A crystal can be older than the rock that carries it. Two minerals from the same mountain can give different ages and both be scientifically useful. An instrument may measure an isotope ratio precisely while leaving the most important question unanswered: what happened at the time represented by that result?
Geochronology establishes numerical ages and rates for geological events. Its gold standard joins clock physics, material history, field relationships and uncertainty in one defensible account. The task is to date an event, explain why the material records it and recognise what later processes may have changed.
The original examples below use deliberately simplified clocks and fictional rock sequences. They are teaching cases, not laboratory determinations or revisions to the geological timescale. They follow The Gold Standard of Education: learning that transfers into life: understand a method, use it and check what changes when the setting changes.
Find your next question
Begin with the route that fits your purpose. Read sequentially for the full argument, or use the chapter index to return to a worked example, an evidence check or a connected guide.
Full chapter index · Geochronology: mechanism library · How X Works
Chapter index
Understand the question · Chapters 1–3
Build the mechanisms · Chapters 4–7
Work through evidence · Chapters 8–11
Use and check the ideas · Chapters 12–15
Transfer and continue · Chapters 16–19
“How old is this rock?” sounds precise until you consider the rock’s history. Its minerals may have crystallised from magma, cooled, been buried, heated, deformed, weathered and redeposited. Different materials and measurements can record different parts of that sequence.
Replace the broad question with an event statement. Do you want the eruption time of a lava, the crystallisation time of a zircon, the deposition time of a sediment or the cooling history of a mountain belt? These are related questions with different evidential requirements.
Field context comes first. Record where the sample sits, what it cuts or overlies, which minerals it contains and whether alteration is visible. A number detached from its sample history loses much of its interpretive value. The laboratory does not automatically restore relationships that were never documented.
USGS’s guide to dating rocks explains why combining crystal-formation information with eruption ages can illuminate the history of a volcanic system. Crystals can grow before eruption. Their ages therefore need not be identical to the time when lava reached the surface.
For a beginner, draw a timeline with separate boxes for formation, transport, deposition and later heating. Put a question mark beside the event of interest. The next job is to choose a clock that might record that event, then test whether its assumptions survived. This sequence makes the eventual arithmetic useful instead of allowing a precise calculation to choose the geological story for you.
Stratigraphy and field relationships establish event order. In an undisturbed sequence, an overlying layer was deposited after the layer below. An intrusion must be younger than the material it cuts. A fragment inside a deposit existed before the deposit incorporated it. These relationships constrain interpretation even before any isotope measurement is made.
Real structures need inspection. Layers can be folded, overturned or faulted, and contacts can represent erosion or long gaps. The useful principle is not to apply a slogan blindly, but to identify the physical relationship that supports the order.
Numerical dating then attaches quantities to suitable events within that framework. A dated ash layer can constrain nearby deposition, while an intrusion can establish that its host was already present. A rock sequence with several independent constraints is usually more informative than a collection of ages whose spatial relationships are unknown.
Consider a sediment between a lower volcanic ash and an upper ash. If the ashes represent primary deposition and their eruption ages are known, the sediment lies between those events. If an ash was reworked from an older deposit, its crystal age may supply only an older bound. The same-looking material has changed its evidential role because its transport history changed.
The broader learning habit is to combine order and quantity. A numerical result should fit the relationships that were independently observed. If it does not, investigate inheritance, disturbance, misidentified contacts or analytical problems. The disagreement can reveal geological information; it is not a reason to erase whichever observation is less convenient.
Radioactive decay is unpredictable for an individual atom but follows a stable statistical law for a large population in the ordinary conditions used by geological clocks. If N₀ parent atoms were present at the clock’s start and the system remains closed, N(t) = N₀ exp(−λt), where λ is the decay constant.
The half-life H is related by λ = ln(2)/H. After one half-life, half the parent remains; after two, one quarter remains; after three, one eighth remains. The fraction lost per equal interval is consistent, while the absolute number lost declines as fewer parents remain.
Use a fictional isotope with half-life 50 million years. If the present parent population is one quarter of a securely known initial population, two half-lives have elapsed. The model age is 100 million years. This is an arithmetic illustration, not the identification of a real isotope system.
Notice the strong assumption hidden in “securely known”. For a natural mineral, the initial parent abundance is usually not directly observed. Dating methods use the relationship between remaining parent, accumulated daughter and other isotope measurements rather than pretending to possess a photograph of the starting atom count.
The half-life is not the age of the specimen. It is a property of the decay process that helps infer elapsed time under a model. Memorising a list of half-lives without understanding that distinction can make a learner choose a number by association rather than follow the physical argument.
In the simplest closed system, each decayed parent contributes one atom of the radiogenic daughter being counted. Let P be present parent atoms and D* the daughter atoms produced by this decay since the clock began. Then P + D* represents the initial parent population, and D*/P = exp(λt) − 1.
Rearranging gives t = ln(1 + D*/P)/λ. Using half-life H, the same relation is t = H × log₂(1 + D*/P). Writing both forms helps reveal that the calculation is a ratio-based population argument rather than a special button on a laboratory instrument.
For a constructed example, H = 80 million years and D*/P = 3. Then 1 + D*/P = 4, log₂(4) = 2 and t = 160 million years. At that time, one quarter of the original parent remains and three quarters has become radiogenic daughter. The atom accounting and exponential calculation agree.
The star on D* matters. It means radiogenic daughter attributable to the relevant decay after the clock began. Total measured daughter may include an initial component or contamination. Substituting it without a justified correction can produce an age with no defensible relation to the intended event.
Check units at the end. If H is in millions of years, t is too. The logarithm uses a dimensionless ratio. If a calculation takes the logarithm of an unnormalised quantity with units, stop and inspect the equation before accepting the output.
A mineral does not begin with an abstract zero in every daughter-isotope box. Its initial composition depends on its chemistry and surroundings. A useful dating system must either make the initial component small enough to constrain or provide measurements that help estimate it.
Zircon is valuable in uranium–lead dating partly because its structure commonly incorporates uranium while initial common lead is often low. That is a favourable material property, not a licence to assume every grain contains no initial lead or experienced no later disturbance.
Other methods use multiple samples or minerals to estimate initial isotope composition and age together. In a simplified isochron relationship, a normalised daughter ratio is plotted against a normalised parent ratio. For coeval materials sharing an appropriate initial composition and remaining closed, the slope depends on elapsed time.
A line is evidence to interpret, not an automatic age certificate. Mixing unrelated components can also generate linear patterns. Examine whether the materials genuinely belong to the same event, whether the common-initial-composition assumption is plausible and whether geological observations support the result.
This is the first major step beyond the half-life exercise. Real measurement does not abandon the decay law; it adds a justified account of the starting conditions and material behaviour. A strong explanation makes those additional steps visible. The learner should be able to say which quantity was measured, which quantity was inferred and which observations defend the inference.
A mineral can exchange relevant elements with its surroundings. At sufficiently high temperatures, diffusion may allow daughter products to escape or redistribute. As a system cools, retention can become effective enough for a dating clock to record a cooling-related event.
“Closure temperature” is a useful model concept, but it is not one universal thermostat setting attached permanently to a mineral name. Grain size, diffusion geometry, cooling rate, damage and the isotope system influence retention. The interpretation needs a physical model suited to the material.
Imagine minerals A and B with different retention behaviour in a fictional cooling rock. A records 90 million years and B records 60 million years. The difference may reflect passage through different temperature windows. It does not require that one laboratory result be wrong, and it does not mean that the entire rock formed twice in the same sense.
Later heating may partly or fully reset a clock. Partial resetting can produce mixed or distributed signals that require analysis rather than a single convenient average. The thermal history may be more complex than steady cooling from one event.
Thermochronology uses such behaviour to investigate cooling histories. Converting cooling into uplift or erosion additionally requires information about temperature structure, burial and heat sources. A cooling age alone is not a direct altitude measurement. This boundary between observation and interpretation is central to professional practice: explain the event a clock can record before using it to infer a larger landscape story.
Sediments are assembled from materials with earlier histories. A zircon eroded from an old rock can travel through a river and become part of a younger sandstone. Its crystallisation age then records its earlier source history, not the moment the sandstone was deposited.
Suppose a deposit contains grains with model ages of 900, 650 and 310 million years. If the 310-million-year result is a reliable source-crystallisation age, deposition must have occurred after that crystal existed. It supplies an older bound on deposition, often described as a maximum depositional age in age-before-present terms.
That phrase can be confusing. A maximum age of 310 million years means deposition is no older than 310 million years. It may be substantially younger. It does not establish that deposition happened exactly 310 million years ago, and it does not mean the deposit’s true age must be numerically greater.
In practice, geologists assess populations, analytical uncertainty, grain textures and possible disturbance rather than automatically treating the youngest single measurement as decisive. A spuriously young analysis or later growth can mislead if its geological meaning is ignored.
Try the transfer question: if a 120-million-year intrusion cuts the sandstone, what follows? The sandstone existed before the intrusion, so it is older than 120 million years. Together, the two constraints place deposition between 310 and 120 million years ago under the stated assumptions. The result is a bracket, which is more honest than selecting an unsupported point in the interval.
Dating methods differ in what they measure and what can reset their signal. Uranium–lead systems can investigate mineral growth and complex older histories. Potassium–argon and argon–argon methods can constrain suitable potassium-bearing materials and events, including volcanic histories when the sample and retention assumptions support that interpretation.
Radiocarbon concerns carbon-bearing material and comparatively recent timescales. It is not a general method for measuring the crystallisation age of an ancient granite. Its useful application also requires an account of carbon source, contamination and calibration.
Luminescence methods investigate accumulated signals in suitable minerals and the resetting associated with light or heat. Cosmogenic-nuclide methods can investigate exposure or burial histories under production and transport models. These are different clocks with different starting conditions, not interchangeable routes to one universal “age”.
Begin method selection with a table containing the target event, suitable material, measured signal, reset mechanism and main ambiguity. Then inspect whether the sample meets those conditions. An impressive analytical precision cannot repair a mismatch between the event and the material.
The USGS Geochron database description illustrates the breadth of methods used in geochronological and thermochronological work. A database record is most useful when its method and sample context travel with the numerical age. Comparing numbers while stripping away those distinctions can produce a false disagreement or an unjustified combined chronology.
A measured radiocarbon age is not automatically the same as a calendar age. The carbon-isotope composition of the relevant reservoir has changed through time, so calibration uses independently constrained records to relate radiocarbon determinations to calendar time.
The published IntCal20 Northern Hemisphere calibration study is a methodological reference. Appropriate calibration depends on the sample’s reservoir and context. Material that acquired carbon in a marine environment requires a suitable marine treatment rather than automatic use of a terrestrial atmospheric curve.
A calibration curve need not be a straight line. A measured value and its uncertainty can correspond to a spread of calendar possibilities, sometimes with more than one interval. Reporting only a transformed central number can hide that structure.
The same general lesson applies beyond radiocarbon. Instruments, reference materials, decay constants and standards connect measured signals with inferred ages. Their uncertainties and shared dependencies matter when comparing results between laboratories or methods.
For a beginner’s analogy, imagine converting a historical price into purchasing power using a changing index. The conversion depends on the reference relationship and its context. That analogy explains why calibration exists; it is not a substitute for radiocarbon physics. The transferable habit is to identify the reference, state the convention and carry its uncertainty into the result instead of treating conversion as a cosmetic change of units.
Consider a fictional landscape. A sedimentary unit lies above an ash dated at 24 million years and below an ash dated at 20 million years. Assume both ashes were deposited at eruption and the sequence is correctly interpreted. Deposition of the intervening unit occurred between those times.
If the unit is 200 metres thick, dividing by four million years gives 50 metres per million years. That is an average based on using the full bounding interval. It does not demonstrate constant deposition, and the unit may occupy only part of the interval.
If deposition lasted two million years instead, the average would be 100 metres per million years. Erosion, compaction and hiatuses introduce further distinctions between preserved thickness, original accumulation and active depositional rate. The arithmetic is easy; choosing what the denominator represents is the scientific work.
Now add a dated horizon inside the unit. It can help divide the interval and test whether accumulation was uniform. It does not automatically justify a straight line through every depth. Examine sedimentary evidence for gaps or abrupt changes before interpolating.
This matters in climate records, landscape evolution and hazard histories. An age model translates depth or sequence into time, so conclusions about rates depend on its structure. Always ask which ages anchor the model, what lies between them and how uncertainty changes away from the anchors. A neat time axis is an inference that deserves inspection, not a neutral decoration beneath a graph.
Analytical uncertainty describes only part of an age interpretation. Counting statistics, instrument behaviour, blanks, calibration and reference values can contribute. Geological interpretation adds other questions: inheritance, alteration, mixed growth zones, reworking and incomplete resetting.
When a result is written as 100 ± 2 million years, ask what the interval means. Is it one standard deviation, another stated confidence convention or a different uncertainty measure? Does it include external reference uncertainty? Without that definition, comparing it with another quoted interval can be misleading.
Shared errors are especially important for differences. Suppose two fictional horizons are dated at 100 and 96 million years. Their difference is four million years. If each has an independent one-standard-deviation error of one million years, the difference has standard deviation √(1² + 1²), about 1.41 million years under the simple independent model.
If both also share a common calibration offset, that component may partly cancel in the difference while still affecting absolute ages. If errors are correlated in another way, the result changes. Do not add or combine uncertainties without identifying their dependence.
An uncertainty interval is not a confession of failed science. It identifies the resolution of the claim. A well-supported broad bracket may be more useful than a narrow number attached to the wrong event. The gold standard preserves both numerical uncertainty and geological meaning so a reader can decide what comparisons the evidence actually supports.
Independent methods can strengthen a chronology when they address compatible events and share few hidden dependencies. Agreement is useful because each method has its own vulnerabilities. However, matching numbers alone does not prove that every assumption is correct.
Suppose a volcanic crystal-growth age predates an eruption age. That can represent genuine residence before eruption. Averaging them into one “best age” would erase the history the measurements were capable of revealing. First ask whether the dates are expected to agree.
Likewise, a broad spread of grain ages can record multiple source regions or episodes of growth. A weighted mean is appropriate only when a defensible common-age model applies and the uncertainty structure is handled correctly. Small analytical errors do not make a mixed population become one event.
Outliers need explanation rather than automatic removal. Inspect textures, chemistry, field setting and analytical quality. A result might be contaminated, inherited or evidence for a previously unrecognised event. Document exclusion criteria so another researcher can see whether the conclusion depends on selecting agreeable data.
The practical reporting pattern is simple: target event; materials and methods; results with uncertainty; interpretation; alternatives; and the observation that would discriminate between them. This makes disagreement productive. It also gives collaborators a clear next investigation instead of a vague instruction to obtain more dates until the picture looks tidy.
A geochronological result begins in a particular place. Preserve location, sample identity, geological relationships and the reason for collection. Through preparation and analysis, keep the connection between the measured fragment and the original material explicit.
Images and descriptions of mineral domains help interpret growth and alteration. Analytical records should identify standards, corrections, uncertainty conventions and processing choices. Raw observations and interpreted ages serve different purposes; retain enough information for meaningful review and reuse.
Collaboration is central. A field geologist may recognise a contact that changes the event interpretation. A mineral specialist may distinguish inherited cores from later rims. An analytical specialist may identify a correction or detection problem. A modeller may reveal that an apparently precise age-depth curve depends heavily on one assumption.
Reproducibility does not mean every laboratory must produce identical last digits. It means methods and provenance are sufficiently clear to understand differences and test the account. Public databases can support this work when entries retain the details needed to connect numbers with material histories.
The transfer into life is a disciplined reading of timestamps. An inscription’s date, a document’s revision date and a file’s copying date answer different questions. Geological clocks make the same general distinction unusually visible: a timestamp belongs to an event under a method. Before using it in a larger argument, establish what started it, what could reset it and how the link to the event was verified.
CHAPTER 14 OF 19
14. Teach the move from half-life exercises to geological reasoning
Start with repeated halving and a labelled population. Then introduce daughter accumulation and derive the ratio equation. Only when the learner can explain the conservation should a calculator replace the simple fractions.
Next, change one assumption at a time. Add initial daughter, allow loss of daughter or place an old crystal inside a younger deposit. Ask whether the earlier calculation remains valid. This is more revealing than assigning ten identical equations with different numbers.
A useful error log separates mathematical errors from interpretive errors. Confusing a logarithm with a linear proportion requires one repair. Treating a crystal-growth age as a depositional age requires another. Correct arithmetic can coexist with a wrong geological conclusion.
For independent practice, give a short field description with two possible materials. Ask the learner to choose which might address the event and identify the missing information. A complete response may be “this provides a bound, not an exact age”. That answer demonstrates control rather than failure to finish.
Later, combine methods in a small timeline and ask which results should agree. Let the learner defend one uncertainty statement and one alternative explanation. The intended outcome is someone who can read an age in a science report and ask useful questions about it. They need not operate a mass spectrometer to understand that the number, material and event must remain connected.
First, a fictional closed system has half-life 40 million years and radiogenic daughter-to-parent ratio 7. The age is 40 × log₂(8) = 120 million years. Three half-lives leave one eighth of the initial parent and produce seven eighths radiogenic daughter under the stated one-to-one model.
Now suppose part of the measured daughter was present initially. Using total daughter as D* would overstate the radiogenic contribution and can overestimate the model age. An initial-component correction is required before the same equation can answer the intended question. The amount of the correction cannot be guessed from the final ratio alone.
Second, a sandstone contains a reliable 180-million-year source crystal and is cut by a 110-million-year intrusion. Deposition occurred after the crystal formed and before the intrusion, giving a bracket between 180 and 110 million years ago. The sandstone is not demonstrated to be exactly as old as either boundary event.
Third, two minerals record 70 and 45 million years under different thermal-retention systems. You cannot infer a 25-million-year eruption duration from that subtraction. Establish which events they record. If they constrain different cooling stages, their difference concerns a thermal history, with further modelling required.
For each case, write a final sentence naming the dated event and one limiting assumption. That sentence is part of the answer. It is the bridge between mathematics and geological interpretation, and it protects the result when someone later copies the number into a table, map or public explanation.
Subtracting two dates appears straightforward, but a geological duration requires that the dated events actually bound the process. Crystal growth before eruption and cooling afterwards do not automatically mark the beginning and end of eruption itself. The timeline needs event labels before subtraction.
Suppose a fictional intrusion crystallises at 130 million years and a lower-temperature system records 100 million years. The thirty-million-year difference is compatible with a long cooling history under a suitable interpretation. It is not evidence that magma was continuously erupting for thirty million years. Different physical histories can connect those endpoints.
Now suppose two ash beds demonstrably bracket sediment accumulation. Their ages can bound the elapsed interval, but gaps in deposition may mean that active accumulation occupied only part of it. A rate calculated with the full interval is a long-term average over the defined history, not necessarily the rate during deposition.
The next useful observation depends on the ambiguity. Additional thermal systems may help reconstruct cooling. More dated horizons may constrain an age-depth model. Field evidence may identify a hiatus. Collecting another highly precise date of the same unsuitable event will not necessarily answer the missing question.
This is an important professional judgement: precision is valuable only after relevance is established. Explain which events bound the interval, how their uncertainties relate and which unobserved periods may lie between them. A duration then becomes an inspectable inference rather than a subtraction whose geological meaning was assumed.
A compact guide to the dated event
| Evidence | Possible interpretation | Question to resolve |
|---|---|---|
| Mineral-growth age | Crystallisation episode | Is the measured domain inherited or new? |
| Retention-related age | Cooling or later thermal change | Which thermal model applies? |
| Detrital grain age | Source history and depositional bound | Was the grain reworked or disturbed? |
| Primary ash age | Eruption-linked horizon | Was the ash deposited at eruption? |
| Radiocarbon result | Carbon-system time estimate | Which reservoir and calibration apply? |
The table describes conditional interpretations. The material’s actual history determines which one is justified.
Is radiocarbon used to date all rocks?
No. Radiocarbon addresses suitable carbon-bearing materials on comparatively recent timescales. Other isotope systems and non-radiometric methods address different materials and events. The sample and question determine the method.
Why can a crystal be older than its host deposit?
The crystal may have formed in an older source rock before erosion, transport and redeposition. Its age can constrain the deposit’s older bound without directly dating deposition. Mineral history and field relationships establish the distinction.
Does a more precise instrument always give a better geological age?
Better measurement can help, but it cannot by itself identify the event. Inheritance, alteration or an unsuitable sample can produce a precise result that does not answer the intended question. Precision and interpretive validity must be assessed together.
What does a maximum depositional age mean?
In age-before-present terms, it is an older bound: deposition should be no older than the reliable constraint. The actual deposit may be younger. State the inequality in words to avoid reversing its meaning.
Can heating reset a clock?
Depending on the mineral, system and thermal conditions, heating can cause partial or complete redistribution. The result may record cooling or later change instead of original crystallisation. The relevant physical model must be examined.
Is the geological timescale just a list of radioactive dates?
No. Geological time integrates relative relationships, numerical constraints and correlation across records. The age of a particular sample and the definition of a time boundary are connected but distinct tasks. Use current authoritative timescale resources for precise boundary values.
Choose a route by the question you want to answer. These connected guides develop the specialist mechanisms and adjacent subjects; they are companions to this explanation rather than substitutes for checking a particular real-world case.
Build a defensible chronology
How Science Works | Geochronology — Isotopic Clocks, Numerical Ages, Closure, Calibration and Measuring Deep Time — Continue into the fuller library of clocks, closure, calibration and event interpretation.
How Science Works | Stratigraphy — Layers, Sequence, Correlation, Unconformities and Reading Earth’s Physical Record — Use layers, contacts, gaps and correlation to establish relative order.
How Science Works | Sedimentology — Weathering, Transport, Deposition, Sedimentary Structures and the Evidence of Ancient Environments — Read transport and deposition evidence before interpreting a sediment’s age.
How Science Works | Petrology — Magma, Sediment, Metamorphism and the Processes That Make Rocks — Connect mineral formation and later alteration with the history of rocks.
Interpret Earth history
How Science Works | Paleoclimatology — Proxies, Ice Cores, Sediments, Tree Rings and Reconstructing Past Climate — Understand how an age model supports a reconstruction of past climate.
How Science Works | Paleontology — Fossils, Deep Time, Evolution, Extinction and Ancient Ecosystems — Place fossil evidence and ancient ecosystems within a chronological framework.
How Science Works | Geochemistry — Elements, Isotopes, Reactions, Reservoirs and Earth’s Chemical Memory — Follow the elemental and isotopic behaviour behind many geological clocks.
How Science Works | Geology — Rocks, Tectonics, Deep Time, Earthquakes and the Dynamic Crust — Return to the broad relationships between rocks, tectonics and deep time.
Continue across Gold Standard subjects
The Gold Standard of Geochemistry provides the connected geochemistry perspective.
The Gold Standard of Sedimentology provides the connected sedimentology perspective.
The Gold Standard of Petrology provides the connected petrology perspective.
The Gold Standard of Palaeoclimatology provides the connected palaeoclimatology perspective.
The Gold Standard of Palaeontology provides the connected palaeontology perspective.
For the wider collection, return to How X Works and How Science Works.
These primary institutional and research sources were checked on 21 October 2026. They support the scientific framing and specialist references; the numerical cases, teaching sequence and interpretation checklists are original educational examples.
USGS: a beginner’s guide to dating rocks — Material, method and event distinctions in volcanic histories.
USGS Geochron database — Geochronological and thermochronological methods, provenance and reuse.
IntCal20 calibration research — Published Northern Hemisphere radiocarbon calibration methodology.
Radiocarbon: IntCal20 calibration issue — Context for distinct atmospheric and marine calibration treatments.
Return to the source appropriate to the specific measurement, method and context before using a classroom model to make a professional decision. Where the article states a conditional example, its conclusion depends on the assumptions explicitly given with it.
Contents · Previous chapter · Continue into the subject mechanisms