Zircon uranium–lead dating estimates the age of a zircon domain from uranium isotopes and the lead produced by their radioactive decay. A defensible result connects the measured domain to a geological event: growth of a crystal, inheritance from an older rock, or a later disturbance. A precise number alone does not establish the age of the rock, its eruption, or its deposition.
This guide takes you from the two decay systems to useful interpretation, with original calculations, a laboratory decision framework and a changed-context exercise. Numerical datasets below are fictional teaching examples, not reported measurements. Begin with Zircon for mineral identity and context. This guide belongs to The Gold Standard and Education: learning that transfers into life; find related mechanisms in How X Works.
Choose your zircon dating route
Chapters 1–3: Understand the clock
Chapters 4–6: Calculate and test
Chapters 7–9: Correct and measure
Chapters 10–12: Interpret and diagnose
Chapter index
Understand the clock · Chapters 1–3
1. Name the event before selecting the clock
Calculate and test · Chapters 4–6
4. Calculate an apparent age and check it forwards
Correct and measure · Chapters 7–9
7. Keep common lead separate from radiogenic lead
Interpret and diagnose · Chapters 10–12
10. Interpret inherited cores and younger growth
Transfer and continue learning · Chapters 13–14
CHAPTER 1 OF 14 · Understand the clock
1. Name the event before selecting the clock
An age measurement answers a narrower question than “How old is this landscape?” The analytical question concerns an isotope system in a particular piece of mineral. The geological question concerns an event in the history of a rock. Connecting them requires observations that are independent of the numerical result.
Consider three samples collected on the same hillside: a granite, a sandstone and a metamorphosed rock. Zircon in the granite might record stages of magma crystallisation. Zircon in the sandstone can have formed long before erosion transported it into that sediment. A zircon in the metamorphic rock can contain an older interior surrounded by younger growth. Giving all three numbers the label “rock age” conceals the distinctions that make the measurements useful.
Before an analysis, complete this sentence: “I want to constrain the time of ___, and I expect this zircon domain to record it because ___.” The second blank might refer to a growth zone, an inclusion relationship, a volcanic ash bed or a field relationship. If the explanation is merely that zircon is datable, the connection remains untested.
Separate three levels of statement
A measured isotope ratio is an analytical observation after calibration and correction. An apparent date is the result of applying an age equation and its assumptions. A geological age interpretation identifies an event that the date represents. These levels should remain distinguishable in notes, figures and prose.
Suppose a laboratory reports an apparent date of 150 million years. That does not yet tell us whether the grain crystallised in a magma at that time, inherited older material, experienced partial disturbance, or combines different domains. The correct first response is to ask which material was analysed and what checks support the interpretation.
This discipline also improves sampling. If the question concerns a short interval between two volcanic layers, sampling random sandstone elsewhere may produce many accurate mineral dates without resolving that interval. Choose evidence for its relationship to the event, not simply for its capacity to produce numbers.
CHAPTER 2 OF 14 · Understand the clock
2. Understand why zircon can preserve a clock
Zircon is a zirconium silicate, commonly written ZrSiO₄. Uranium can enter its structure during growth, while initial lead is often low. Over time, radioactive uranium generates radiogenic lead. The mineral can preserve this parent–daughter information, but its durability is a property to investigate in a real sample rather than an unconditional guarantee.
The basic distinction is between lead produced within the dated system and other lead. An asterisk in Pb* denotes the radiogenic component used in a simplified age equation. “Common lead” means lead that is not attributed to the relevant in-place decay history in that calculation. It must not automatically be counted as daughter product.
A crystal is also not necessarily a single historical unit. An old core can survive while younger zircon grows around it. A crack, inclusion or damaged region can alter what enters an analysis. A spot crossing a boundary can combine material that the geological interpretation needs to keep separate.
Draw the sample before calculating
For a teaching exercise, sketch a circle for an old core and a ring for a younger rim. Place one analytical spot wholly inside each region and a third across their boundary. The first two measurements could answer different growth questions. The third creates a mixing question even if the instrument performs perfectly.
Now add a crack through the rim. The sketch does not prove lead loss, but it identifies a feature that deserves examination. Images, analytical positions and chemical observations provide a way to test possible histories. A spreadsheet of dates without those connections removes much of the evidence needed to interpret the dates.
Use a domain identifier that survives every stage: specimen, grain, image, spot or fragment, result and interpretation. If an analyst cannot reconnect a number to its physical location, the apparent precision of the result cannot repair the missing context. Traceability is part of scientific meaning, not merely administrative organisation.
CHAPTER 3 OF 14 · Understand the clock
3. Derive the age equation from atom accounting
The two principal uranium–lead systems connect ²³⁸U with ²⁰⁶Pb and ²³⁵U with ²⁰⁷Pb through decay chains. For the simplified closed-system treatment, write the remaining parent population as P = P₀ exp(−λt). Here P₀ is the initial parent population, λ is the decay constant, and t is elapsed time.
If the system began with no relevant radiogenic daughter and retained the products, the daughter population is D* = P₀ − P. Rearranging the parent equation gives P₀ = P exp(λt). Substitution therefore gives D*/P = exp(λt) − 1. Calling this isotope ratio R, the inverse relation is t = ln(1 + R)/λ.
The derivation explains a common denominator error. The ratio uses the parent remaining now, not the initial parent. After one half-life, half the original parent remains and half has produced daughter. Daughter divided by remaining parent is one, even though daughter divided by initial parent is one-half.
Units and assumptions
A ratio of isotope atom amounts is dimensionless. If λ is expressed per year, the calculated t is in years. A result in years becomes millions of years, or Ma, by dividing by one million. Confusing per-year and per-million-year constants introduces a factor of a million, not a small rounding error.
For the examples here, use λ₂₃₈ = 1.55125 × 10⁻¹⁰ per year and λ₂₃₅ = 9.8485 × 10⁻¹⁰ per year. These conventional values and the two growth equations are documented in UCL’s U–Pb teaching chapter. The examples ignore decay-constant uncertainty and initial decay-chain disequilibrium to isolate the arithmetic.
That last simplification matters. The equation is not permission to ignore all intermediate nuclides under every condition. A professional interpretation must state which corrections are needed for the age range, precision and material being studied. The simple model is valuable because its assumptions are explicit and its predictions can be tested.
CHAPTER 4 OF 14 · Calculate and test
4. Calculate an apparent age and check it forwards
Suppose a hypothetical domain has a corrected ²⁰⁶Pb*/²³⁸U ratio R = 0.1000. With the stated λ₂₃₈, t = ln(1.1000)/(1.55125 × 10⁻¹⁰). Since ln(1.1) is approximately 0.0953102, the result is approximately 614.4 million years.
This is an apparent date from one isotope system. It becomes a stronger geological interpretation only after the other evidence has been considered. The arithmetic itself is nevertheless easy to audit: insert the calculated t into exp(λt) − 1 and recover approximately 0.1000.
Compare an approximation with the full equation
For sufficiently small λt, exp(λt) − 1 is approximately λt, so t is approximately R/λ. Using R = 0.1000 gives approximately 644.6 Ma. That is about 30.2 million years higher than the logarithmic answer. The approximation is not suitable merely because the ratio looks like a small decimal.
At R = 0.0100, the exact expression gives about 64.14 Ma, while R/λ gives about 64.46 Ma. The absolute discrepancy is smaller, but whether it is acceptable still depends on the question. A rough classroom estimate and a high-precision comparison of geological events have different tolerances.
The direction of the error follows from ln(1 + R) being less than R for positive R. The linear estimate therefore overestimates the exact age in these examples. Knowing this direction gives a useful debugging check: if a program using the stated approximation produces a smaller age, inspect the implementation.
Avoid unjustified digits
A calculator may display many decimal places, but the supplied ratio has limited precision and the real measurement would have uncertainty. Reporting 614.4 Ma is enough for this exercise. Adding a long tail of digits makes the arithmetic look more certain without supplying additional evidence.
For independent practice, choose a hypothetical age of 500 Ma, generate both uranium–lead ratios with the exponential equations, then invert each ratio. Both inversions should return the chosen age, apart from rounding. This is a test of the calculation, not a demonstration that a real sample remained closed.
A hypothetical closed domain of age 1000 Ma gives ²⁰⁶Pb*/²³⁸U of approximately 0.1678 and ²⁰⁷Pb*/²³⁵U of approximately 1.6774 using the constants above. The ratios differ greatly because the decay constants differ. Different isotope ratios do not imply different ages.
A concordia curve collects the paired ratios predicted for successive ages. In a conventional Wetherill diagram, the coordinates are the two radiogenic lead-to-parent ratios. Every point on the ideal curve represents the same elapsed time in both systems. A measured point is compared with the curve together with its uncertainty.
Read the ellipse as part of the result
An analytical result is not an exact dot. Uncertainty in the ratios, including their correlation, determines the shape and orientation of its uncertainty region. A visual impression of closeness to a line can be misleading when the scale changes or the uncertainty is omitted.
Concordance is evidence of compatibility with the model under the chosen treatment. It is not a certificate that the desired geological event has been dated. An inherited grain can be beautifully concordant while recording an event much older than the host rock’s final assembly.
For an original diagnostic exercise, imagine two groups. Group A has mutually consistent dates near 1000 Ma from cores. Group B has mutually consistent dates near 600 Ma from rims. Combining them into a single mean of 800 Ma would destroy the observed structure. The two groups invite an interpretation of distinct domains and events, subject to the images and field evidence.
The more focused concordia and discordia guide develops the geometry and ambiguity of disturbed systems. Use it after you can explain why the two ratios can agree on one time even though their numerical values differ.
A system can depart from the simple model if material moves in or out, if distinct domains are mixed, or if a correction is inadequate. A discordant result tells us that the simple interpretation needs investigation. It does not identify one unique cause by itself.
Consider an explicitly artificial case: start with the 1000 Ma ratios above, then remove 20% of both radiogenic lead populations immediately before measurement while leaving uranium unchanged. Multiply each ratio by 0.8 and invert. The resulting apparent dates are roughly 812 Ma and 864 Ma. Neither is the original growth age, and their difference is a warning about the model.
Why a recent loss example is deliberately simple
The phrase “immediately before measurement” removes subsequent ingrowth from the exercise. If lead were lost long ago, new daughter lead would accumulate after the loss. The equations would need to represent both intervals. A realistic disturbance history cannot be recovered by applying the same percentage shortcut without checking what history it represents.
A fitted discordia may intersect concordia at two ages, but assigning those intercepts to crystallisation and a single disturbance requires a suitable geological and analytical model. Continuous loss, multiple events and mixtures can complicate that interpretation. A neat line is a mathematical pattern; its physical meaning needs evidence.
The 2026 research paper Modelling diffusion, decay and ingrowth of U–Pb isotopes in zircon explicitly treats transport alongside decay. It illustrates why temperature, duration, position and material properties matter when evaluating possible lead redistribution. Its model conditions should not be converted into a universal temperature threshold for every zircon.
A useful repair sequence is to inspect images and spot positions, examine common-lead treatment, review calibration, and compare domains before fitting increasingly elaborate histories. More complicated mathematics cannot compensate for a spot that unknowingly crosses two growth zones.
CHAPTER 7 OF 14 · Correct and measure
7. Keep common lead separate from radiogenic lead
The clock equation needs daughter lead generated by the parent history being modelled. Measured lead can also contain an initial or introduced component. Treating all measured ²⁰⁶Pb as radiogenic can make an apparent date too old in a simple case where the additional lead increases the numerator.
Suppose the measured ²⁰⁶Pb/²³⁸U ratio is 0.1100, but independent evidence establishes a common-lead contribution of 0.0100 in the same ratio units. The corrected radiogenic ratio is 0.1000. The uncorrected calculation gives approximately 672.7 Ma, whereas the corrected calculation gives approximately 614.4 Ma. The difference comes from the assumed component separation, not a change in the decay constant.
A correction needs its own evidence
In that exercise the common component is supplied as known. Real work has to estimate it, measure relevant signals, or adopt a justified model. Uncertainty in the correction belongs in the resulting uncertainty. A subtraction written in a spreadsheet is not enough to establish that the subtracted quantity is correct.
Ask how the common-lead estimate was obtained, which isotopes or assumptions were used, and how sensitive the result is to alternative plausible values. If a small change in the common component strongly shifts the age, the interpretation should reflect that sensitivity rather than presenting the preferred correction as uniquely established.
Blank contributions and contamination also deserve attention. The laboratory procedure can introduce material that must be characterised. This is one reason analytical records include more than the final isotope ratios: preparation, backgrounds and reference measurements affect the meaning of the reported result.
A useful language repair is to replace “The lead gives the age” with “The corrected radiogenic lead-to-uranium system gives this apparent date under the stated model.” The longer sentence names the inferential steps. Once those steps are secure, a concise geological conclusion becomes much more informative.
CHAPTER 8 OF 14 · Correct and measure
8. Match the analytical method to the geological question
Three widely used approaches are laser ablation inductively coupled plasma mass spectrometry, secondary ion mass spectrometry, and isotope dilution thermal ionization mass spectrometry. Their abbreviations are LA-ICP-MS, SIMS and ID-TIMS. They differ in sample treatment, spatial targeting, throughput and attainable precision. There is no single choice that answers every question equally well.
A microanalytical approach can target a visible domain, subject to spot size, depth and the actual structure intersected. Dissolution-based analysis can measure a selected grain or fragment with high precision, but the selected material may combine growth histories. Spatial selectivity and numerical precision are separate properties.
Schoene’s U–Th–Pb geochronology overview covers these methods and their interpretive tradeoffs. For the preparation-to-interpretation chain in high-precision work, see the 2024 CA-ID-TIMS methods overview.
Design a two-stage investigation
Imagine a rock in which images suggest inherited cores and younger rims. The first priority is to establish which domains exist and how their dates relate. An analysis that averages the entire grain might produce an extremely precise intermediate result while obscuring that structure.
Now imagine a well-characterised volcanic layer where the question is the duration between two closely spaced events. Once the relevant zircon population has been established, improved analytical precision may directly help resolve the interval. The best method follows the unresolved question rather than a general ranking of instruments.
Chemical abrasion is used to reduce the influence of damaged material in appropriate zircon analyses. It is not a guarantee that every remaining domain is unaffected. Experimental research on chemical abrasion reports residual lead-loss concerns and tests how treatment changes results. The proper lesson is to evaluate performance and sample behaviour, not to regard a technique name as a substitute for quality assessment.
This guide explains method selection, not laboratory handling procedures. The practical learning task is to read a methods section and identify what material the reported number actually integrates.
An uncertainty statement must specify its coverage or convention and the components included. Analytical uncertainty, tracer calibration and decay constants need not contribute in the same way to every comparison. Two numbers with identical printed plus-or-minus values may therefore not be directly comparable.
For a simplified independent ratio uncertainty, differentiation of t = ln(1 + R)/λ gives σt approximately σR/[λ(1 + R)], treating λ as exact. At R = 0.1000 and σR = 0.0010, this gives approximately 5.86 Ma as a one-standard-deviation age uncertainty. This is a local propagation exercise, not a complete laboratory uncertainty budget.
Separate random and shared components
Suppose two hypothetical dates are 100.0 ± 1.0 Ma and 103.0 ± 1.0 Ma, with independent one-standard-deviation uncertainties. The difference is 3.0 Ma and its standard uncertainty is √(1² + 1²), approximately 1.41 Ma. The comparison is about 2.12 standard uncertainties from zero under those assumptions.
If both dates share a calibration uncertainty, simply treating every error component as independent can misrepresent uncertainty in their difference. Some shared effects can cancel partly in a relative comparison while still affecting absolute placement on the timescale. The covariance structure determines the appropriate calculation.
Likewise, averaging several dates reduces uncertainty only under a suitable statistical model. Repeated measurements of a shared bias do not eliminate that bias. Measurements of distinct growth events should not be forced into a common-age mean just because averaging produces a smaller reported error.
Inspect scatter before celebrating precision
If nominally equivalent domains scatter more than expected, investigate geological variation, analytical effects and the uncertainty model. An overdispersion statistic is a diagnostic, not an instruction to delete inconvenient measurements until the statistic looks attractive.
Write exclusion rules in terms of evidence: an identified mixed domain, a documented analytical failure, or a stated quality criterion. Keep the original data and explain the consequences of exclusion. A reader should be able to distinguish a defensible selection from a result created by choosing whichever numbers support the desired story.
CHAPTER 10 OF 14 · Interpret and diagnose
10. Interpret inherited cores and younger growth
A zircon can preserve a sequence rather than a single birthday. Imagine a grain with a core giving approximately 1200 Ma and a rim giving approximately 450 Ma. If imaging and analytical checks support distinct domains, the data suggest at least two episodes represented within the grain. They do not automatically establish the complete history between those episodes.
The older core might have been inherited from pre-existing material. The younger domain might relate to later growth. Choosing between geological explanations requires petrographic and field evidence, and often complementary chemistry. Numerical age order establishes that one recorded domain is older; it does not by itself explain the physical process that formed each domain.
Test a mixed-domain interpretation
Suppose an analysis overlaps the boundary. Its ratios can combine contributions from both domains. The resulting apparent date need not equal a simple arithmetic mean of their ages because isotope ratios mix according to element amounts, and the age transformation is nonlinear.
A useful classroom demonstration is to compare two operations: average two ages, or first mix their isotope inventories and then calculate an apparent date. These operations answer different questions. To perform the second correctly, specify how much uranium and lead each component contributes; a visual estimate of area alone is insufficient.
If a published interpretation relies on one boundary-crossing spot, a sensible next measurement is a better-isolated domain or an independent observation that tests the proposed history. Repeating the same mixed analysis more precisely does not remove the mixing problem.
The reporting sentence should preserve the domain: “The analysed rim population supports growth near 450 Ma” is more informative than “The zircon is 450 Ma.” It leaves room for the older core and tells the reader which part of the evidence carries the conclusion.
CHAPTER 11 OF 14 · Interpret and diagnose
11. Use detrital ages without inventing a depositional date
A detrital zircon formed before it was incorporated into the sediment under investigation. Its date can constrain deposition, but the time between crystallisation and deposition may include erosion, storage and recycling. The youngest reliable population can help establish a maximum depositional age; it does not automatically equal the deposition age.
For a hypothetical sandstone with robust youngest grains near 90 Ma, deposition must be younger than or approximately contemporaneous with those grains, allowing for uncertainty and correct interpretation. In numerical Ma notation, younger ages are smaller. This reversal between “maximum age” and “earliest permissible time” is a frequent source of verbal mistakes.
Add an independent boundary
Suppose a cross-cutting intrusion is securely dated at 70 Ma and demonstrably intrudes the already-deposited sandstone. Then, ignoring uncertainty for this teaching case, deposition lies between 90 and 70 Ma. The interval comes from two different relationships: the sediment postdates its detrital grains and predates the intrusion.
If the intrusion relationship is uncertain, the lower numerical bound cannot simply be retained. Similarly, if the 90 Ma grain is affected by lead loss, it may not supply a reliable depositional constraint. The strength of a bracket depends on both geological relationships and the validity of the clocks.
The specialist youngest detrital zircon guide develops this distinction in more detail. For source interpretation, consult recovering source-rock fractions from unequal detrital grain yields.
An abundant age peak represents many measured grains under the sampling and selection procedure. It does not directly establish the same fraction of sediment mass from that source. A source can yield zircon abundantly or sparsely, and transport and selection can alter the observed population. Keep the age-distribution question separate from the mass-contribution question.
CHAPTER 12 OF 14 · Interpret and diagnose
12. Diagnose a doubtful interpretation systematically
Begin with identity. Can each result be traced to a specimen, grain and analytical domain? If not, recover the connection before interpreting geological structure. Next inspect the ratio definition, units and constants. A wrong isotope denominator or a factor-of-a-million unit error is easier to fix than an elaborate geological model built around it.
Then inspect correction and calibration. Was common lead addressed appropriately? Are uncertainties comparable? Are reference measurements and analytical performance documented? These questions determine whether the numerical inputs are suitable for the proposed calculation.
Use competing explanations
If dates are scattered, list at least two plausible causes and the observation that would distinguish them. Mixed growth zones predict relationships with domain boundaries. A session-wide calibration issue might affect reference materials or multiple unrelated grains. Lead-loss interpretations need evidence consistent with open-system behaviour rather than scatter alone.
If a sandstone date is being described as a depositional date, ask whether the zircon was deposited shortly after growth or recycled. If a granite date is being called an eruption date, ask whether eruption is even the event represented by that rock and population. Repair the event label before refining the decimals.
Finally, test whether the conclusion changes when a reasonable alternative assumption is used. A correction, selection rule or grouping choice can sometimes dominate the interpretation. Showing that sensitivity helps readers understand what is well established and what remains conditional.
Preserve an audit trail
A useful report includes sample context, analytical domain images where available, method and reference information, isotope ratios, uncertainty conventions, correction choices, rejected analyses with reasons, and the reasoning that connects the accepted results to an event. The reader should be able to follow the chain without reconstructing hidden decisions.
This does not require pretending that every uncertainty has been eliminated. “These domains support one growth episode, while the youngest disturbed analyses remain unexplained” is a legitimate conclusion. It is more useful than a single smooth date that conceals incompatible observations.
CHAPTER 13 OF 14 · Transfer and continue learning
13. Transfer the method when the context changes
Work through this fictional case before reading the answer. A sandstone contains zircon populations near 1000, 300 and 120 Ma. Three apparently clean young rims cluster near 120 Ma; an isolated cracked grain gives an apparent 80 Ma date and is discordant. An intrusion cutting the sandstone is independently established at 100 Ma. Ignore numerical uncertainty initially and assume the geological relationship is secure.
Question one: does the sandstone have an age of 120 Ma? Question two: should the isolated 80 Ma result replace the young cluster as the maximum depositional constraint? Question three: what age interval is supported if the 120 Ma rims record genuine pre-depositional growth and the 80 Ma result is unsuitable for that purpose?
Answer and reasoning
The sandstone is not demonstrated to be exactly 120 Ma. The young valid grains existed before deposition. The isolated 80 Ma result cannot be promoted merely because it is youngest: discordance and the damaged domain require investigation. Under the stipulated conditions, deposition lies between 120 and 100 Ma.
The older populations remain valuable for provenance and recycling questions. They are not failed measurements just because they do not provide the tightest depositional bound. Conversely, the intrusion age brackets deposition because of the field relationship, not because 100 is conveniently between other numbers.
Change one condition
Now suppose the 120 Ma rims are shown to have grown after burial during metamorphism. The previous maximum depositional interpretation fails because the domains no longer predate deposition. The same numerical dates now answer a different event question. Reassess the older detrital domains and the intrusion relationship; do not preserve the old bracket by habit.
Change a second condition: the supposed intrusion is actually a transported igneous clast in the sandstone. Its crystallisation predates deposition, so the direction of its constraint reverses. This exercise shows why a field label can matter more than another decimal place in a date.
The transfer test is passed when you can explain which inequality changed and why. Memorising “youngest zircon equals maximum depositional age” is insufficient unless you can establish that the zircon domain was genuinely detrital and its date is suitable for the comparison.
CHAPTER 14 OF 14 · Transfer and continue learning
14. Build a practical learning and reading route
Start with atom accounting. Explain aloud why one half-life gives a daughter-to-remaining-parent ratio of one. Then derive the logarithmic age equation, calculate a synthetic ratio, invert it, and check the units. These tasks establish whether the mathematical mechanism is understood rather than merely copied.
Next draw a core and rim and assign distinct fictional ages. Decide which spots isolate domains and which mix them. Finally, place the grains in a geological setting: granite, sandstone or metamorphic rock. State what additional evidence is needed before naming an event.
Frequently asked questions
Does concordance prove the rock age? No. It supports compatibility between isotope systems under the treatment used. The dated domain can still be inherited, and its relationship to the host must be established.
Does a discordant grain have no scientific value? No. It can contain evidence about disturbance or mixing. The problem is assigning it a simple growth or depositional meaning without resolving that behaviour.
Is a smaller uncertainty always better? It is useful when the measurement targets the right material and event. An extremely precise mixture can answer the wrong question more precisely. Improve domain selection and geological interpretation alongside analytical performance.
Can all zircon dates be averaged? No. A common-age calculation requires a justified population and an appropriate uncertainty model. Distinct growth episodes should remain distinct unless the scientific question explicitly calls for another statistic with a clear meaning.
Is common lead always negligible in zircon? No. Low initial lead is one reason zircon is useful, but the relevant sample, signals and precision requirements must be evaluated. A general mineral property cannot replace a sample-specific correction assessment.
Can zircon date sediment deposition exactly? Sometimes a suitable volcanic contribution closely constrains deposition, but detrital grains can be much older than their sediment. State the evidence for a close connection instead of assuming it.
Read a real dataset with a question
The USGS New Hampshire uranium–lead data release offers an example of documented zircon analyses and geological context. Use a real dataset to identify sample names, methods, uncertainty conventions and the interpretation being tested. Do not import its ages into an unrelated rock merely because the mineral is the same.
For a first independent review, write four sentences: what was sampled; what was measured; which assumptions connect the ratios to dates; and which observations connect the dates to an event. If any sentence is missing, you have identified the next useful reading or measurement. That is the working standard: a result whose numerical calculation and geological meaning can both be followed.
Continue with a worked weighted-mean check
Separate weighted-mean age uncertainty from excess scatter across zircon analyses with original inverse-variance, MSWD and uncertainty-component examples.