use the geological methods (independent evidence, provenance, cross-cutting logic) as discipline, not as claims that firms behave like sediment.
- Demand concordance: trust results only when independent measures agree.
- Track provenance and ask what reset the clock for every recorded event.
- Treat absence carefully, and map local faults before applying general models.
- Separate a useful descriptor from a ratified standard.
Glossary
| Term | Meaning |
|---|---|
| Acritarch | Organic-walled microfossil of uncertain origin, mostly algal cysts |
| Anoxia | Absence of dissolved oxygen in water or sediment |
| Apatite | Calcium phosphate mineral; cements the pebble's dark stripes |
| Avalonia | Microcontinent born as an island arc about 1 billion years ago; includes Wales, England and Newfoundland |
| Cable bacteria | Filamentous bacteria that conduct electrons through sediment (discovered 2012) |
| Chitinozoan | Flask-shaped organic microfossil, probably the egg case of an unknown animal |
| Chondrule | Melt droplet from the early Solar System, flash-heated to about 1500°C |
| Clathrate | Methane trapped in an ice lattice, stable only at certain pressures and temperatures |
| Cleavage (slaty) | Closely spaced planes along which slate splits, formed by tectonic squeezing |
| Concordant | A zircon whose two uranium-lead clocks give the same age |
| Cosmogenic dating | Dating exposure of a rock surface from isotopes made by cosmic rays |
| Cross-cutting relationship | Principle that a feature cutting another is younger |
| Diagenesis | Changes to sediment after burial, at low temperature and pressure |
| Flute cast | Scour mark on the base of a bed that shows current direction |
| Framboid | Raspberry-like cluster of pyrite microcrystals |
| Graptolite | Extinct colonial pterobranch with collagen tubes; key Ordovician and Silurian time fossil |
| Half-graben | Basin bounded by a fault on one side and a hinge on the other |
| Term | Meaning |
|---|---|
| Hemipelagite | Fine mud that settles slowly from suspension between turbidity currents |
| Iapetus and Rheic oceans | Oceans north and south of Avalonia that closed in the Palaeozoic |
| Isochron | Line through several samples on an isotope plot whose slope gives an age |
| Isorenieratane | Pigment fossil of a photosynthetic, oxygen-intolerant sulphur bacterium |
| Large low-velocity provinces | Two continent-sized dense regions at the base of the mantle |
| Monazite | Rare earth phosphate; grown in the pebble during burial |
| OSL | Optically stimulated luminescence: dates when grains were last exposed to light |
| Palaeomagnetism | Record of the ancient magnetic field preserved in rocks |
| Pyrite | Iron sulphide, FeS2, "fool's gold" |
| Residence time | Average time a dissolved element spends in the sea |
| Sm-Nd model age | Age at which crustal material separated from the mantle |
| Tempestite | Storm-deposited sediment layer |
| Theia | Mars-sized planet whose collision with Earth formed the Moon |
| Turbidite | Layer deposited by a turbidity current, a dense underwater sediment flow |
| Uniformitarianism | Principle that present processes, given time, explain the past |
| Varve | Annual layer pair in lake or basin sediment |
| Zircon | Durable mineral that hosts uranium and excludes lead; the premier geological clock |
Who's Who
The book is full of people, past and present, who supplied a method or a metaphor. In order of appearance:
Amedeo Avogadro, Count of Quaregna and Cerreto, gave the constant that lets Zalasiewicz count the pebble's 10^24 atoms. William Blake, who saw "a world in a grain of sand," is the book's poetic patron. Charles Lyell coined uniformitarianism and danced at Darwin's coral-atoll theory; James Hutton saw "no vestige of a beginning, no prospect of an end." Henry Clifton Sorby of Sheffield invented the thin section and was mocked for studying mountains with a microscope. Henri Becquerel discovered radioactivity in 1896, the foundation of every clock in the book.
Pliny the Elder, Seleucus and the Venerable Bede each glimpsed how tides work; the Marquis de Laplace solved the "spiniest problem." Xenophanes guessed that winds needed "the great sea"; Evangelista Torricelli found their true cause in air density. John Marr saw that the Silurian sea had been "poisoned below the 100 fathom line," and Rachel Carson named marine snow the ocean's "eternal snowfall."
Alfred Eisenack discovered chitinozoans between two spells as a prisoner of war. Charles Lapworth, a Galashiels schoolteacher, turned graptolites into a clock; Joachim Barrande of Bohemia proposed the rival idea of "colonies." Sir Alister Hardy wrote The Open Sea, which showed the author how much plankton leaves no trace. William Gilbert realised in 1600 that the Earth is a magnet. Robert Brown described the motion that bears his name, and Albert Einstein (whom the book calls "Alfred") helped explain it. Hayao Miyazaki celebrated Welsh miners in film, Jules Verne imagined the underground world the veins disprove, and Charles Darwin supplies the closing words, "endless forms most beautiful."
Jan Zalasiewicz himself is a field geologist, palaeontologist and stratigrapher, based at the University of Leicester after working at the British Geological Survey, author of The Earth After Us (2008) and founding chair of the Anthropocene Working Group.
About this book
The Planet in a Pebble: A Journey into Earth's Deep History by Jan Zalasiewicz, 2010. This is a chapter-by-chapter study of the book, with summaries, maps, tables and diagrams; short quotations keep their page references. It is shared free for personal study, is not for sale, and is not endorsed by the author or publisher. Please read and buy the original.
Insight Editions · rajeshrmahapatra.com/insight-editions