The geologic time scale
The International Commission on Stratigraphy’s chart, in full: the fifty units that divide 4.54 billion years, from eon to age, each with its bounds and what those bounds are worth. Because a boundary is not a round number. Some are fixed by a golden spike, a marker driven into a layer of rock at one precise place on the globe; others by a dating that a better measurement will move. Charts publish the figures without saying which is which. What happened inside each unit is on its own page: this one gives the frame.
50 units, across five ranks
A period boundary is a decision before it is a date. Where the two disagree, this entry says which is which.
Hadean
4.54 Ga → 4.03 GaEon
Five hundred million years without a single rock left to show for them. The eon takes its name from the underworld, and it is the only division of Earth's history defined by the absence of what should record it.
Its older bound is the age of Earth's accretion, measured on meteorites rather than on Earth, because no terrestrial material is that old.
See it on the timelineArchean
4.03 Ga → 2.5 GaEon
A billion and a half years in which the planet acquires everything that matters, and almost none of it leaves a mark: continents, oceans, a magnetic field, and life.
Neither bound was placed at an event. The younger one, 2.5 billion years, is a round number, agreed because nothing in the record singles out the passage.
See it on the timelineEoarchean
4.03 Ga → 3.6 GaEra
The dawn era, and the first four hundred million years for which any rock survives at all. What it holds is thin, contested, and among the most important evidence there is.
An era whose beginning is set by what has survived rather than by what happened, and whose end, 3.6 billion years, is a round number with no event attached.
See it on the timelinePaleoarchean
3.6 Ga → 3.2 GaEra
Four hundred million years in which life stops being a contested trace and becomes a structure you can walk up to and touch.
The era's bounds, 3.6 and 3.2 billion years, are two of the round numbers that cut the Archean into four roughly equal parts.
See it on the timelineMesoarchean
3.2 Ga → 2.8 GaEra
Somewhere in this era the crust may start behaving as it does today, one plate sliding beneath another and driving the machinery that builds continents.
Nothing in the rock marks either end of this era; 3.2 and 2.8 billion years are conveniences. What the era is best known for, the onset of plate tectonics.
See it on the timelineNeoarchean
2.8 Ga → 2.5 GaEra
Three hundred million years of oxygen made and swallowed at once: cyanobacteria are already splitting water, and the ocean takes back everything they release.
A long gap separates the invention from its consequence, and neither end of that gap is sharply dated. What earns the era a place here is a process that began.
See it on the timeline
Proterozoic
2.5 Ga → 539 MaEon
Two billion years, the longest eon of all, and the one in which oxygen enters the air, the cell becomes complex, and the planet freezes over more than once.
Its younger bound, 538.8 million years, is one of the best defined on the chart: a golden spike in Newfoundland at the first appearance of a particular burrow.
See it on the timelinePaleoproterozoic
2.5 Ga → 1.6 GaEra
Nine hundred million years, and the most violent change the planet has undergone: the air fills with oxygen, and almost everything alive is poisoned by it.
Both of the era's bounds were set at round figures, and the first falls a few tens of millions of years before the event that gives the era its interest.
See it on the timelineSiderian
2.5 Ga → 2.3 GaPeriod
The period in which the air acquires oxygen for good, and the planet, stripped of its methane blanket, freezes over.
Nothing in the rock marks either bound: 2,500 and 2,300 million years are round numbers, fixed by agreement. The oxygenation they are meant to frame is a process.
See it on the timelineRhyacian
2.3 Ga → 2.05 GaPeriod
Oxygen overflows, then withdraws; magma pours into the crust in quantities never seen again, and something large and flat may already be living.
The two bounds are conventional round numbers. What the period contains is dated far better than what defines it: the Bushveld intrusion to within five million years.
See it on the timelineOrosirian
2.05 Ga → 1.8 GaPeriod
Mountains rise on every continent, the first supercontinent that can be mapped takes shape, and a cell swallows a bacterium and keeps it.
The bounds are conventional. The events the period is known for are dated by their rocks and not by it, and the most consequential of them.
See it on the timelineStatherian
1.8 Ga → 1.6 GaPeriod
The period that gives the planet its long calm: continents stabilise, oxygen settles low, and the deep sea turns to sulfide for a billion years.
Two conventional round numbers bound it, and they mark nothing. The boring billion they open is a description after the fact, with no agreed beginning or end.
See it on the timeline
Mesoproterozoic
1.6 Ga → 1 GaEra
Six hundred million years in which, as far as anyone can tell, remarkably little happens. They sit at the heart of what the literature calls the boring billion, and the nickname is a confession.
An era named for what it seems to lack, and the name has outlived the evidence for it. What looks like stasis may be the resolution of the record rather.
See it on the timelineCalymmian
1.6 Ga → 1.4 GaPeriod
Life becomes big enough to see with the naked eye, under an air that carries at most a thousandth of today's oxygen.
Neither bound marks an event: both come from a decision. The Gaoyuzhuang fossils are firmly dated but not firmly identified: their size and the regularity.
See it on the timelineEctasian
1.4 Ga → 1.2 GaPeriod
Two hundred million years with almost no landmarks, and the stretch of the chart where sex, arguably the most consequential of inventions, is usually placed.
Behind its two round numbers there is nothing at all. The origin of sex has no date either: it is placed here from the spread of meiotic genes among living lineages.
See it on the timelineStenian
1.2 Ga → 1 GaPeriod
Narrow belts of crushed rock cross every continent: the mountains that welded Rodinia, and the first organism we can give a modern name.
The period's limits were fixed on the chart, not read in the rock; what is measured here belongs to the mountains, the Grenvillian collision falling between 1.09.
See it on the timeline
Neoproterozoic
1 Ga → 539 MaEra
Four hundred and sixty million years in which ice reaches the tropics more than once, a supercontinent breaks apart, and the first animals appear.
Its end is a golden spike and one of the best defined boundaries on the chart; its beginning is the round one billion years.
See it on the timelineTonian
1 Ga → 720 MaPeriod
Rodinia comes apart, oxygen begins its second rise, and cells start eating other cells: the Neoproterozoic opens.
The base is a conventional round number; the top, by contrast, is tied to the onset of the Sturtian glaciation and moves as that dating improves.
See it on the timelineCryogenian
720 Ma → 635 MaPeriod
Twice, ice reached sea level at every latitude, tropics included; between the two, algae took the open ocean from bacteria.
This is one of the few Precambrian periods whose base was set to catch an event, the onset of the Sturtian glaciation, rather than fixed at an arbitrary round figure.
See it on the timelineEdiacaran
635 Ma → 539 MaPeriod
The first new period on the chart in a hundred and twenty-five years, and the one in which large, soft, unarmed bodies appear and then vanish.
Its base is a defined section and its top the base of the Cambrian at 538.8 million years, so both bounds are firm. What happens between them is much less secure.
See it on the timeline
Phanerozoic
539 Ma → 2026Eon
The eon of visible life, and the last twelve per cent of Earth's history. Everything a museum can display happens here, which is why the timeline is so crowded at this end and so bare everywhere else.
Its older bound is not in doubt; its name is. It says that life becomes visible here, when what becomes common is hard parts, and the eons before were full of life.
See it on the timelinePaleozoic
539 Ma → 252 MaEra
Two hundred and eighty-seven million years, from the first shelled animals to the worst extinction in the record. Life leaves the water in this era and never fully returns.
Both of its bounds are golden spikes, boundaries pinned to a chosen outcrop rather than to a round number, which makes this era one of the best bracketed on the chart.
See it on the timelineCambrian
539 Ma → 485 MaPeriod
Nearly every way of building an animal appears within a few tens of millions of years, and the radiations that followed explored those designs rather than adding new ones.
A golden spike in Newfoundland at 538.8 million years, placed on the first appearance of a burrow rather than of a body, makes its base one of the best defined.
See it on the timelineOrdovician
485 Ma → 444 MaPeriod
The Cambrian settled the body plans; the Ordovician filled the sea with them, then the ice carried much of that life away.
Its bounds are golden spikes and firm. What happens inside is less so: the size of the cooling is reconstructed from the isotopic composition of fossil shells.
See it on the timelineSilurian
444 Ma → 419 MaPeriod
A short period of recovery in which vertebrates acquire jaws, land animals leave their first fossils, and the sky becomes a filter.
Two golden spikes hold it. The dates inside are less settled: the oldest body fossil of a land animal gives 425 million years, molecular clocks give considerably older.
See it on the timelineDevonian
419 Ma → 359 MaPeriod
Plants invent wood and become trees, vertebrates walk out of water, and the reefs reach a size never equalled since.
The bounds are golden spikes. What the forests did to the climate is where the argument sits: they were long credited with drawing carbon dioxide down and so cooling.
See it on the timelineCarboniferous
359 Ma → 299 MaPeriod
Forests bury carbon faster than decay returns it: oxygen reaches the highest level of the past half-billion years, insects grow enormous, and an egg frees vertebrates from water.
Two golden spikes bound it. The oxygen figure is modelled from the carbon and sulfur cycles rather than measured, and the models disagree on the peak.
See it on the timelinePermian
299 Ma → 252 MaPeriod
The period of a single continent, when the relatives of mammals ruled the land, closed by the worst extinction in the history of life.
Its top is among the most precisely dated boundaries in the Phanerozoic, the extinction interval bracketed to within a few tens of thousands of years by uranium.
See it on the timeline
Mesozoic
252 Ma → 66 MaEra
A hundred and eighty-six million years between two catastrophes: the only era on the chart that both begins and ends with a mass extinction.
Both of its bounds are dated more tightly than most on the chart, because each falls on an event that left its trace across the whole planet at once.
See it on the timelineTriassic
252 Ma → 201 MaPeriod
A world rebuilt from almost nothing, in which dinosaurs and mammals both begin, small and unremarkable, before another eruption ends it.
A golden spike opens it; its top is dated by uranium and lead in the basalts of the Central Atlantic province, which is why it is known to better.
See it on the timelineJurassic
201 Ma → 145 MaPeriod
The supercontinent comes apart, the largest animals that ever walked reach their full size, and small feathered dinosaurs take to the air.
A golden spike opens it. Its top, at 145 million years, is one of the few Phanerozoic boundaries still without one, and its estimated age has moved.
See it on the timelineCretaceous
145 Ma → 66 MaPeriod
The longest period of the Phanerozoic: flowers spread across the land, the world grows hot enough to lose all its ice, and an asteroid closes it in a single day.
Its top is among the sharpest boundaries on the chart, an iridium layer traceable worldwide. Its base is not: at 145 million years it still lacks a golden spike.
See it on the timeline
Cenozoic
66 Ma → 2026Era
The last sixty-six million years, in which mammals inherit a world emptied by an asteroid, grasslands spread, ice returns, and one primate lineage ends up drawing timelines.
Only one of its bounds is fixed. The older is among the sharpest on the chart; the recent end stays open, because this is the era we are in.
See it on the timelinePaleogene
66 Ma → 23 MaPeriod
Forty-three million years in which the world turns from a greenhouse all but free of ice into one with a polar cap, and mammals grow into the space the asteroid left.
Its base is the Cretaceous boundary, one of the sharpest lines on the chart; its top, at 23 million years, is an ordinary stratigraphic boundary and nothing like.
See it on the timelinePaleocene
66 Ma → 56 MaEpoch
Ten million years in an emptied world: the survivors of the asteroid grow large fast, and the epoch ends in the sharpest carbon spike of the Cenozoic.
It begins at the Cretaceous boundary, among the sharpest on the chart. Its top is defined on the carbon isotope excursion itself, which is to say.
See it on the timelineEocene
56 Ma → 33.9 MaEpoch
The warmest stretch of the Cenozoic, in which bats fly, whales enter the sea and primates begin, and which closes with the freezing of Antarctica.
Both ends are well defined, the base on the carbon spike and the top near the first great Antarctic glaciation. How much the Azolla bloom contributed to the cooling.
See it on the timelineOligocene
33.9 Ma → 23 MaEpoch
The epoch nobody can quite explain: Antarctica carries an ice sheet close to its modern size, and the rest of the world stays warm anyway.
Its own specialists call it an enigma in a paper title. The bounds are defined, but the climate between them is poorly documented compared with what lies on either side.
See it on the timeline
Neogene
23 Ma → 2.6 MaPeriod
Grasslands spread across the continents, the Mediterranean dries out and refills, and a branch of African apes begins to walk upright.
Its bounds are golden spikes. The C4 expansion is dated by carbon isotopes: in fossil teeth, which record what the animals ate, and in ancient soils.
See it on the timelineMiocene
23 Ma → 5.3 MaEpoch
The longest epoch of the Neogene, in which a plateau rises and a monsoon strengthens with it, apes are many and various, and the Mediterranean dries out at the very end.
Golden spikes hold both ends. The plateau rose and the monsoon strengthened over the same stretch of time, which does not prove that one made the other.
See it on the timelinePliocene
5.3 Ma → 2.6 MaEpoch
The last warm epoch before the ice ages: apes walk upright, strike the oldest stone tools yet found, and the two Americas join.
Both boundaries are held by golden spikes; the argument is inside them. The closing of the Panama isthmus is genuinely contested: a review that weighs geology.
See it on the timeline
Quaternary
2.6 Ma → 2026Period
The period of the ice ages, and the only one in which our own species exists: ice sheets advance and retreat dozens of times to an orbital beat.
Its base was moved from 1.8 to 2.58 million years by a ratification in 2009, a reminder that a boundary is a decision before it is a date.
See it on the timelinePleistocene
2.6 Ma → 11.7 kaEpoch
The ice ages themselves: two and a half million years of advancing and retreating ice, crossed by at least three kinds of human being.
Its base was lowered from 1.8 to 2.58 million years in 2009, so the epoch gained three-quarters of a million years by a vote rather than by a discovery.
See it on the timelineGelasian
2.6 Ma → 1.8 MaAge
The age that begins the Quaternary, added to it in 2009: the northern ice sheets take hold, and hominins are already leaving Africa.
The boundary is well defined, though its stated relation to the Gauss-Matuyama magnetic reversal has itself been corrected in the literature since it was first published.
See it on the timelineCalabrian
1.8 Ma → 774 kaAge
An age defined by a golden spike that had already served once: it carried the base of the Quaternary until 2009, and was given this new job in 2011.
The boundary is firm, since it was an established marker before it was a Calabrian one. The shift in glacial rhythm that begins in this age has no agreed cause.
See it on the timelineChibanian
774 ka → 129 kaAge
The only unit of the geological time scale named after a place in Japan, and the age in which our own species appears.
Its boundary is fixed, and the magnetic reversal beside it is one of the best-dated events of the Quaternary. The appearance of our species is not a date.
See it on the timelineLate Pleistocene
129 ka → 11.7 kaAge
The last hundred and twenty thousand years of the ice ages: our species leaves Africa, settles the last habitable continents, and paints.
Alone among the Pleistocene subseries it still has no golden spike: its base is fixed by convention at the onset of the last interglacial rather than by a section.
See it on the timeline
Holocene
11.7 ka → 2026Epoch
Eleven thousand seven hundred years of unusual climatic calm, in which farming, cities and writing all appear.
The base is precisely defined. The top is the present, or so the chart says: a proposal to close the epoch in the mid-twentieth century and open.
See it on the timelineGreenlandian
11.7 ka → 6,237 BCEAge
The first age of the Holocene, defined in a Greenland ice core: the world warms for good, farming begins, and the Sahara is grassland.
Counting the ice layers dates the boundary to the year, a precision no rock section can offer, but that precision belongs to the core and not to the world.
See it on the timelineNorthgrippian
6,237 BCE → 2,251 BCEAge
An age that opens on the sharpest cold spell of the Holocene, and in which some adults keep the ability to digest milk.
Here too the layer count gives a boundary good to the year, but the event under it is a regional signal read as global: how far the 8.2 ka cooling reached beyond.
See it on the timelineMeghalayan
2,251 BCE → 2026Age
The briefest unit on the whole geological time scale, and the only one defined in a stalagmite: it begins with a drought and runs to today.
Precise at its base and contested in its status: ratified in 2018, it was criticised at once, partly because the 4.2 ka event is a regional monsoon signal used.
See it on the timeline
The chart, and its sources
- Cohen, K. M., Finney, S. C., Gibbard, P. L., Fan, J.-X. (2013). The ICS International Chronostratigraphic Chart. Episodes 36(3), 199–204, mise à jour continue, voir Cohen et al. (2025), Episodes.
- Walker, M., Head, M. J., Berkelhammer, M., et al. (2018). Formal ratification of the subdivision of the Holocene Series/Epoch: two new GSSPs and three new stages/subseries. Episodes 41(4), 213–223.
- Head, M. J. (2019). Formal subdivision of the Quaternary System/Period: Present status and future directions. Quaternary International 500, 32–51.