Skip to content
Earth & Climate

The Rock Cycle

The mountain, the beach sand, and the marble countertop can be the same atoms of rock caught at different moments of a journey tens of millions of years long.

10 min read·July 9, 2026

ISM
On this page

The same atoms, caught mid-journey#

Stand on a granite mountaintop, pick up a handful of quartz sand on a beach, run your palm across a polished marble countertop. Three things that could not feel more different — one immovable, one loose enough to pour through your fingers, one cut and shipped as the very emblem of a solid, finished surface. And yet, statistically, some of the atoms in all three have been the same rock. They are the same journey, photographed at three different moments.

We treat rock as the definition of permanence. "Set in stone" means settled forever; "solid as a rock" means it will not move. It is the most durable material most people ever touch, and the instinct that it is fixed and finished is almost irresistible. That instinct is wrong. Every rock you have ever seen is partway through a transformation. The mountain is being ground down grain by grain and carried to the sea. The beach sand is on its way to being buried, cemented, and turned back to stone. The marble was once mud on a seafloor, was once a living reef, and before that was atmospheric carbon; it may yet be dragged down into the mantle, melted, and erupted as fresh lava. Given enough time, and Earth has almost incomprehensible amounts of it, rock is endlessly melted, ground down, buried, and reborn.

The name for that restless recycling is the rock cycle. It sits alongside the carbon cycle and the water cycle as one of Earth's master cycles — and, like them, it is best understood not as a story about substances that stay put but as bookkeeping: a fixed stock of material shuffled between reservoirs by a set of driving forces. Here the reservoirs are the three great families of rock, and the forces are Earth's internal heat, delivered by plate tectonics, and the sun-and-gravity machinery that wears the surface down.

Three ways to make a rock#

Every rock on Earth belongs to one of three families, and the families are defined not by what a rock is made of but by how it formed. That is the single idea to hold onto: the classification is a classification of processes.

Igneous rock forms when molten rock cools and crystallises. Melt — called magma underground, lava once it erupts — is a hot liquid of silicates. As it loses heat, atoms lock into crystal lattices and the liquid becomes solid. The crucial variable is how fast. Magma that cools slowly, kilometres down inside the crust, gives crystals a long time to grow, producing coarse-grained rock with interlocking crystals you can see with the naked eye — granite is the type example. This is called intrusive or plutonic rock. Lava that erupts and cools in hours to days at the surface freezes into fine-grained rock whose crystals are microscopic — basalt is the type example, and it is extrusive or volcanic. Same process, opposite speeds, visibly different rock.

Sedimentary rock forms when older rock is broken down and the pieces are reassembled. Weathering and erosion break exposed rock into fragments — sediment — ranging from boulders to clay and dissolved ions. Water, wind, ice, and gravity transport that sediment, then deposit it wherever the energy of the transporting medium drops: a river mouth, a lake bed, the floor of a sea. Layer settles on layer, and over time the weight of overburden compacts the pile and mineral cement precipitates in the pore spaces, gluing loose grains into solid rock. This lithification is why sedimentary rocks come in layers (strata) and why they, almost uniquely, can preserve fossils — the buried remains and traces of once-living things. Sandstone (cemented sand) and limestone (largely the calcium-carbonate remains of marine organisms) are the classic examples.

Metamorphic rock forms when existing rock is transformed in the solid state by heat and pressure — without fully melting. This is the family people find least intuitive, because nothing obvious "happens": no melt, no erosion, just rock going in and different rock coming out. But under the heat and pressure of deep burial or a tectonic collision, the minerals in a rock become unstable and recrystallise into new minerals, often growing larger and aligning into bands. The rock's chemistry may barely change while its texture and mineralogy are wholly remade. Bury limestone and it recrystallises into marble; squeeze the mud-rock shale and it becomes slate, then schist, then gneiss as conditions intensify. The defining line is melting: cross it and you are back to making igneous rock; stay just below it and you are metamorphosing.

Three families, three processes: cool a melt, cement some sediment, or cook and squeeze what is already there. Everything else is detail.

Watching the cycle turn#

Here is the step that overturns the "rocks are permanent" intuition for good. None of these three states is a destination. Each is a doorway to the others, and the processes above are the doors. Cool a melt and you have igneous rock — but weather it and it becomes sediment; bury that sediment and it becomes sedimentary rock; cook it and it becomes metamorphic rock; push it deep enough and it melts back to magma, ready to start again as something else entirely.

The widget is that web of transformations. The three rock families sit at the corners, with magma and loose sediment as the two in-between states, and the arrows are the processes that carry material from one to another — cooling, weathering and erosion, deposition and compaction, heat and pressure, and melting.

Things to try:

  • Pick a starting rock and a single process, and follow one step. Start on igneous rock and choose weathering: watch it break down into sediment. From sediment choose deposition and compaction: it lithifies into sedimentary rock. Each door leads somewhere, and the label tells you which process is opening it.
  • Chase a long path and notice there is no "correct" one. Take metamorphic rock and melt it to magma, cool it to igneous, weather it to sediment, compact it to sedimentary, then cook it straight back to metamorphic. You have returned to where you started without ever repeating a step in a fixed order. Now try a completely different route to the same place. The point lands physically: any rock can become any other, and there is no single loop it must follow.
  • Watch which arrows are surface processes and which are deep ones. Weathering, erosion, and deposition happen up in the sunlight, powered by water and gravity. Melting, deep burial, and the heat-and-pressure of metamorphism happen kilometres down, powered by Earth's internal heat and the motions of plate tectonics. The cycle straddles two engines, and every complete circuit uses both.

The most important thing the animation makes concrete is a negative: it has no fixed direction. There is no law that says igneous must come first or sedimentary last. The old textbook diagram with a tidy clockwise loop is a simplification bordering on a lie. The real cycle is a network, and material takes whatever path conditions allow.

The mathematics of deep time#

The rock cycle runs on timescales that defeat intuition, and a little arithmetic — the same reservoir-and-flux arithmetic that organises the carbon and water cycles — is what makes those timescales legible.

Residence time = reservoir ÷ flux#

Treat each family of rock as a reservoir holding some mass of material, connected to the others by fluxes (the rates at which material transforms and moves). For any reservoir of size RR drained by a through-flux FF, the average time a given parcel of rock spends there before moving on is the residence time:

τ  =  RF\tau \;=\; \frac{R}{F}

This is exactly the definition used for a molecule of water in the atmosphere or a tonne of carbon in the ocean — reservoir divided by flux — and it behaves the same way: a large reservoir drained by a small flux turns over slowly, a small reservoir flushed by a large flux turns over fast, and it is the ratio, not the size, that decides.

Put in rough numbers. The sedimentary shell of the continents holds on the order of R2.5×1018R \sim 2.5\times10^{18} tonnes of rock, and rivers deliver sediment to the sea — the flux that erodes old rock and builds new sedimentary rock — at very roughly F2×1010F \sim 2\times10^{10} tonnes per year. Then

τsed  =  RF    2.5×1018 t2×1010 t/yr    108 years,\tau_{\text{sed}} \;=\; \frac{R}{F} \;\approx\; \frac{2.5\times10^{18}\ \text{t}}{2\times10^{10}\ \text{t/yr}} \;\approx\; 10^{8}\ \text{years,}

on the order of a hundred million years for the sedimentary reservoir to cycle through once. These figures are order-of-magnitude estimates, and geologists quote ranges rather than points, but the scale is the lesson: the hook's "tens of millions of years" is not hyperbole, it is the residence time.

The contrast between the two kinds of crust sharpens the point. Oceanic crust is manufactured at mid-ocean ridges and destroyed at subduction zones, so its reservoir is small and its flux large: the entire ocean floor is recycled in roughly 150–200 million years, which is why no seafloor older than about 180 Myr exists anywhere on Earth. Continental crust is too buoyant to subduct, so it accumulates: its average age is around 2 billion years, and some grains are over 4 billion years old. Same planet, same cycle, residence times differing by more than an order of magnitude — because the fluxes differ, exactly as τ=R/F\tau = R/F predicts.

Grain size and cooling rate: an inverse law#

The igneous branch hides a second, cleaner relationship. When a melt crystallises, two things compete: the nucleation of new crystals and the growth of existing ones. Cool slowly and few nuclei form but each has a long time to grow, so the rock ends up coarse; cool quickly and a swarm of nuclei freeze in before any can enlarge, so the rock ends up fine. The final grain size dd therefore falls as the cooling rate T˙\dot{T} rises:

d    T˙n,n1312d \;\propto\; \dot{T}^{-n}, \qquad n \approx \tfrac{1}{3}\text{–}\tfrac{1}{2}

an inverse relationship: faster cooling, smaller grains. The exact exponent depends on the melt, but the direction is universal and it is why you can read a rock's cooling history straight off its texture. Coarse, interlocking crystals (granite) mean slow cooling kilometres down; a glassy or microscopic texture (basalt, or the extreme case of volcanic glass, obsidian, where cooling was so fast that essentially no crystals formed) means rapid cooling at or near the surface. The grain size is a thermometer for how the rock was born.

Three routes to rock, side by side#

The cycle diagram shows how the families connect. The second widget zooms in on the three formation processes themselves, running them next to each other so the differences are concrete.

Each panel is one route to solid rock:

  • Igneous — a body of magma cools and crystals nucleate and grow. Drag the cooling-rate control. Slow it down and watch a few large, interlocking crystals fill the frame (a granite texture); speed it up and watch a dense mat of tiny crystals freeze in (a basalt texture). This is the dT˙nd \propto \dot{T}^{-n} law made visible: you are trading grain size against cooling rate with your own hand.
  • Sedimentary — grains rain down and stack into layers, older at the bottom, and compaction squeezes them into rock. Watch a fossil get buried in a layer and preserved. The layering and the trapped fossils are the signatures no other family carries, and they are why sedimentary rock is the pages of Earth's history book.
  • Metamorphic — a pre-existing rock is squeezed and heated, and its crystals recrystallise and realign into bandswithout melting. Watch the grains rotate and grow into foliation while the rock stays solid throughout. Nothing pours; nothing erodes; the rock is remade in place.

Trigger each and compare. The three panels never converge on the same picture, because they are genuinely different physics: freezing a liquid, gluing a pile, and re-cooking a solid. That is the whole taxonomy, seen in motion.

Why it matters: the cycle that runs the planet#

The rock cycle is not a museum curiosity. It is coupled to the other master cycles and to the habitability of the planet itself.

The tightest coupling is to the carbon cycle, and the link is weathering. Rain is slightly acidic because it carries dissolved CO2 as carbonic acid, and that acid chemically attacks silicate rock on the continents. The reaction consumes CO2, liberates dissolved ions, and delivers them to the ocean, where organisms and chemistry lock the carbon away as carbonate rock — limestone — on the seafloor. So the surface half of the rock cycle is quietly drawing CO2 out of the atmosphere and burying it as stone. Over hundreds of thousands to millions of years this is the planet's long-run thermostat: warmer, wetter climates weather rock faster, pulling down more CO2, which cools things back off. The rock cycle and the carbon cycle are the same machine seen on two timescales, and silicate weathering is the gear that couples them.

The deep half of the cycle is driven by plate tectonics, which supplies the heat and the motion. Subduction drags old crust and its cargo of sediment down into the mantle, where it heats, dehydrates, and triggers the melting that feeds volcanic arcs — turning sedimentary and oceanic rock into new igneous rock. Collision and burial cook rock into metamorphic belts and thicken the crust into mountains. Uplift then raises deep rock back to the surface where weathering can attack it, closing the loop between the two engines. Without plate tectonics the deep half of the cycle would stall; the Moon and Mars, which lost their internal heat engines long ago, have surfaces where the rock cycle has largely stopped, preserving craters billions of years old that on Earth would have been erased many times over.

That is the deepest reason the "permanence" intuition fails. A rock feels eternal on a human timescale precisely because our timescale is so short compared to τ=R/F\tau = R/F. Wait long enough and the mountain is sand, the sand is stone, the stone is marble, and the marble is magma. The granite peak, the beach, and the countertop really are one substance, photographed at three moments of a journey that never ends.

Key takeaways
  • Rocks fall into three families by how they form: igneous from cooling and crystallising melt (slow cooling underground makes coarse grains like granite, fast cooling at the surface makes fine grains like basalt), sedimentary from weathered sediment that is transported, deposited, and compacted into fossil-bearing layers like sandstone and limestone, and metamorphic from existing rock transformed by heat and pressure without melting, like limestone to marble or shale to slate.
  • Rock is not permanent. Over deep geological time every rock is continuously melted, ground down, buried, and reborn as one of the other families.
  • The rock cycle has no fixed one-way sequence — it is a network, not a loop, and any rock type can transform into any other given the right conditions.
  • The same residence-time logic as the carbon and water cycles applies: τ=R/F\tau = R/F explains why buoyant continental crust averages ~2 billion years old while subductable oceanic crust recycles every ~150–200 million years. Igneous grain size follows an inverse law, dT˙nd \propto \dot{T}^{-n}: faster cooling, finer grains.
  • The cycle is powered by two engines and coupled to the others: Earth's internal heat via plate tectonics (uplift, burial, subduction, melting) and surface processes driven by sun, water, and gravity (weathering, erosion, deposition). Silicate weathering draws CO2 out of the air, linking the rock cycle to the carbon cycle as the planet's long-term thermostat.
Check your understanding
1. A granite and a basalt can have almost identical chemical compositions, yet granite is coarse-grained (crystals you can see) while basalt is fine-grained (crystals you need a microscope for). What sets the difference?
2. Limestone buried and squeezed becomes marble, and marble can be uplifted, weathered into grains, redeposited, and lithified back into limestone. What does this illustrate about the rock cycle?
3. Rain that is slightly acidic with dissolved CO2 chemically weathers silicate rock on the continents. Over hundreds of thousands of years, what effect does this have on the atmosphere?
0 / 3 answered

Share this article

Share on X