Rock Cycle Diagram, Rock Identification Key and QAPF Classifier

How to use the three tools on this page

Start with the cycle diagram. The five boxes are reservoirs, the places rock sits between events, and the eleven arrows are the processes that move material between them. Click any arrow and the panel gives you the mechanism, the temperature and depth window it operates in, a rate with units, and the field evidence that tells you it happened. That last part matters. A process you cannot recognise in outcrop is a process you have memorised rather than understood.

The identification key is a live dichotomous key over 92 rock types. It is not a fixed flowchart. After each answer it recalculates which remaining question splits the surviving candidates most evenly, so the path through it depends on your rock. Most specimens resolve in six or seven questions. If you cannot answer one, use the skip button rather than guessing, because a wrong answer removes the correct rock permanently and the key will then confidently walk you to the wrong place.

The QAPF classifier is the third tab and it exists because the key deliberately stops short on one group. Answer the key honestly on a coarse igneous rock and you will land on a shortlist of granite, granodiorite, tonalite, diorite, monzonite and syenite, and no further field test separates them. They are separated by the ratio of quartz to alkali feldspar to plagioclase, which is what QAPF plots. Estimate those three percentages, recalculate them so they sum to 100, and click the point.

The input people get wrong is the acid test. Put the drop on a freshly broken surface, not a weathered one, and watch for at least ten seconds. A weathered limestone surface can be so case hardened it barely reacts, and a dolostone will do nothing at all until you scratch it to powder first. The key treats "fizzes" and "fizzes only when powdered" as different answers, because that single distinction separates limestone from dolostone and those two rocks behave differently in almost every engineering context that matters.

The rock cycle is not a cycle

The diagram everyone is taught shows a circle: igneous becomes sedimentary, sedimentary becomes metamorphic, metamorphic melts and returns to igneous. It is drawn as a loop with arrows all running the same way round. Almost every part of that picture is wrong, and the wrongness is not pedantic, it changes what you predict.

There is no fixed direction. Sediment can be buried, lithified, uplifted, weathered and redeposited without ever going near a metamorphic condition, and most sediment does exactly that. Roughly half the sand in a large river system is recycled from older sandstone rather than freshly liberated from granite. There is also no requirement to pass through every box. An igneous rock can be metamorphosed directly, which is what happens to essentially all new ocean floor within a few million years of its formation. A sedimentary rock can melt directly if you push it deep enough down a subduction zone.

The honest way to think about it is as a budget with pathways rather than a loop. Material sits in five reservoirs. Eleven processes move it between them at rates that differ by many orders of magnitude, and the direction any particular parcel of rock takes depends entirely on where it is on the planet.

Two independent engines drive the whole thing, and only one of them is usually mentioned. Earth's internal heat flow is about 47 terawatts, and it powers plate motion, metamorphism and melting. Solar input is about 174,000 terawatts, and it powers weathering, erosion and transport. Those are not comparable numbers. The surface half of the cycle runs on an energy supply roughly 3,700 times larger than the deep half, which is why erosion can strip a mountain range faster than tectonics can build it, and why the rate limiting step in the cycle is almost never the weathering.

The cycle is also not closed. Material goes down subduction zones and some of it stays down. Some returns in arc magmas within a few million years, some is stored in the mantle for a billion years or more, and some is never coming back on any timescale that matters to us. That is why the continental crust has grown over Earth history rather than staying constant, and it is why the phrase "the rock cycle" should be read as an accounting identity rather than a promise that everything returns.

The eleven pathways, and what actually happens on each

The panels in the tool carry the full detail. What follows is the reasoning behind the ones that are most often taught as slogans.

Crystallisation, and why slow cooling is not the reason for big crystals

Every geology course teaches that intrusive rocks are coarse because they cooled slowly and volcanic rocks are fine because they cooled fast. The correlation is real. The mechanism given for it is not.

Grain size is set by the ratio of crystal growth rate to nucleation rate, not by elapsed time. If a melt nucleates a very large number of crystals, you get a fine rock no matter how long you wait, because all those nuclei compete for the same material. If it nucleates very few, each one grows large. What controls the nucleation rate is undercooling, the amount by which the melt sits below its liquidus, together with the presence of fluxing components such as water, boron and fluorine that lower viscosity and speed up diffusion.

The decisive evidence is pegmatite. Granitic pegmatites contain feldspar and mica crystals metres across, and the textbook inference is that they must have cooled over millions of years. Experimental and textural work shows the opposite: they grow at large undercooling in fluxed, water rich melt, and crystals that size can form in days to years. Slow cooling usually accompanies coarse grain size because both follow from being deep in the crust. It does not cause it.

The consequence for identification is that you cannot read cooling time off grain size. You can read the environment, which is what you actually wanted.

Weathering and erosion, the fastest arrow on the diagram

Chemical weathering attacks silicates through hydrolysis, and the rate depends on temperature, water availability and how much fresh mineral surface is exposed. Physical erosion controls that last term, which is why the two are coupled rather than independent. A steep wet tropical mountain belt strips at 1,000 metres per million years and above. A stable cratonic shield in a dry climate runs below 5 metres per million years. That is a spread of more than two orders of magnitude, and it is set mostly by relief and rainfall rather than by rock type.

The important asymmetry is that this arrow is fast. Given a mountain range and no tectonic renewal, erosion flattens it in tens of millions of years. Continents persist not because erosion is slow but because the crust keeps being rebuilt and because the isostatic root rises as the top is removed, which returns roughly five metres of rock for every six removed.

Lithification, which is not caused by pressure

Ask where sedimentary rock comes from and you will be told that sediment is compacted by the weight of overlying material until it becomes rock. Compaction is real and it does expel water, but on its own it never makes a rock.

Pile clean quartz sand as deep as you like and mechanical compaction stalls near 26 percent porosity. The grains have reached a packing limit. They are pressed together, they are not bonded, and if you removed the overburden you would still have sand. What turns sand into sandstone is mass transfer through the pore fluid: silica dissolves at stressed grain contacts where solubility is higher, migrates a short distance, and precipitates in the pore space as cement. The rock is welded chemically, not squeezed together.

This explains a fact that the compaction story cannot. Shallow marine carbonate can lithify into hard rock in decades, because aragonite is highly soluble and reprecipitates as calcite almost immediately in seawater. Siliciclastic sand under the same burial takes 10 to 100 million years, because quartz is barely soluble. If pressure were the mechanism, the deeper rock would always be the harder one, and it is not.

Prograde metamorphism, and what heat and pressure actually means

"Heat and pressure" is a label, not an explanation. What happens is a sequence of specific mineral reactions, each with a defined temperature and pressure at which it runs, and each releasing or consuming volatiles.

Take a mudrock and bury it. Near 200 degrees the clays begin to react and you get slate, with a cleavage that has nothing to do with the original bedding and everything to do with mica growing perpendicular to the maximum stress. Around 400 degrees the micas coarsen enough to give phyllite its sheen. By 500 to 600 degrees you have visible mica and often garnet, which is schist. Above roughly 650 degrees the feldspars and quartz segregate into bands and you have gneiss. Push past about 700 degrees with water present and the rock begins to melt, giving migmatite, which is the boundary case where metamorphism hands over to igneous.

The single most important constraint is that metamorphism is solid state by definition. If it melted, it is not metamorphism any more. That is the line between gneiss and migmatite, and it is why the answer to "does metamorphism melt the rock" is no, right up until it does, at which point we call it something else.

Anatexis, and why granite is not melted basalt

Partial melting is the normal case and it is the reason the continental crust exists in the form it does. When you melt a rock, the first liquid to form is richer in silica, aluminium, potassium and sodium than the bulk source, because the minerals with the lowest melting points are the feldspars and quartz. Take that melt away and the residue left behind is more mafic than it started.

This is why melting the mantle, which is ultramafic, gives basalt rather than more mantle, and why melting basaltic or greywacke crust gives granite rather than more basalt. Each round of partial melting moves the melt toward the felsic end and the residue toward the mafic end. The continental crust is the accumulated melt fraction of a planet, which is exactly why it floats.

There is a mass balance argument worth knowing, because it kills a common shortcut. If you tried to make granite by fractionally crystallising basalt, you would need roughly 90 percent crystallisation, and you would have to leave behind about nine times as much ultramafic cumulate as the granite you produced. Those cumulates are not found under continents in anything like the required volume. Crustal melting, not fractionation of mantle melt, makes most granite.

Slab dehydration, the arrow that is usually drawn wrong

Almost every diagram of a subduction zone shows the slab melting and the melt rising to feed the volcanoes. In most arcs that is not what happens.

As the slab descends it heats up, and hydrous minerals in it break down in sequence. Chlorite, amphibole and lawsonite release water between roughly 80 and 150 kilometres depth. That water rises into the overlying mantle wedge, and water dramatically lowers the melting point of peridotite. The wedge melts, not the slab. The resulting basaltic melt then rises, ponds at the base of the crust, and differentiates or triggers crustal melting on its way up, which is why arc volcanoes erupt andesite rather than basalt.

Sediment on the slab is the exception that gets misused. At about 100 kilometres the slab top reaches 700 to 850 degrees, which is above the wet solidus for sediment but below that for basalt. So subducted sediment genuinely does melt, and its distinctive isotopic signature shows up in arc lavas. But it contributes signature, not volume. If you take away one thing: the slab dehydrates, the wedge melts.

Ocean floor metamorphism, the biggest arrow nobody draws

Here is a fact that should reorder your mental picture. The largest volume of metamorphic rock produced on Earth today is not made in mountain belts. It is made at mid ocean ridges.

New oceanic crust forms at about 3 cubic kilometres a year, and essentially all of it is hydrothermally altered within a few million years of forming. Seawater circulates through the fractured, hot basalt, and the primary igneous minerals react to chlorite, epidote, actinolite and albite. That is greenschist facies metamorphism, running continuously beneath 60 percent of the planet's surface. By volume it dwarfs the metamorphism happening in every collisional orogen combined.

It appears on no standard rock cycle diagram, because the standard diagram was drawn before plate tectonics and never updated. The tool on this page includes it as the igneous to metamorphic pathway, and it is worth clicking, because it also explains why so much of the seafloor is serpentinite and why subducting slabs carry so much water in the first place. The ridge hydrates the crust, and the trench delivers that water to the mantle wedge 100 kilometres down. Those two arrows are the same water.

Exhumation, and three words that are not synonyms

Uplift, surface uplift and exhumation get used interchangeably and they mean different things. Uplift of rock is displacement of rock relative to the geoid. Surface uplift is displacement of the land surface relative to the geoid. Exhumation is the displacement of rock relative to the surface, which is to say the depth of material removed from above it.

The distinction matters because thermochronology, which is how these numbers are actually measured, records exhumation only. A cooling age tells you when a rock passed upward through a given isotherm. It tells you nothing directly about whether the mountain got taller. A range can exhume rapidly while its summit elevation stays constant or even falls, which is the usual steady state for an active orogen. Confusing the three is how people end up claiming a mountain belt rose by an amount that is actually the thickness of rock eroded off it.

Typical exhumation runs 0.1 to 1 millimetre a year in active orogens. Ultrahigh pressure terranes, the ones that carry coesite and prove burial to 100 kilometres and back, require rates near or above 1 centimetre a year, and how they achieve that is still argued about.

The three arrows that get left off

Cannibalistic recycling is sedimentary rock weathering straight back to sediment without any metamorphic or igneous step. It is quantitatively one of the largest fluxes in the system and it appears on essentially no classroom diagram, because the diagram is drawn as a loop and a loop has no room for a shortcut. Roughly half the quartz sand in a mature river system has been through at least one previous sandstone. Each pass destroys more feldspar and lithic grains, so the sand becomes purer quartz with every cycle. A quartz arenite is not evidence of one intense weathering event, it is usually evidence of several ordinary ones.

Remelting of igneous crust is the arrow from igneous straight back to magma without any intervening step. Underplated basalt at the base of an arc reheats and melts, and so does older granite when a new intrusion arrives beneath it. This is where a large fraction of granite actually comes from, and the isotopic evidence for it is unambiguous, because the melts inherit the radiogenic signature of old crust rather than of mantle.

Burial and contact metamorphism take igneous rock to metamorphic rock directly. Contact metamorphism bakes an aureole around an intrusion, producing hornfels with no foliation because there is no differential stress, just heat. Burial metamorphism affects thick volcanic piles that subside under their own weight. Neither requires mountain building, and together with the ocean floor case they mean the igneous to metamorphic arrow is one of the busiest on the diagram rather than an afterthought.

Where the cycle actually runs

The circular diagram has no geography in it, and that is its worst failure. Every arrow belongs to a place.

At a mid ocean ridge, decompression melting of rising mantle produces basalt, which erupts as pillow lava and freezes as gabbro below. Seawater alters it immediately. That is three pathways operating within a few kilometres of each other.

Across the ocean basin, pelagic sediment settles at 1 to 20 metres per million years, a rate so slow that the whole abyssal sediment pile is usually thinner than a single delta foreset bed. It rides the plate toward the trench.

At the trench, that sediment either scrapes off into an accretionary wedge, where it is deformed and lithified and sometimes metamorphosed to blueschist, or it goes down with the slab. Which one happens depends on convergence rate and sediment supply, and both occur in different segments of the same margin.

Beneath the arc, the sequence described above runs: dehydration at depth, wedge melting, ascent, differentiation, eruption. New igneous rock is added to the crust here, and this is the main site of continental growth.

In a collision belt, crust doubles in thickness. The lower half of the pile goes through prograde metamorphism, and if it gets hot enough with water present, its base melts and makes the granite that cores most large mountain ranges. Meanwhile the top is being stripped by erosion at rates that can exceed the rate of thickening, feeding sediment into the foreland basin where lithification starts the sedimentary arrow again.

How the identification key decides what to ask

The key is adaptive, so it is worth knowing what it is optimising. At every step it scores each remaining question by how evenly that question would split the surviving candidates, weighted by how many candidates it actually applies to, and weighted again by a preference for structural observations over mineral identification. It then asks the highest scoring one.

The preference for structure is deliberate. A question like "does it have a planar fabric" can be answered by anyone holding the rock, and it is close to a perfect binary split across the dataset. A question like "do you see pyroxene" requires you to already know what pyroxene looks like, and it is meaningless on a rock too fine to resolve. So mineral questions are downweighted and get asked late, once the candidate set is small enough that they are doing real work.

Three observations carry most of the load, and they are worth internalising even without the tool.

Fabric first. Does the rock have a planar or linear fabric that cuts across any original layering? If yes, it is metamorphic, and you have removed roughly two thirds of the dataset in one question. Be careful not to confuse foliation with sedimentary bedding: bedding is compositional layering laid down flat, foliation is a mineral alignment produced by stress, and where you can see both, they are usually at an angle to each other.

Fragments or crystals second. Is the rock made of pieces that were once separate, cemented together, or of crystals that grew interlocked in place? Fragmental means sedimentary or pyroclastic. Interlocking means igneous or metamorphic or a chemical precipitate. This is a texture question and it survives on a weathered surface better than most.

Acid third. A carbonate answer collapses the candidate set immediately, and the powdered variant separates limestone from dolostone. Twelve of the 92 rocks in the dataset react to cold dilute acid in some form, so a positive result is genuinely informative.

A worked identification you can repeat

Take a specimen and follow along in the tool. This is the exact sequence the key produced on a common rock, with the candidate count after each answer.

Start with all 92 rock types. The key opens with the fragmental question. Our specimen is built of rounded pebbles set in a finer matrix, so the answer is fragmental, and the count drops from 92 to 25.

Second question, will it scratch window glass. The pebbles are quartzite and quartz vein material and they scratch glass easily, so yes. That takes 25 down to 11, and it has quietly removed every carbonate and every soft evaporite from consideration.

Third, are any crystals larger than a centimetre. The pebbles are two to five centimetres, so yes. 11 down to 5.

Fourth, are most grains coarser than 2 millimetres. Yes again, and now there are 3 candidates left: volcanic breccia, conglomerate and sedimentary breccia.

That is where a well built key should stop, because those three rocks are genuinely similar and the thing that separates them is a single observation the key has not asked for. Volcanic breccia is made of angular volcanic fragments, usually with a fine ash matrix. Sedimentary breccia is angular fragments of any composition, which means little transport. Conglomerate is rounded fragments, which means a lot of transport. Our pebbles are rounded, so it is conglomerate.

Notice what the roundness told you. Angular means the clasts have not travelled, so you are looking at a scree, a fault zone, or a collapse deposit. Rounded means kilometres to hundreds of kilometres of river or beach transport. The name is the small part. The transport history is what you actually extracted.

Repeat the run and skip the glass question instead of answering it. The key will ask a different second question and still arrive at the same group, because skipping removes a constraint rather than adding a wrong one. That is the whole reason the skip button exists.

Igneous classification, and what QAPF is really for

Igneous rocks are classified on two things: where they cooled, and what they are made of. Cooling location you get from texture, in one look. Coarse and interlocking means it crystallised at depth. Fine grained means it erupted. Glassy means it was quenched. Large crystals in a fine matrix means both, a two stage history with slow growth at depth followed by eruption.

Composition is the hard part, and QAPF is the IUGS answer for rocks where you can actually see the minerals. You count the modal percentages of quartz, alkali feldspar, plagioclase and feldspathoid, recalculate them to 100 percent, and plot. The upper triangle handles quartz bearing rocks, the lower handles feldspathoid bearing rocks, and no rock plots in both because quartz and feldspathoids are chemically incompatible. A silica undersaturated melt cannot precipitate quartz, which is why the diagram is two triangles sharing an edge rather than one square.

Two things about this diagram catch people out constantly.

The first is that the diorite and gabbro boundary is not on it. Both plot in the same field, at the plagioclase rich end with less than 5 percent quartz. They are separated by the anorthite content of the plagioclase, less than 50 for diorite and 50 or more for gabbro, and in practice in the field by colour index, since gabbro is much darker. If you are looking for the line on the diagram that divides them, there is not one, and that is not an error in the diagram.

The second is that QAPF requires modal mineralogy, which means it only works where you can identify and count minerals. That restricts it in practice to plutonic rocks and coarse volcanics. For fine grained and glassy volcanic rocks you cannot count what you cannot see, so those are classified chemically on total alkalis against silica instead. The two schemes do not always agree on the same rock, and when they disagree the chemical one is usually the one to trust for a fine grained rock.

There is also a naming trap worth flagging. "Quartz monzonite" in older North American usage means what IUGS calls monzogranite, which sits inside the granite field. If you are reading a paper written before about 1980, check which convention it uses before you compare it to anything.

Estimating the modes without a point counter

QAPF asks for volume percentages, and in the field you have a hand lens and a flat surface. Point counting a thin section is the rigorous method, but you can get to the right field on the diagram with a disciplined estimate, and getting to the right field is usually all you need.

Work on a cut or broken surface, not a weathered one, and pick an area at least ten times the diameter of the largest crystal so you are not sampling one patch. Identify the three phases first, before estimating anything. Quartz is glassy and grey, has no cleavage, and breaks with curved surfaces that catch the light in a scattered way. Alkali feldspar is usually pink or white with flat cleavage faces that flash all at once when you rotate the specimen, and it often shows fine parallel exsolution stripes. Plagioclase is white to grey with the same flat flashing cleavage, but it frequently shows sharp parallel twin striations on one cleavage face, which alkali feldspar does not.

Then estimate percentages against a comparison chart rather than by intuition, because untrained estimates are biased high for whatever is most visually obvious. Dark minerals in particular get systematically overestimated because they contrast. Once you have quartz, alkali feldspar and plagioclase as percentages of the whole rock, discard the dark minerals and recalculate the three so they total 100. That recalculation is the step people skip, and skipping it moves the point substantially.

A worked case: you estimate 25 percent quartz, 30 percent alkali feldspar, 25 percent plagioclase and 20 percent biotite and hornblende. Discard the 20 percent mafics. The remaining 80 recalculates to quartz 31, alkali feldspar 38, plagioclase 31. The plagioclase ratio is 31 divided by 69, which is 45. Plot Q at 31 and a plagioclase ratio of 45 and you land in granite. Had you forgotten to recalculate and plotted Q at 25, you would still be in granite, but a rock near a boundary would have moved across it, which is exactly when the answer matters.

Sedimentary classification, and the mismatch that explains a lot

Sedimentary rocks divide first by origin. Siliciclastic rocks are made of transported fragments of older rock. Carbonates are made largely in place by organisms and chemical precipitation. Evaporites precipitate from concentrated brine. Organic rocks are accumulated carbon.

Siliciclastics are then named by grain size, and the boundaries are conventions worth memorising because everything else hangs off them. Coarser than 2 millimetres is gravel grade, giving conglomerate if the clasts are rounded and breccia if they are angular. Between 2 millimetres and 62.5 micrometres is sand grade, giving sandstone. Between 62.5 and 4 micrometres is silt, giving siltstone. Finer than 4 micrometres is clay, giving claystone, and if the rock splits into sheets it is shale.

Sandstones then subdivide by what is in them, and the composition is a provenance signal. Quartz arenite is almost pure quartz, which means intense weathering, long transport, or repeated recycling. Arkose carries more than 25 percent feldspar, which means rapid erosion of a granitic source and burial before the feldspar could weather away. Greywacke is poorly sorted with a muddy matrix and abundant rock fragments, which means rapid deposition, typically by turbidity currents.

Carbonates use a different logic because their grains are made where they sit. The Dunham scheme asks whether the rock is mud supported or grain supported, and whether the organisms are bound in growth position. That distinction is about depositional energy, which is the thing you want to know.

Now the mismatch. Sedimentary rocks cover 70 to 75 percent of the land surface but make up only about 5 to 8 percent of the crust by volume. They are a thin skin over a mostly igneous and metamorphic crust. Almost every intuition people have about how common rock types are comes from this, because you are standing on the skin. It is also why so many people assume most sand comes from weathering granite, when in fact a large proportion of it is recycled from older sandstone that was itself recycled.

Metamorphic classification, grade and facies

Metamorphic rocks split on one observation: is there a foliation or not.

Foliated rocks form under differential stress, and the foliation records the stress orientation. The classic series runs slate, phyllite, schist, gneiss with increasing grade, and all four can come from the same mudrock. Slate has a cleavage but no visible mica. Phyllite has a sheen because the mica has coarsened but is still not resolvable individually. Schist has visible platy minerals. Gneiss has segregated light and dark bands, and by then the fabric is compositional rather than simply aligned.

Non foliated rocks form either without differential stress, as in contact metamorphism around an intrusion, or from a protolith with only one mineral, where there is nothing to align. Marble from limestone and quartzite from sandstone are the second case. Hornfels around a pluton is the first.

Grade is temperature. Facies is the combination of temperature and pressure, defined by which mineral assemblage is stable in a basaltic bulk composition, and it is more useful because it captures the tectonic setting. Blueschist facies, high pressure and low temperature, only forms where cold material is pushed down fast, which means a subduction zone. Its presence in a terrane is direct evidence of subduction. Granulite facies, high temperature and moderate pressure, means deep crust. Eclogite means very high pressure, and if it contains coesite you are looking at a rock that reached at least 90 kilometres depth and came back.

Two cautions. Metamorphic grade tells you the peak conditions the rock reached, not how long it sat there, and not the path it took to get there. And retrograde metamorphism, the reaction back down as the rock cools and rises, is generally incomplete, because the reactions need water and the water usually left during the prograde leg. That incompleteness is a gift, because it is why we can still read peak conditions from a rock sitting at the surface.

What is taught wrong

The tool has a tab listing fourteen of these with corrections. These are the ones that cause the most downstream damage.

That the cycle has a direction. It does not, and the shortcuts matter more than the loop. Weathering of sedimentary rock straight back to sediment is quantitatively one of the largest arrows on the whole diagram, and it appears on no classroom version.

That metamorphic rock comes from sedimentary rock. Metamorphic rock comes from whatever was there. The single largest volume of metamorphic rock on the planet has a basaltic igneous protolith and is sitting under the oceans.

That metamorphism means the rock partially melted. It means the opposite. Metamorphism is solid state by definition, and the moment melt appears you have crossed into migmatite and then igneous.

That melting produces magma of the same composition as its source. Partial melting fractionates, always, and this is the single most important fact in igneous petrology. It is why the crust is felsic and the mantle is not.

That pressure turns sediment into rock. Cementation does, through a pore fluid, which is why marine carbonate can lithify in decades and quartz sand takes tens of millions of years under identical burial.

That the cycle is closed. It is a budget with leaks into the mantle and inputs from it, running on two different energy sources at two very different scales.

Mistakes that catch people out

Reading bedding as foliation. A layered sedimentary rock is not foliated. Foliation is mineral alignment produced by stress, and the test is whether the fabric is defined by aligned platy minerals or by compositional layers that were deposited. Where a rock has both, they usually intersect at an angle, and that angle is itself a useful structural measurement.

Testing acid on a weathered surface. Case hardened carbonate can give almost no reaction. Break the sample and test the fresh face. If you still get nothing, powder a small area with the point of a knife and try again, because that is the dolostone signature and it is diagnostic.

Calling anything dark and fine grained basalt. Plenty of fine dark rocks are not basalt: andesite, some hornfels, fine grained diabase, dark shale and slate all get misidentified this way. If it has a fabric it is not basalt. If it fizzes it is not basalt. If it is genuinely too fine to see minerals and shows vesicles or a chilled margin, then you are on solid ground.

Using grain size to estimate cooling time. Covered above. Grain size records the environment, not the clock.

Assuming a granite is granite. Most rocks sold and described as granite are not granite in the IUGS sense. Commercial granite includes granodiorite, tonalite, monzonite, gabbro and gneiss. If the distinction matters for what you are doing, plot it on QAPF rather than trusting the label.

Trusting a key you answered by guessing. Every question you answer wrongly removes the correct rock permanently, and a key cannot detect that it has happened. It will still hand you a confident answer. Skip anything you are not sure of.

The numbers behind the budget

Treating the cycle as an accounting problem rather than a picture makes several things obvious that the picture hides.

Rivers deliver roughly 20 billion tonnes of sediment to the oceans each year, although the pre human figure was lower and the modern one is complicated by dams trapping a large fraction of it before it arrives. Continental denudation averages something like 30 to 60 metres per million years globally, with the spread by setting covering two orders of magnitude either side.

New oceanic crust is created at about 3 cubic kilometres a year and destroyed at very nearly the same rate, which it has to be, because the Earth is not changing size. Sediment entering trenches runs somewhere between 1.5 and 2 cubic kilometres a year, and that number varies by a factor of two between published compilations, so treat it as an order of magnitude rather than a measurement.

Arc magmatism adds new crust at roughly 1 cubic kilometre a year. Set against the erosional flux, that is the fundamental tension in the whole system: the surface is being removed faster than the deep processes replace it, and the continents persist because of isostasy and because most eroded material is redeposited on continental crust rather than being lost to the oceans.

The residence times are where the skew shows. Oceanic crust averages 60 to 65 million years old and nothing survives past about 200 million, because it all gets subducted. Continental crust contains material more than 4 billion years old. Sedimentary rock has a mean age of order 100 million years, but that mean is generated by a distribution with a very long tail, because the probability of a sedimentary rock being destroyed is roughly constant per unit time, which produces exponential survival.

That last point is worth pausing on. It means the sedimentary record is not a uniform sample of Earth history. It is heavily weighted toward the recent, and it becomes sparse rapidly as you go back. Any statement about how conditions changed over geological time has to account for that bias before it means anything.

What this means if you are not a geologist

If you are specifying materials, the rock cycle is what determines the properties you care about, and the connection is usually direct.

Grain size and cement type set the strength and porosity of a sandstone, and both come from its diagenetic history. A quartz cemented arenite is strong and impermeable. The same sand with a clay matrix, deposited rapidly as greywacke, is weaker and behaves differently when wet. Same starting material, different pathway, different engineering rock.

Foliation is a mechanical weakness plane. A schist or slate is markedly weaker across the foliation than along it, and that anisotropy is often a factor of two or more in strength. Orient a tunnel or a slope against it at your peril. This is also why slate makes good roofing and terrible aggregate.

Carbonates dissolve. Limestone and marble are chemically attacked by anything acidic, including rainwater over long enough periods, which is where karst comes from and why a limestone foundation needs a cavity survey. The same reaction that makes them fizz in the acid test is the one that undermines them in the ground.

Reactive silica causes concrete to fail. Chert, opal, volcanic glass and strained quartz can react with the alkalis in cement to form an expansive gel. Knowing that your aggregate source is a chert bearing gravel rather than a clean quartzite is the difference between a structure that lasts and one that cracks in twenty years.

Quartz content drives wear. Every drilling and excavation cost estimate depends on how abrasive the rock is, and abrasiveness tracks quartz content closely. This is exactly what equivalent quartz content and the Cerchar index quantify, and it is why identifying the rock correctly is the first step in any tunnelling cost model.

Common questions about the rock cycle

What are the three types of rock?

Igneous, sedimentary and metamorphic. Igneous rock crystallises from molten material, either underground as plutonic rock or at the surface as volcanic rock. Sedimentary rock forms from fragments, chemical precipitates or organic material that accumulate and are cemented together. Metamorphic rock forms when any existing rock is changed in the solid state by heat, pressure or chemically active fluids, without melting.

What is the rock cycle in simple terms?

It is the set of processes that convert rock from one type into another and move material between the three families. Melting and crystallisation make igneous rock. Weathering, transport, deposition and cementation make sedimentary rock. Heating and burial without melting make metamorphic rock. The important qualifier is that these processes do not run in a fixed order, and rock can skip stages or repeat one many times.

Does the rock cycle have a beginning or an end?

No. It has no start point and no fixed direction. Any reservoir can feed any other that it has a pathway to, and the tool on this page shows eleven such pathways among five reservoirs. The nearest thing to a beginning is the mantle, which supplied the original material, and material is still being exchanged with it in both directions today.

How long does the rock cycle take?

There is no single figure, and the range spans about ten orders of magnitude. Lithification of marine carbonate can happen in decades. A basalt erupted at a mid ocean ridge is typically subducted within 60 to 180 million years. Cratonic gneiss can sit unchanged for two to three billion years. The average residence time of sedimentary rock is of order 100 million years, but averages hide the fact that the distribution is extremely skewed.

What is the difference between magma and lava?

Magma is molten rock below the surface, and it usually contains dissolved gases and suspended crystals. Lava is magma that has erupted. The distinction matters because the loss of dissolved gas on eruption changes viscosity and crystallisation behaviour, so the same melt produces different textures depending on which side of the surface it froze on.

Can a metamorphic rock become a sedimentary rock?

Not directly, and this is the arrow people expect that does not exist. A metamorphic rock has to be exhumed and weathered to sediment first, and only then can that sediment be deposited and lithified. There is no process that converts metamorphic rock into sedimentary rock in place, because sedimentary rocks are defined by having been transported and deposited.

What is the hardest rock?

Rock hardness is not a single defined property the way mineral hardness is, because a rock is an aggregate. In practice the hardest common rocks are those made almost entirely of quartz, so quartzite and chert, both effectively Mohs 7. Some metamorphic rocks such as eclogite are harder to break but not harder to scratch, which is toughness rather than hardness, and the two are separate properties.

How do you tell limestone from dolostone?

Cold dilute hydrochloric acid. Limestone fizzes vigorously on a fresh surface. Dolostone gives little or no reaction until you powder it, and then it fizzes. That is the whole test, and it works because the dolomite reaction rate depends on surface area in a way the calcite reaction does not. Both rocks otherwise look nearly identical in hand specimen.

What rock is formed by heat and pressure?

Metamorphic rock, but "heat and pressure" is a description rather than a mechanism. What actually happens is a series of specific mineral reactions at defined conditions. A mudrock passes through slate near 200 degrees, phyllite near 400, schist near 500 to 600, and gneiss above about 650, and each of those steps is a set of named reactions rather than a vague transformation.

Is granite igneous or metamorphic?

Granite is igneous, crystallised from a silica rich melt at depth. The confusion arises because granitic gneiss looks similar and because granite and gneiss can occur together in the same terrane, with gneiss sometimes derived from granite. The test is fabric. Granite has randomly oriented interlocking crystals. Gneiss has segregated light and dark bands.

What is the difference between intrusive and extrusive rock?

Intrusive rock crystallises below the surface and is generally coarse grained because few crystals nucleate and each grows large. Extrusive rock erupts and freezes quickly, so it is fine grained or glassy. The textural difference is reliable. The usual explanation for it, that slow cooling produces large crystals, is not the actual mechanism, as covered above.

Where does most metamorphic rock form?

At mid ocean ridges, not in mountain belts. About 3 cubic kilometres of new oceanic crust forms each year and nearly all of it is hydrothermally metamorphosed to greenschist facies within a few million years. By volume this exceeds all collisional metamorphism combined, and it appears on almost no rock cycle diagram.

What is QAPF used for?

QAPF is the IUGS scheme for naming igneous rocks from their modal mineralogy. You count quartz, alkali feldspar, plagioclase and feldspathoid, recalculate to 100 percent, and plot on a double triangle. It requires countable minerals, so it applies to plutonic rocks and coarse volcanics. Fine grained and glassy rocks are classified chemically instead.

Why does the same magma make different rocks?

Because the melt changes as it crystallises. Early formed crystals are richer in magnesium, iron and calcium, and if they settle out or are otherwise separated, the remaining liquid becomes progressively richer in silica, sodium and potassium. One parent magma can therefore produce a series of rocks. The magma also assimilates wall rock and can mix with other melts on the way up.

What drives the rock cycle?

Two independent energy sources. Earth's internal heat, about 47 terawatts, drives plate motion, metamorphism and melting. Solar energy, about 174,000 terawatts, drives weathering, erosion and transport. The surface half of the cycle runs on roughly 3,700 times more power than the deep half, which is why erosion can outpace mountain building.

Is the rock cycle still happening?

Continuously. Lava is erupting somewhere at all times, rivers are moving several billion tonnes of sediment to the oceans annually, subduction is carrying crust into the mantle at a few square metres per second globally, and metamorphic reactions are running at depth under every mountain belt. None of it stops.

Related Calculators and Reference

• Pangea Map and Continental Drift Reconstruction puts the tillites, coals and red beds of the cycle back at the latitudes they formed at.
• Subduction Zone Explorer and Megathrust Calculator works the descending limb of the cycle in detail, including the blueschist and eclogite path and where arc magma actually comes from.
• Seafloor Spreading Calculator and Magnetic Stripe Explorer works the mid ocean ridge end of this cycle in detail, with the magnetic record that dates the crust and the depth model that tracks it as it cools.
• Interactive Geological Time Scale puts every process on this page on a dated axis, which is the other half of the picture once you know what happened.
• Mohs Hardness Scale is the scratch test the key leans on, with 229 minerals and materials and the field references that bracket them.
• Rock Density Calculator turns a modal mineral estimate into a bulk density, so the QAPF percentages you read off here feed straight into it.
• RQD Calculator takes you from rock type to rock mass quality, which is where identification starts paying for itself in ground engineering.
• Sediment Settling Velocity Calculator quantifies the transport step between weathering and deposition, the arrow this page treats qualitatively.
• USCS Soil Classification Tool handles the material once weathering has gone all the way to soil.
• International Union of Geological Sciences is the body behind the QAPF scheme and the formal rock nomenclature this classifier follows.