How are fossils dated?

Palaeontologists do not normally point an instrument at a dinosaur bone and read its age. Most fossils have lost the original material needed by common radiometric methods. Instead, investigators date geological events around the find, establish the order of the rocks, and test whether several independent clocks tell the same story.

The result may be a numerical age, such as 75.4 million years, or a bounded interval between two dated ash beds. Both are useful only when the sampled material is demonstrably related to the fossil. A precise laboratory number cannot rescue a sample whose geological context is misunderstood.

First ask what event is being dated

A number always belongs to an event. Uranium–lead dating of a zircon crystal records when that crystal formed, usually as magma cooled or volcanic ash crystallised. Argon dating can record cooling or eruption. Luminescence may estimate when mineral grains were last exposed to sunlight or heat. None of these events is automatically identical to the death of an animal.

A dinosaur could die on a floodplain, be buried soon afterwards, and lie between eruptions that left ash below and above its horizon. Dating those two ashes does not date the bone directly. It establishes that burial occurred after the lower eruption and before the upper one. The distinction is simple, but it prevents a laboratory age from being assigned to the wrong biological event.

Context is more valuable than an isolated bone

Before a fossil is lifted, a field team records its exact level, orientation, surrounding sediment, nearby structures and relation to mapped beds. These observations determine whether the specimen belongs where it was found. A bone weathering out on a slope may have fallen from a higher layer. A river may erode an older deposit and redeposit its fossils inside younger sediment.

This is why a detached specimen with no locality record may be anatomically valuable yet almost useless for precise dating. The critical evidence is not only the object but its connection to a measured section. The same principle applies when reconstructing how a fossil formed: transport, burial and later erosion can separate an organism from the layer in which it originally fossilised.

Relative dating establishes the order of events

In an undisturbed stack of sedimentary rocks, lower beds were deposited before the beds above them. This principle of superposition gives a relative sequence without assigning years. Cross-cutting relationships add more clues: a fault or intrusion must be younger than the rocks it cuts, while an eroded surface is older than the sediment laid over it.

Real outcrops are rarely perfect layer cakes. Folding can tilt or invert beds. Faults can repeat or remove part of a sequence. Burrows mix sediment across a boundary, roots penetrate older layers, and erosion creates gaps representing unrecorded time. Geologists therefore use sedimentary structures, regional mapping and multiple exposures to decide which way was originally up and whether the sequence is continuous.

Biostratigraphy correlates rocks with fossils

Some organisms were geographically widespread but existed for a relatively short interval. Their fossils can act as index fossils. If the same diagnostic species appears in two separated sections, the fossil-bearing beds may be broadly equivalent in age even when their rock types differ.

Biostratigraphy is correlation, not a universal stopwatch. A species may appear earlier in one region than another, survive locally after disappearing elsewhere, or be reworked into younger sediment. Good correlations use assemblages rather than a single specimen and are calibrated where possible against independently dated volcanic material. Marine microfossils are especially useful because they can be abundant and widespread.

U–Pb zircon dating measures mineral crystallisation

Zircon crystals incorporate uranium when they form but strongly exclude lead. Over time, uranium isotopes decay through known chains into lead isotopes. Measuring parent and daughter isotopes allows a crystallisation age to be calculated. Two decay systems within the same crystal provide an internal check, and resistant zircon can retain its record through long geological histories.

Volcanic ash is particularly valuable. Tiny zircons formed in magma may be dispersed across a wide region during an eruption and deposited almost instantaneously by geological standards. However, not every grain dates the eruption. Ash can contain older inherited zircons, while later alteration can disturb part of the isotopic system. Analysts examine individual grains, textures and concordance rather than averaging every measurement without interpretation.

Ar–Ar dating records volcanic and thermal events

Potassium-40 decays to argon-40. The argon–argon method compares isotopes released from a mineral after irradiation and commonly dates volcanic feldspars or micas. It is widely used for Mesozoic volcanic rocks and ash beds that lack suitable zircon.

Argon is a gas, so heating and alteration matter. A later thermal event may let some argon escape and partly reset the clock. Excess argon can produce an apparently old result. Step-heating experiments release gas in stages and help reveal whether a mineral retained a consistent age signal. As with U–Pb dating, the quoted number is accepted only after mineral behaviour and geological setting agree.

Two ash beds can bracket a fossil

Suppose a fossil layer lies above ash dated to 101.6 ± 0.1 million years and below ash dated to 101.2 ± 0.1 million years. If the section is intact and the ashes represent primary fallout, burial occurred within that interval. The fossil need not be chemically datable for its age to be constrained closely.

A single dated layer can still set a maximum or minimum age. A primary ash below the fossil means the burial is younger than the eruption. An ash above it means the burial is older. The inference fails if ash was redeposited from an older source, if faults juxtaposed unrelated beds, or if the fossil itself was reworked. Field evidence decides whether the bracket is real.

Magnetostratigraphy reads polarity reversals

Earth's magnetic field has repeatedly reversed. Iron-bearing grains in sediment or cooling lava can acquire a remanent magnetisation aligned with the field at that time. Sampling a section at close intervals may reveal a pattern of normal and reversed polarity zones.

The pattern is matched to the geomagnetic polarity time scale, but the match needs an anchor. Several parts of the global sequence may share similar short patterns. Radiometric ages, index fossils or a securely identified boundary narrow the alternatives. Chemical alteration can also replace the original magnetisation with a later one, so laboratory demagnetisation tests whether the signal is primary.

Chemical and orbital signals refine a time scale

Changes in stable isotopes or elemental composition can mark events recognised across distant sections. A carbon-isotope excursion, for example, may help correlate marine and terrestrial records when its shape and stratigraphic position are well established. These signals can be altered by local environments, diagenesis or gaps, so they work best alongside fossils and numerical ages.

Sedimentary cycles may reflect long-term variations in Earth's orbit and axial orientation. Astrochronology matches repeated bedding or chemical patterns to calculated orbital rhythms, potentially dividing a sequence more finely than isolated dates. It requires continuous deposition and a defensible link between the observed cycle and an astronomical driver. A visually regular banded rock is not automatically an orbital calendar.

Carbon-14 is only for young organic remains

Radiocarbon dating measures carbon-14 left in once-living material. Its half-life is about 5,730 years, which makes it useful for archaeology and late Quaternary palaeontology but not for non-avian dinosaurs. After roughly 50,000 years, the remaining original isotope is so scarce that tiny modern contamination dominates the measurement.

Non-avian dinosaurs disappeared 66 million years ago, more than a thousand times beyond the practical radiocarbon range. A carbon-14 signal reported from a dinosaur fossil would indicate contamination, younger material, instrument background or a mistaken context, not surviving Mesozoic carbon. Dinosaur-bearing rocks are dated with long-lived isotope systems and stratigraphic correlation instead.

Other methods date deposition or later change

Optically stimulated luminescence estimates when quartz or feldspar grains were last exposed to light before burial. Electron spin resonance measures trapped charges in minerals, including tooth enamel. Uranium-series techniques track disequilibrium among uranium decay products and can be useful for cave deposits, carbonates and relatively young fossil contexts.

Each method has a particular clock, usable age range and vulnerability to environmental change. Uranium may enter a tooth after burial. Luminescence grains may not have been fully reset by sunlight before deposition. Electron spin resonance depends on the radiation dose received through time. These tools are powerful when their assumptions are tested, not interchangeable substitutes for any desired age.

Uncertainty is part of the result

An age written as 100.0 ± 0.2 million years is not a disguised exact date. The uncertainty summarises analytical precision under stated assumptions, often at a defined confidence level. Additional uncertainty can come from decay constants, calibration standards, sample alteration and the geological link between the dated mineral and the fossil.

Precision and accuracy are different. A machine may measure isotope ratios very precisely in a reworked crystal that is millions of years older than the ash deposit. The resulting number can have a small analytical error yet answer the wrong question. Repeated analyses, reference standards and independent methods expose this kind of false precision.

Why published ages are revised

Dates change when laboratories improve calibration, analysts remove contaminated or inherited grains, mapping reveals an unnoticed fault, or a formation is correlated differently. A newer age does not necessarily mean the earlier scientists were careless. It may result from higher precision or from new evidence that tests an assumption unavailable to the first study.

Revisions can also move a fossil from one stage of the geological time scale to another when formal boundary ages are recalibrated. The rock and fossil have not moved in time. The numerical framework used to describe their position has become better constrained.

How to read an age in a dinosaur article

Look for four details: what material was measured, which method was used, where the dated sample sits relative to the fossil, and what uncertainty or interval was reported. “About 76 million years old” may summarise a robust bracket, a stage-level correlation or merely an old regional estimate. The underlying evidence determines how much weight the number deserves.

A range such as 75–71 million years can describe the known duration of a formation or taxon, not the lifespan of one individual. A single skeleton may occupy one horizon within that range. Likewise, a dinosaur profile that says “Late Cretaceous” communicates a broader temporal placement than a directly bracketed locality. Both can be accurate at different levels of resolution.

It is also worth checking whether an age refers to the original study or to a later recalibration. Older publications may use superseded decay constants or an earlier version of the time scale. Modern authors sometimes recalculate the same laboratory measurements so that several localities can be compared consistently. That changes the numerical expression of the evidence without inventing a new fossil horizon.

The short answer

Fossils are dated by reconstructing their geological context. Rock order establishes sequence, isotope systems date suitable minerals, and fossils, magnetic reversals and chemical signals correlate one section with another. The strongest age is not the most impressive solitary number but the interval supported by independent evidence that all refers to the same event.