Radiometric Anomalies | Berean Archive

"If the present data are representative, argon of slightly anomalous composition can be expected in approximately one out of three volcanic rocks." — Brent Dalrymple, 1969¹

"The inert gas 'ages' then, are of uncertain significance for these samples." — Funkhouser and Naughton, on the Hualalai xenoliths, 1968²

"Nevertheless, if we use the Dalrymple (1969) study as a guide, he reported that about one-quarter of his results were anomalous (see below). At least he knew which samples gave anomalous results." — Andrew Snelling, 2025³

Summary

Radiometric dating sometimes gives the wrong answer. Lavas that erupted within living memory have returned potassium-argon ages of hundreds of thousands of years; diamonds have returned an isochron older than the earth; a human skeleton was dated four ways to 62,000 years and then redated to 40,000. The young-earth literature has collected these cases since 1979 and draws one inference from them: if the methods fail on rocks of known age, they cannot be trusted on rocks of unknown age. This article takes the cases at full strength, reads the primary papers behind them, and tests the inference. It absorbs the "when dating goes wrong" section of the radiometric concordance article, which now keeps a summary and a link.

The findings, stated first.

The famous known-age failures are real, small, and two-directional. Dalrymple's 1969 study of 26 historic lava flows, obtained in full for this article, found 18 indistinguishable from air, five with excess argon giving 0.10 to 1.19 million years, and three with the opposite anomaly giving negative ages down to −0.20 million years.¹ The creationist retellings never mention the negative ones. Scatter of both signs about the atmospheric value is the signature of fractionation, which Krummenacher identified in 1970.
The most-cited creationist table prints argon concentrations as ages, and has since 1996. Austin 1996, Snelling 1999 and Snelling 2025 list seven of Dalrymple's flows with "ages" of 1.6, 1.41, 0.25, 0.35, 0.11, 0.27 and 0.25 million years. Those are the excess-argon concentrations from the second column of Dalrymple's Table 2, in 10⁻¹² moles per gram. The ages in the fourth column are 1.19, 1.05, 0.10, 0.15, 0.13, 0.22 and 0.21. Henke's 2003 charge is confirmed in every digit. ICR conceded it in an addendum; the 2025 review reprints the table uncorrected.
The spectacular numbers come from material the original authors said was undatable. Hualalai's billions of years are from mantle xenoliths, whose measurers wrote that the ages "are of uncertain significance". The 22-million-year Kilauea figure is from deep-ocean glass whose measurers wrote "Caution is urged". Beryl has been known since 1958 to carry excess argon; glauconite is distrusted in the mainstream literature to this day, and a quarter of Woodmorappe's founding catalogue of anomalies is glauconite.
Excess argon is additive, and that is the whole argument. A fixed quantity of inherited argon that makes a fresh lava read 300,000 years makes a 50-million-year lava read 50.3 million. The young-earth inference from "it fails on new rocks" to "it fails on old rocks" requires the error to scale with age. It does not, and no young-earth source found for this article addresses the point.
The field throws away a third to half of its raw dates by rules published in advance, and then checks the survivors against a different clock. The 2021 argon-argon standard defines a plateau, sets an explicit excess-argon veto, and forbids trimming data to improve statistics. On the Mesa Falls Tuff, 147 single-crystal dates leave 36 to 55; the outliers are old, from excess argon and inherited crystals; the filtered result, 1.3009 million years, matches zircon uranium-lead on the same rock at 1.3004.⁴
The known-age test, done with the right method, passes. Sanidine from the AD 79 eruption of Vesuvius gave an argon-argon isochron age of 1,925 ± 94 years against a calendar age of 1,918. The same crystals carried excess argon, and their total-gas age was 3,300 years. The isochron saw the excess and removed it.⁵ The single potassium-argon measurements Austin and Snelling commissioned cannot do that, and almost all of them fell below the laboratory's own published floor of 2 million years.
Concordance can fail, and the article says where. The Zaire diamonds were concordantly wrong by two argon methods, and a chemical correlation caught what reproducibility could not. Lake Mungo was one uranium measurement reported three ways. Discordance, by contrast, is the alarm working: Jinmium's 116,000-year thermoluminescence age fell to under 10,000 because the radiocarbon disagreed.
Where the young-earth side is right. Snelling's 2016 critique of the potassium-40 decay constant is accurate; there is a real 0.4 per cent floor under every argon age from the choice of constants. The glauconite problem is real. Snelling's 2025 review correctly names the Hualalai xenoliths, reports Surtsey giving a correct zero age, and concedes that radiometric ages "usually match their stratigraphic order."³ Both real effects are one-per-cent effects. A young earth needs a million-fold one.

The verdict. Anomalous dates are not evidence against radiometric dating. They are the reason the method has rejection rules, and the mainstream literature found, named and quantified every mechanism in the creationist catalogue before the catalogue was compiled. The creationist argument survives only by omitting what the primary papers say about their own samples, by printing one column of a table as another, and by an inference from additive error to proportional error that fails on arithmetic. What remains true is narrower and worth saying: concordance between methods that share a failure mode proves nothing, single dates from single laboratories carry wider real uncertainty than their error bars, and the field discards a great deal of data by rules it publishes. Those are limits on precision. They are not a factor of a hundred thousand.

Part 1: The known-age lavas

The young-earth case starts here, and it deserves to be stated in full before it is answered. If a lava that erupted in 1801 gives a potassium-argon age of over a million years, something in the method is wrong, and a reader is entitled to ask why the same method should be trusted at a billion.

Dalrymple's 26 historic flows, read from the paper. In 1969 Brent Dalrymple measured the argon isotope ratio in 26 lava flows of known historical age, chosen precisely "to test the hypothesis [1] that excess radiogenic 40Ar could be a serious problem in potassium-argon dating studies of young volcanic rocks."¹ His result: "18 of the samples have 40Ar/36Ar values that are not significantly different from atmospheric argon."¹ Of the rest, "five have 40Ar/36Ar ratios greater than atmospheric argon and three less."¹ Table 2 of that paper, which the Berean Archive obtained in full for this article after two earlier attempts failed, lists the apparent ages. The five flows with excess argon give 0.10 to 1.19 million years. The three with the opposite anomaly give negative ages, down to −0.20 million years.¹ The young-earth literature cites the first group constantly and never mentions the second. Dalrymple's own conclusion: "With the exception of the Hualalai flow, the amounts of excess 40Ar and 36Ar found in the flows with anomalous 40Ar/36Ar ratios were too small to cause serious errors in potassium-argon dating of rocks a few million years old or older. However, these anomalous 40Ar/36Ar ratios could be a problem in dating very young rocks. If the present data are representative, argon of slightly anomalous composition can be expected in approximately one out of three volcanic rocks."¹ That is the honest mainstream statement, from 1969: one rock in three carries a small anomaly, of either sign, and it matters for rocks younger than a few million years.

The miscopied table, now settled. Steve Austin's 1996 paper lists seven of Dalrymple's flows under the heading of rocks that "yielded excessively old K-Ar 'ages'," with the values 1.6, 1.41, 0.25, 0.35, 0.11, 0.27 and 0.25 million years.⁶ Snelling reproduced the same seven numbers in 1999 and again in his 2025 review.⁷³ Kevin Henke charged in 2003 that these were not ages but the excess-argon concentrations from the neighbouring column, in units of 10⁻¹² moles per gram.⁸ With Table 2 in hand the charge is confirmed in every digit. The second column reads 1.60 ± 0.16, 1.41 ± 0.08, 0.25 ± 0.08, 0.35 ± 0.14, 0.11 ± 0.03, 0.27 ± 0.09 and 0.25 ± 0.15, in moles per gram. The fourth column, the apparent ages, reads 1.19, 1.05, 0.10, 0.15, 0.13, 0.22 and 0.21 million years.¹ The creationist table printed the measurement as the conclusion, and has done so for twenty-nine years. ICR's web reprint of Snelling's 1999 article carries an addendum conceding that the numbers "reflect the amounts of argon rather than the ages of the rocks listed in Dalrymple's paper," adding that "this in no way changes the argument in question."⁷ It changes it somewhat. The flows did carry excess argon, which is what Dalrymple's title said; but the ages were smaller than printed, four other flows in the same table came out negative, and the paper's author concluded the effect was too small to matter for old rocks.

Hualalai, 1801: the xenoliths. The spectacular figure in the popular tables, up to 2.96 or 3.0 billion years for a Hawaiian flow erupted in 1801, comes from Funkhouser and Naughton's 1968 study of ultramafic xenoliths, chunks of mantle rock carried up in the lava. Their abstract: "Apparent ages calculated from the measurement of radiogenic argon and helium have very high values ... From gas extraction by crushing and decrepitation, it is shown that the inert gases occur in fluid and gaseous inclusions".² Dalrymple 1969 says the same: "this flow is noted for its abundance and variety of ultramafic xenoliths."¹ His 1986 restatement is blunt: "Funkhouser and Naughton were quite careful to point out that the apparent 'ages' they measured were not geologically meaningful. Quite simply, xenoliths are one of the types of rocks that cannot be dated by K-Ar techniques."⁹ The other big numbers in the popular tables, up to 22 million years for Hawaiian basalts, are from Noble and Naughton's deep-ocean samples, whose abstract reports that ages "increase with sample depth up to 22 million years for lavas deduced to be recent. Caution is urged in applying dates from deep-ocean basalts".¹⁰ The studies that produced the famous numbers were studies of whether the material was datable, and they concluded it was not.

Mount St Helens, 1986. Austin submitted dacite from the ten-year-old lava dome to Geochron Laboratories and received potassium-argon ages from 0.34 to 2.8 million years on five preparations of one rock.⁶ He states in print that the laboratory was told only that the samples "came from dacite, and that the lab should expect 'low argon'". The results are internally discordant by a factor of eight, and 96 to 99 per cent of the argon measured was indistinguishable from air. And Geochron's own web page for the method, archived from 1998 to 2003 and obtained for this article, says: "Samples less than 5 M.Y. old, or containing less than 0.1%K will incur a 50% surcharge, reflecting the special care and additional analyses required. We cannot analyze samples expected to be younger than 2 M.Y."¹¹ Four of Austin's five results are below the laboratory's own floor. The one above it, 2.8 million years, is on pyroxene, a mineral known to hold excess argon.

Ngauruhoe, 1949 to 1975. Snelling submitted andesite from five recent flows of the New Zealand volcano to the same laboratory and got ages from below detection to 3.5 million years.¹² His own Table 4 shows the argon ratios scattering both ways about air: nine splits above the atmospheric value, four below. Duplicate splits gave under 0.27 against 1.3 million years, and 3.5 against 0.8. Snelling's own words: "Clearly, reproducibility was not obtained."¹² Twelve of the thirteen splits are below Geochron's 2 million year floor. In the same paper he quotes Krummenacher's 1970 finding that "fractionation, along with contamination by non-atmospheric argon inherited from the parent magma, is an important cause of anomalous 40Ar/36Ar ratios in these rocks",¹³ and does not apply it to his own two-directional data. Fractionation of atmospheric argon produces anomalies of both signs. Inherited radiogenic argon produces only one.

Part 2: Why ages go wrong

Excess argon is a quantified trace-element problem. Simon Kelley's 2002 review states the phenomenon and its cause in one abstract: the methods "occasionally produce anomalously old ages attributed to excess argon", and argon partitions so weakly from fluids into minerals, "as low as 10⁻⁶", that "fluid inclusions ... dominate the radiogenic argon budget, particularly in low potassium minerals. Melt inclusions are less dominant but become critical in dating younger samples."¹⁴ That predicts, in advance, where the worst cases will be: young, low-potassium, inclusion-rich material. The Mount St Helens dacite, the Ngauruhoe andesite, the Hualalai xenoliths and the deep-sea glasses are all exactly that. Ozawa and colleagues showed in 2006 that on 28 historic Hawaiian lavas the excess argon lives in the whole rock and not the groundmass, is "up to 0.5 my" in apparent age, and disappears when the groundmass is sieved to the 180–250 micron fraction.¹⁵ A defect that a stated preparation step removes is a systematic, not a refutation.

Materials nobody dates. Beryl has been known since 1958 to carry helium and argon "in great excess over that which can be accounted for from radioactive decay", some crystals more than 99.9 per cent excess.¹⁶ Cordierite, tourmaline, and glauconite are on the same list, and a 2025 mainstream review concedes that glauconite and illite ages "are mainly younger than those obtained from other methods".¹⁷ A list of wrong ages is not an argument until each entry is checked against what the original study was trying to do.

Inheritance and open systems in uranium-lead. Zircon crystals grow around older cores, and lose lead from radiation-damaged zones. The Treatise on Geochemistry states the limit plainly: "many samples may have experienced open-system behavior but are still statistically concordant".¹⁸ The field's answer is to destroy the damaged material rather than average over it. Mattinson's chemical-abrasion method anneals the crystal and dissolves away the lead-lost zones step by step; once that is done, "an age progression from step to step is an extremely sensitive detector of inheritance".¹⁹

Dates of a later event. A mineral records the time it cooled below the temperature at which it retains its daughter product, not necessarily when it formed. Schaen and colleagues: "the date we calculate based on 40Ar/39Ar analysis of a slowly cooled mineral is neither the crystallization age nor the time of complete system closure but some time intermediate between the two."⁴ A skeptic will say this can excuse any discordance. It cannot, because closure temperatures are measured in diffusion experiments without reference to any age and they fix an ordering: in one rock the zircon uranium-lead date must be older than the hornblende argon date, which must be older than the biotite date, which must be older than the apatite fission-track date. That ordering can fail. It does not.

Part 3: When concordance failed, and when discordance worked

The six-billion-year diamonds. Zashu, Ozima and Nitoh obtained a potassium-argon isochron on ten Zaire cubic diamonds giving 6.0 billion years, older than the earth.²⁰ Argon-argon analysis of four more diamonds reproduced the same figure.²¹ Two argon methods, concordantly wrong. The authors' proposed rescue, that the potassium in these diamonds was isotopically odd, was tested by two laboratories and failed; the ratio is normal to about 1 per cent.²² Three years later Ozima himself found the answer: excess argon in sub-micrometre mantle-fluid inclusions, identified by its correlation with chlorine, with "no age significance."²³ Reproducibility did not catch it. A second argon method did not catch it. A chemical correlation did. That is a real limit on what concordance proves.

Lake Mungo. In 1999 the age of Australia's oldest human skeleton was put at 62 ± 6 thousand years by combined uranium-series and electron-spin-resonance analyses on the bone, "in close agreement with OSL age estimates on the sediment into which the skeleton was buried of 61,000 ± 2,000 years."²⁴ In 2003 twenty-five new optical ages gave 40 ± 2 thousand.²⁵ The concordant estimate was wrong by 55 per cent, because the uranium-series and resonance ages all read the same uranium in the same bone through the same uptake model. One measurement had been reported three ways. Concordance is evidence only across methods whose failure modes are unrelated.

The KBS Tuff. The most-cited "they changed the date to fit the fossil" story, and the real sequence is worse for both sides than either tells it. Fitch and Miller's 41 determinations on the tuff ran from 223 to 0.91 million years; the famous 2.61 was a selection from that scatter.²⁶ Curtis found "numerous yellow grains of rounded and altered K-spar, clearly detrital in origin and probably derived from ancient bedrock" in the mineral separate; clean crystals gave 1.6 to 1.8.²⁷ The modern figure is 1.869 ± 0.021 million years.²⁸ The fission-track "confirmation" of the old age was not independent, and its author said so in print: the earlier ages "apparently agree closely with each other but this is mainly due to the close communication between these authors on track identification".²⁹ His own redetermination gave 1.87. Inherited detrital grains produced the wrong answer; a microscope found them.

Jinmium. Thermoluminescence gave an Australian rock shelter 116,000 to 176,000 years. Radiocarbon disagreed. Single-grain optical dating gave under 10,000 and vindicated the radiocarbon.³⁰ The disagreement between methods was the alarm, and it worked.

Part 4: What the field throws away, and what survives

The rules are written down in advance. The 2021 community standard for argon-argon dating requires a plateau to "consist of at least five or more consecutive steps that comprise at least >50% of the 39Ar released", to have no slope, and to have "an isochron regressed through all of the plateau steps with a (40Ar/36Ar)i that is indistinguishable from the atmospheric value at the 95% confidence level".⁴ The last rule is an explicit excess-argon veto. Isochron scatter is policed by a statistic with a published threshold: if the mean squared weighted deviation exceeds 1 + 2√(2/f), "there is only <5% probability that the data define an isochron."³¹ And the standard forbids the trick a skeptic would suspect: "Trimming an overdispersed data set by selectively rejecting outliers until achieving a MSWD ≈ 1 is also ill-advised ... Outlier identification and rejection must always be accompanied by full disclosure of the specific criteria used".⁴

How much is discarded. A great deal, and the numbers are published. On the Mesa Falls Tuff, 147 single-crystal argon dates "range from 1.280 Ma to 2.052 Ma and show a large tail towards older ages"; the three published filtering methods retain 53, 36 and 55 of them.⁴ The outliers are old, and are attributed to "excess Ar or undegased, inherited crystals."⁴ On the Bishop Tuff, 49 dates yield 31, 25 or 32.⁴ In detrital zircon work "most current detrital zircon studies apply a 10 %–30 % relative age cutoff."³² There is no field-wide rejection rate, and this article does not invent one. What there is, for modern work, is the raw data: the reporting standards require every analysis to be published, rejected ones included, which makes the file-drawer charge testable for work after about 2010 and untestable for work before 1990.

The filtered answer is checked against a different clock. The Mesa Falls argon result after filtering is 1.3009 ± 0.0006 million years. Uranium-lead on zircon from the same sample gives 1.3004 ± 0.0007.⁴ Two decay systems with unrelated failure modes, each after applying its own published rejection rule, agree to 0.04 per cent.

The known-age test, done properly. Renne and colleagues dated sanidine from the AD 79 eruption of Vesuvius by argon-argon step heating. The isochron gave "1925 ± 94 years ago", against a calendar age of 1,918.⁵ The same sample's total-gas age was "3300 ± 500 years, clearly distinct within error from the known calendar age", because the sanidine really does carry excess argon, with a trapped ratio of 306.9 against air at 295.5.⁵ The isochron saw the excess component and removed it. A single potassium-argon measurement, which is what Austin and Snelling commissioned, would not have. One paper contains the young-earth premise and its answer.

What step heating cannot do. It detects excess argon; it does not correct for it. Lanphere and Dalrymple, on minerals with known excess argon: "saddle-shaped age spectra are diagnostic", the minima "approach but do not reach the crystallization age", and "Neither the age spectrum diagram nor the 40Ar/36Ar versus 39Ar/36Ar isochron diagram reliably reveal the crystallization age in such samples."³³ The correct response to that spectrum is rejection, not rescue. That is a real limit, published by the people who built the method.

Laboratory against laboratory. Blind interlaboratory comparisons exist, and they are the test the "labs tune the answer" charge calls for. Two argon laboratories on one plagioclase separate: 259.9 ± 1.2 and 259.8 ± 1.2 million years.³⁴ Under that sits a systematic floor of about 0.4 per cent from the choice of potassium decay constants and standard ages.³⁵ Laser-ablation uranium-lead has a community limit of "ca. 2 % 2s".³⁶ Radiocarbon across fifty laboratories has a between-laboratory scatter of a few tenths of a per cent for accelerator labs, but about 10 per cent of all returned results are outliers, and one laboratory reported 4,810 years on a bone whose consensus was 9,528.³⁷ Anyone quoting a single date from a single laboratory without a replicate, on either side, is quoting a number with a wider real uncertainty than its error bar.

Part 5: The catalogue, and the inference

Woodmorappe, 1979. The founding document is John Woodmorappe's "Radiometric Geochronology Reappraised" in the Creation Research Society Quarterly, a table of over 300 dates that disagree with the stratigraphic position of their rocks. His selection rule: "With few exceptions, Table 1 is confined to datings on material that is considered to be reliable."³⁸ Three pages later: "Usually, however, glauconite K-Ar and Rb-Sr values are 'too young' for their biostratigraphic position. Many such anomalously young dates are entered into Table 1."³⁸ About one entry in four is a glauconite or illite date, on a mineral the standard textbook of the day said gives argon ages that "are not geologically meaningful."³⁹ His statistical claim, that "somewhat less than half of all dates agree within 10% of accepted values", is assembled from an estimate about glauconite and a remark about basalt, then generalised. His own next sentence concedes the gap: the mainstream "may contend that there are many more dates in agreement with accepted values than there are anomalies such as all of Table 1. Even if this were true, it would not appear to be an overwhelming majority, and a significant minority of discrepant dates would probably be sufficient to discredit all of radiometric dating."³⁸ That is the inference this article tests, stated by its author. A creationist reviewer of his 1999 book answered it from inside the movement. Robert Brown, in the Geoscience Research Institute's journal: "such a list, however long, is not proof that daughter/parent isotope ratios are never time significant",⁴⁰ and "A choice between an 'old-earth' concept and a 'young-earth' concept must be made on a basis other than success in the development of a reasonable explanation for radioisotope daughter/parent ratios."⁴⁰

The popular tables. Henry Morris's 1974 textbook put the Hualalai figure into circulation as a modern rock dated at billions of years. Snelling's 2025 review, to its credit, attributes it correctly to "fluid and gaseous inclusions in olivine, plagioclase, and pyroxene in ultramafic xenoliths".³ The same review's Table II lists fourteen "anomalous" lavas, four of them East Pacific Rise or seamount basalts at 580 to 700 million years and two of them Kilauea submarine glass at 21.8 and 42.9 million; none of the rows carries the word "submarine".³ Its Table I reprints the seven miscopied Dalrymple numbers as "Anomalous K-Ar ages for historic lavas calculated from their K-Ar analyses", with the 1792 Etna flow dated "AD 1972" and no mention of Henke or of ICR's own 1999 addendum.³ The popular channel goes further. Snelling in Answers magazine in 2009: lavas tested soon after eruption "invariably contained much more argon-40 than expected."⁴¹ His own 2025 technical paper: "about one-quarter of his results were anomalous ... At least he knew which samples gave anomalous results."³ Henke's count of Dalrymple's table is 18 right, 5 too old, 3 too young. "Invariably" is the version the readers get.

Crinum: wood in basalt. Snelling's most-cited own result is fossil wood inside a Queensland basalt, the basalt dated by potassium-argon at 37 to 47 million years, the wood by radiocarbon at 29,500 to 44,700 years.⁴² He draws the inference: "The presence of the radiocarbon in the fossilised wood demonstrates that the enclosing basalt cannot be millions of years old and that the radioisotopic 'dating' is grossly in error."⁴² Two things in his own paper answer it. First, he does not believe the radiocarbon age either: "While not providing the true age, the excessive radiocarbon 'age' is consistent with a stronger magnetic field".⁴² Both numbers in the "conflict" are rejected by the man who produced them. Second, the radiocarbon result is what a carbon-dead sample looks like under the pretreatment his laboratories used. Bird and colleagues ran charcoal independently known to be older than 50,000 years through the standard acid-base-acid procedure and got 46,300 years at 0.32 per cent modern carbon; only a harsher oxidation procedure pushed it past 51,000.⁴³ Snelling's wood, under the acid-base-acid procedure, returned 0.4 to 2.5 per cent modern. He cites the Bird paper and answers it with a hypothesis: "it could be that the graphite is not '14C-dead', and the 14C in it is there because the graphite is in fact not all that old."⁴² The radiocarbon article treats the process-blank question in full.

The inference itself. Snelling 2025 states it cleanly: "if these methods fail when the correct ages of the rocks are known, how can we trust these methods on rocks of unknown ages? After all, such failures are likely to be systematic".³ The answer is arithmetic. Excess argon is a quantity of gas trapped in a rock at eruption. It does not grow with the rock's age. Henke quotes the point with approval: "in 25 times out of 26 tests there is no excess argon or there is so little excess argon that it will make only a tiny error ... If the identical rock had been formed 50 million years ago, the K-Ar would give a 'false' age of a little over 51 million years."⁸ Dalrymple said the same in 1969: the excess was "too small to cause serious errors in potassium-argon dating of rocks a few million years old or older."¹ An error that is fatal at zero age is a two per cent error at fifty million years. The creationist inference requires the error to be proportional to age. Nothing in the physics of trapped gas makes it so, and no young-earth source found for this article argues that it is. Snelling's review, in fact, offers two incompatible accounts of the record: the wrong dates are excess argon, which predicts scatter, and the right dates, whose agreement with stratigraphic order he concedes, are accelerated decay, which predicts systematic inflation.³ The paper never says which dates belong to which mechanism.

What the young-earth side has right. Snelling's 2016 paper on the potassium-40 decay constant argues that the two competing half-life values "ultimately depend on adjusting (that is, massaging) K-Ar and Ar-Ar ages to conform to U-Pb and Pb-Pb ages".⁴⁴ That is a fair description of how the constants were optimised, and the consequence is real: the same standard mineral is 28.18, 28.20 or 28.29 million years depending on the constants chosen, a 0.4 per cent floor under every argon age.¹⁷ The glauconite complaint is real and conceded in the 2025 mainstream literature. Tas Walker's concession runs the other way and deserves quoting too: "rocks in the same area which give similar 'dates' are likely to have formed from the same lava at about the same time during the Flood. So, although the assumptions behind the calculation are wrong and the dates are incorrect, there may be a pattern in the results".⁴⁵ The pattern is the stratigraphic order. Explaining why it exists is the young-earth side's problem, not the mainstream's.

Objections and responses

"Dalrymple found anomalies in a third of his flows; that is not rare." Correct, and he said so. They were of both signs, small, and confined to material a few hundred years old. His conclusion in the same paragraph was that they were too small to matter for rocks a few million years old or older.

"The laboratory gave Austin and Snelling millions of years on rocks it was told nothing about." It gave them results below its own printed floor of 2 million years, at a 50 per cent surcharge for samples under 5 million, on a technique that cannot see excess argon. Wiens' summary is fair: "Geologists have known for over forty years that the potassium-argon method cannot be used on rocks only twenty to thirty years old."⁴⁶ The argon-argon isochron on Vesuvius shows what the right technique does with the same problem.

"Geologists 'interpret' any date to fit." Walker's version is that "No matter what the radiometric date turned out to be, our geologist would always be able to 'interpret' it."⁴⁵ The rejection rules in Part 4 are published before the data are collected, the raw analyses must be published including the rejected ones, and the filtered answer is checked against a second decay system. Where the interpretation is unfalsifiable, as the Treatise concedes of "recrystallization",¹⁸ this article says so.

"Anomalous dates go unpublished, so the true failure rate is hidden." Woodmorappe asserted it in 1979; Snelling repeats it in 2025. It is unmeasured on both sides. For modern work the reporting standards make it testable, because every analysis must be printed; for work before 1990 it is not. That asymmetry is the honest statement.

"Lake Mungo and the diamonds show concordance is worthless." They show concordance between methods that share a failure mode is worthless. The Mesa Falls and Vesuvius checks are between methods that do not.

"Excess argon shows the mantle is full of inherited argon, so every old basalt is inflated." It shows the mantle carries a small, fixed component, which the isochron intercept measures and removes. A fixed component is a large fraction of the argon in a young rock and a negligible fraction in an old one.

Issues and unanswered questions

No creationist source found addresses the additive-versus-proportional point, and no mainstream source states it as directly as it should be stated.
Whether the distribution of anomalous argon ratios across all historic-flow studies is symmetric about air has not been computed; Dalrymple's Table 1, Snelling's Table 4, and Ozawa's data would settle it.
The Puetz database of 604,553 zircon analyses could give a field-wide concordance failure rate. Nobody has published it.
Funkhouser and Naughton's statement that the Hualalai matrix without xenoliths gave zero age is cited by Henke and Isaak to p. 4603 and not yet read from the paper.
Lanphere and Dalrymple's warning that neither the spectrum nor the isochron reliably recovers a crystallisation age from a sample with excess argon deserves a fuller treatment than this article gives it.

Verdict

Every mechanism in the creationist anomaly catalogue was found, named and quantified by the mainstream literature before the catalogue existed: excess argon in inclusions in 1958 and 1968, its two-directional scatter in 1969 and 1970, its trace-element physics in 2002, undatable minerals in every textbook, inheritance and lead loss since 1955. The catalogue's most-cited table has printed concentrations as ages for twenty-nine years, was corrected by its own publisher in 1999, and was reprinted uncorrected in 2025. Its most spectacular numbers come from samples whose measurers said in their abstracts that the ages meant nothing.

What survives is a set of limits. A quarter to a third of young volcanic rocks carry a small argon anomaly. The field discards a third to half of its raw single-crystal dates by published rules. Concordance between correlated methods proves nothing. A 0.4 per cent floor sits under every argon age, and glauconite should not be dated. None of these is in dispute, and none of them reaches past the second decimal place of a billion.

What would change this verdict. A young-earth argon-argon step-heating experiment on the Mount St Helens or Ngauruhoe material producing a plateau that meets the published criteria at a non-zero age. A mechanism by which trapped argon scales with the age of its host. A known-age lava that fails the isochron test the way Vesuvius passed it. None has been published.

Sources

1.
Dalrymple, G. Brent. “⁴⁰Ar/³⁶Ar analyses of historic lava flows.” Earth and Planetary Science Letters 6(1):47–55. 1969. A laboratory study of historic lava flows that measured the argon isotope ratio in each sample, to test whether extra radiogenic argon could spoil potassium-argon dates of young volcanic rocks. Most samples had argon like the air, and eight had an anomaly, five with a higher argon ratio than air and three with a lower one. The article uses it to show that the known-age failures are real, small and run in both directions, and the authors judged the anomalies too small to matter for rocks a few million years old, except at Hualalai.
1. 1a.
  Page 47: “This work was done to test the hypothesis [1] that excess radiogenic 40Ar could be a serious problem in potassium-argon dating studies of young volcanic rocks.” ↩
2. 1b.
  Page 49: “The results (table 1) show that 18 of the samples have 40Ar/36Ar values that are not significantly different from atmospheric argon.” ↩
3. 1c.
  Page 49: “Of the eight samples with anomalous argon compositions, five have 40Ar/36Ar ratios greater than atmospheric argon and three less.” ↩
4. 1d.
  Table 2, page 51: “Calculated 'excess' argon and apparent age for samples with anomalous 40Ar/36Ar ratios.” Columns Excess Ar: 40Ar (10⁻¹² mol/g), Excess Ar: 36Ar (10⁻¹⁵ mol/g), Apparent age (10⁶ years). Hualalai 1800–1801: 1.60 ± 0.16 and 1.41 ± 0.08, ages 1.19 and 1.05. Mt. Etna 122 B.C.: 0.25 ± 0.08, age 0.10. Mt. Etna 1792: 0.35 ± 0.14, age 0.15. Mt. Lassen 1915: 0.11 ± 0.03, age 0.13. Sunset Crater 1064–1065: 0.27 ± 0.09 and 0.25 ± 0.15, ages 0.22 and 0.21. Glass Mountain: excess 36Ar, ages −0.13 and −0.03. Mt. Mihara 1951: −0.07. Sakurajima 1946: −0.20. ↩ ↩ ↩
5. 1e.
  Page 49: “The occurrence of excess 40Ar in the Hualalai flow is not surprising, because this flow is noted for its abundance and variety of ultramafic xenoliths.” ↩
6. 1f.
  Page 52: “With the exception of the Hualalai flow, the amounts of excess 40Ar and 36Ar found in the flows with anomalous 40Ar/36Ar ratios were too small to cause serious errors in potassium-argon dating of rocks a few million years old or older. However, these anomalous 40Ar/36Ar ratios could be a problem in dating very young rocks. If the present data are representative, argon of slightly anomalous composition can be expected in approximately one out of three volcanic rocks.” ↩ ↩ ↩
2.
Funkhouser, J. G., and J. J. Naughton. “Radiogenic helium and argon in ultramafic inclusions from Hawaii.” Journal of Geophysical Research 73(14):4601–4607. 1968. A study of the inert gases helium and argon in xenoliths, which are chunks of mantle rock carried up in lava, from two Hawaiian sites including the 1800–1801 Kaupulehu flow of Hualalai. It finds that the gases sit in fluid and gaseous inclusions inside the minerals, so the apparent ages they give are of uncertain significance. The article uses it as the source of the Hualalai figure of billions of years that appears in popular creationist tables.
1. 2a.
  Section Abstract: “Isotopic studies have been made of the inert gases present in ultramafic xenoliths from two sites in Hawaii, the 1800–1801 Kaupulehu flow (Hualalai Volcano, Hawaii) and Salt Lake Crater (Oahu). Apparent ages calculated from the measurement of radiogenic argon and helium have very high values. … From gas extraction by crushing and decrepitation, it is shown that the inert gases occur in fluid and gaseous inclusions in the minerals of the nodules. … The inert gas 'ages' then, are of uncertain significance for these samples.” ↩ ↩
3.
Snelling, Andrew A. “Age-Dating of Volcanic Rocks: A Review.” Creation Research Society Quarterly 61(4):283–300. 2025. A review of published work on age-dating volcanic rocks, issued in the spring 2025 issue. It argues that radioisotope ages of such rocks are unreliable, and that the inflated ages often match the rock layers, which the author ties to accelerated radioactive decay during the Genesis Flood. The article uses it for its concessions, for its reprint of the miscopied Dalrymple table, and for its Table II. ↩
1. 3a.
  Table I, page 284: “Anomalous K-Ar ages for historic lavas calculated from their K-Ar analyses [Dalrymple, 1969]”. It reprints 1.6, 1.41, 0.25, 0.35, 0.11, 0.27 and 0.25 Ma, with the Etna row dated AD 1972. Henke's name does not appear in the paper. ↩
2. 3b.
  Table II, page 285: “Anomalous K-Ar ages for lavas with historic or expected recent ages [sources in Snelling, 2000]”. Fourteen rows, four East Pacific Rise or seamount basalts at 580–700 Ma, two Kilauea submarine glasses at 21.8 and 42.9 Ma, none labelled submarine. ↩
3. 3c.
  Page 283: “How serious this problem may really be is hard to gauge or quantify because it would seem that not all discrepant results are published, for obvious reasons. Nevertheless, if we use the Dalrymple (1969) study as a guide, he reported that about one-quarter of his results were anomalous (see below). At least he knew which samples gave anomalous results.” ↩ ↩
4. 3d.
  Page 286: “Funkhouser and Naughton (1968) found that the excess 40Ar* in the 1800–1801 Hualalai flow, Hawaii, resided in fluid and gaseous inclusions in olivine, plagioclase, and pyroxene in ultramafic xenoliths in the basalt, and was sufficient to yield ages of 2.6 Ma to 2960 Ma.” “a plagioclase from the 1964 eruption of Surtsey only had argon whose isotopic composition matched that of air.” ↩
5. 3e.
  Pages 283–284: “That then raises the question—if these methods fail when the correct ages of the rocks are known, how can we trust these methods on rocks of unknown ages? After all, such failures are likely to be systematic, so that even when the methods provide ages compatible with the presumed ages, the compatibility may likewise be flawed.” ↩
6. 3f.
  Page 297: “there is often agreement between the ages of volcanic rock units obtained by these radioisotope methods and their presumed ages based on their relative stratigraphic and/or biostratigraphic positions in the rock record, plus the radioisotope ages of the volcanic rocks usually match their stratigraphic order.” Pages 297–298: “these highly-inflated radioisotope ages are reconciled by the accelerated decay of the parent radioisotopes during the recent global Genesis Flood cataclysm.” ↩ ↩
4.
Schaen, Allen J. et al. “Interpreting and reporting ⁴⁰Ar/³⁹Ar geochronologic data.” GSA Bulletin 133(3–4):461–487. 2021. The community reporting standard for argon-argon dating, written by a large group of argon-dating specialists and published in a themed issue of GSA Bulletin on reporting geochronologic data. It suggests what a plateau should look like before it counts as a date, and says that outlier rejection must come with full disclosure of the criteria used. The article uses it for the field's written rules, and for its published counts of how many dates the filters throw away.
1. 4a.
  Page 476: “the date we calculate based on 40Ar/39Ar analysis of a slowly cooled mineral is neither the crystallization age nor the time of complete system closure but some time intermediate between the two.” ↩
2. 4b.
  Page 475: “we suggest that a plateau (1) consist of at least five or more consecutive steps that comprise at least >50% of the 39Ar released; (2) not have a slope (i.e., the majority of consecutive plateau steps do not have ascending or descending ages; Sharp and Renne, 2005); and (3) have an isochron regressed through all of the plateau steps with a (40Ar/36Ar)i that is indistinguishable from the atmospheric value at the 95% confidence level (i.e., for terrestrial samples only).” ↩
3. 4c.
  Page 471: “Trimming an overdispersed data set by selectively rejecting outliers until achieving a MSWD ≈ 1 is also ill-advised because this risks the loss of geologically valuable information and biasing the results. Outlier identification and rejection must always be accompanied by full disclosure of the specific criteria used for such evaluation, and not simply to improve the statistics of a data set.” ↩
4. 4d.
  Page 474: “Single-crystal 40Ar/39Ar model dates for the Mesa Falls Tuff sanidine range from 1.280 Ma to 2.052 Ma and show a large tail towards older ages. Application of the filtering methods described above yielded inverse-variance-weighted mean dates of 1.2985 ± 0.0006 Ma (n = 53/147) for method 1, 1.2957 ± 0.0008 Ma (n = 36/147) for method 2, and 1.3009 ± 0.0006 (n = 55/147) for method 3, respectively (Fig. 7B).” ↩
5. 4e.
  Page 474: “…a group of the youngest single-crystal 40Ar/39Ar dates, whereas older dates reflect excess Ar or undegased, inherited crystals.” ↩
6. 4f.
  Page 474: “Bishop Tuff single-sanidine plateau dates (Andersen et al., 2017) also show a distribution with a tail towards older dates. Application of methods 1, 2, and 3 yielded inverse-variance-weighted mean dates of 765.2 ± 0.14 ka (n = 31/49) for method 1, 764.8 ± 0.17 ka (n = 25/49) for method 2, and 765.4 ± 0.13 (n = 32/49) for method 3 (Fig. 7C).” ↩
7. 4g.
  Page 474: “Ellis et al. (2017) noted that the inverse-variance-weighted mean date derived from method 3 (1.3009 ± 0.0006 Ma) agrees well with the zircon U/Pb data for this sample, 1.3004 ± 0.0007 Ma, and suggested that this value represents the best eruption age estimate for this sample.” ↩ ↩
5.
Renne, Paul R. et al. “⁴⁰Ar/³⁹Ar dating into the historical realm: calibration against Pliny the Younger.” Science 277(5330):1279–1280. 1997. A 1997 Science report that dated sanidine crystals (a potassium-rich feldspar) from the pumice of the eruption of Vesuvius in 79 A.D. by argon-argon step heating, which heats the sample in steps and measures the argon released at each step. The isochron (a line through the measurements whose slope gives the age) gave 1925 ± 94 years ago, against a calendar age of 1918 years ago, while the total gas gave 3300 ± 500 years because the crystals carried excess argon. The article uses it as a known-age test that the method passed, because the isochron removed the excess argon.
1. 5a.
  Section Abstract, page 1279: “Laser incremental heating of sanidine from the pumice deposited by the Plinian eruption of Vesuvius in 79 A.D. yielded a 40Ar/39Ar isochron age of 1925 ± 94 years ago. Close agreement with the Gregorian calendar–based age of 1918 years ago demonstrates that the 40Ar/39Ar method can be reliably extended into the temporal range of recorded history. Excess 40Ar is present in the sanidine in concentrations that would cause significant errors if ignored in dating Holocene samples.” ↩ ↩
2. 5b.
  Page 1279: “The initial 40Ar/36Ar ratio defined by the isochron is 306.9 ± 1.3, significantly higher than the atmospheric value and indicative of a trapped component of extraneous 40Ar.” “the 40Ar/39Ar apparent age (10) calculated from the sum of all gas released is 3300 ± 500 years, clearly distinct within error from the known calendar age.” ↩
6.
Austin, Steven A. “Excess Argon within Mineral Concentrates from the New Dacite Lava Dome at Mount St Helens Volcano.” Creation Ex Nihilo Technical Journal 10(3):335–343. 1996. Page 5: “were submitted to Geochron Laboratories with the statement that they came from dacite, and that the lab should expect 'low argon.' No information was given to the lab concerning where the dacite came from or that the rock has a historically known age”. Table 3: The ages are 0.34 ± 0.06 to 2.8 ± 0.6 Ma (million years) on five preparations, with 40Ar*/total 40Ar of 1.05 to 3.60 %. A 1996 report of potassium-argon dates on five preparations of dacite from the new lava dome at Mount St Helens, sent to Geochron Laboratories for dating. The article uses it because five preparations of one rock gave dates from 0.34 to 2.8 million years, which the article calls internally discordant. ↩
1. 6a.
  Page 6: The list of historic flows is under the heading of rocks that. “yielded excessively old K-Ar 'ages'”. The list includes. “Mt. Etna basalt (Sicily, AD 1792) 0.35±0.14Ma”. ↩
7.
Snelling, Andrew A. “'Excess Argon': The 'Archilles' Heel' of Potassium-Argon and Argon-Argon 'Dating' of Volcanic Rocks.” Impact (307). Institute for Creation Research. Jan 1999. Section Addendum on the web reprint: “the numbers given for the five historic lava flows reflect the amounts of argon rather than the ages of the rocks listed in Dalrymple's paper '40Ar/36Ar Analyses of Historic Lava Flows.' Nevertheless, this in no way changes the argument in question. While the apparent ages might be slightly different, they are still excessive because of excess argon and are in the same ballpark as those listed above.” A 1999 article that reproduces Austin's seven numbers for historic lava flows, with the 1792 Etna flow misdated to AD 1972. The article uses it as the second appearance of those seven numbers, after Austin's 1996 paper. ↩ ↩
8.
Henke, Kevin R. “Blind Leading the Blind: Austin, Snelling and Swenson Misinterpret Dalrymple's K-Ar Dating of Historical Volcanics.” No Answers in Genesis. 2003. A web essay by Kevin R. Henke, Ph.D., who checks the creationist use of Dalrymple's 1969 study of historic lava flows against Dalrymple's own data. He argues that the creationist writers miscopied the data, and that in 25 of 26 tests excess argon is absent or too small to matter for rocks millions of years old. The article uses it for the miscopy charge and for the point that excess argon adds a fixed error, not one that grows with age.
1. 8a.
  “Snelling mistakenly listed the concentrations of 40Ar (in 10 to the -12 moles/gram) for the Hualalai, Mt. Etna (2 samples), Mt. Lassen, and Sunset Crater samples as their apparent K-Ar dates!! … Austin, Snelling and Swenson all list the 'apparent K-Ar date' for the Hualalai basalt as '1.60 million years' instead of 1.19 million years. In reality, the Hualalai basalt had 1.60 x 10 to the -12 moles/gram of 40Ar.” ↩
2. 8b.
  Henke quoting, with approval, an anonymous online author: “in 25 times out of 26 tests there is no excess argon or there is so little excess argon that it will make only a tiny error, if any, in the final date for rocks millions of years old. … If the identical rock had been formed 50 million years ago, the K-Ar would give a 'false' age of a little over 51 million years.” ↩
9.
Dalrymple, G. Brent. “Radiometric dating, geologic time, and the age of the Earth: a reply to 'scientific' creationism.” U.S. Geological Survey Open-File Report 86-110. 1986. A reply by G. Brent Dalrymple of the U.S. Geological Survey to creationist writers such as Morris (1974, 1977) and Kofahl and Segraves (1975), who say the Earth is only about 6,000 to 10,000 years old. Its discussion of criticisms of K-Ar dating, on pages 26 to 28, covers the Hawaiian xenoliths and the submarine basalts. The article uses it because Dalrymple there reports that Funkhouser and Naughton themselves called the xenolith 'ages' not geologically meaningful.
1. 9a.
  Page 26: “The study by Funkhouser and Naughton (1968) was on the xenoliths, not on the lava. … Funkhouser and Naughton were quite careful to point out that the apparent 'ages' they measured were not geologically meaningful. Quite simply, xenoliths are one of the types of rocks that cannot be dated by K-Ar techniques.” Page 28: On the submarine basalts: “the findings clearly indicated that these rocks were not suitable for dating”. ↩
10.
Noble, C. S., and J. J. Naughton. “Deep-Ocean Basalts: Inert Gas Content and Uncertainties in Age Dating.” Science 162(3850):265–267. 1968. A 1968 report on inert gases in deep-ocean basalts, which gave ages rising with sample depth to 22 million years for lavas deduced to be recent. The authors urge caution in using such dates in studies of ocean-floor spreading. The article cites it beside the Hualalai case, as a second study of inert-gas ages in lavas thought to be recent.
1. 10a.
  Section Abstract: “Ages calculated from these measurements increase with sample depth up to 22 million years for lavas deduced to be recent. Caution is urged in applying dates from deep-ocean basalts in studies on ocean-floor spreading.” ↩
11.
Geochron Laboratories (Krueger Enterprises). Potassium-Argon Age Determinations. n.d. A commercial laboratory's web page describing its potassium-argon dating service and its charges. It says that samples under 5 million years old, or with less than 0.1 per cent potassium, cost a 50 per cent surcharge, and that the laboratory cannot date samples expected to be younger than 2 million years. The article uses it as the laboratory's own statement of its limit for young rock.
1. 11a.
  “Samples less than 5 M.Y. old, or containing less than 0.1%K will incur a 50% surcharge, reflecting the special care and additional analyses required. We cannot analyze samples expected to be younger than 2 M.Y.” ↩
12.
Snelling, Andrew A. “The Cause of Anomalous Potassium-Argon 'Ages' for Recent Andesite Flows at Mt Ngauruhoe, New Zealand, and the Implications for Potassium-Argon 'Dating'.” Proceedings of the Fourth International Conference on Creationism, 503–525. Creation Science Fellowship, Pittsburgh. Edited by R. E. Walsh. 1998. A paper given at a creationist conference on why potassium-argon dates for recent andesite lava flows at Mt Ngauruhoe in New Zealand came out anomalous. Its Table 4 gives thirteen splits of the rock, with 40Ar/36Ar ratios from 283 to 311 against 295.5 for air, nine above air and four below. The article uses it because the author himself reports that the Ngauruhoe results could not be reproduced. ↩
1. 12a.
  Page 509: “The model ages listed in Table 4 range from <0.27 Ma to 3.5±0.2 Ma.” Page 511: On the duplicate splits of the June 30, 1954 flow, which returned <0.27 against 1.3 ± 0.3 Ma and 3.5 ± 0.2 against 0.8 ± 0.2 Ma: “Clearly, reproducibility was not obtained.” ↩
13.
Krummenacher, D. “Isotopic composition of argon in modern surface volcanic rocks.” Earth and Planetary Science Letters 8(2):109–117. 1970. A 1970 study that measured argon isotope ratios in 27 modern surface volcanic rocks. It found that fractionation, along with argon inherited from the parent magma, is an important cause of the anomalous ratios. The article cites it for the 1970 findings that Snelling quotes in his own paper.
1. 13a.
  Section Abstract: “A study of the 40Ar/36Ar and 38Ar/36Ar ratios in argon extracted from 27 modern surface volcanic rocks shows that fractionation, along with contamination by non-atmospheric argon inherited from the parent magma, is an important cause of anomalous 40Ar/36Ar ratios in these rocks.” ↩
14.
Kelley, Simon. “Excess argon in K–Ar and Ar–Ar geochronology.” Chemical Geology 188(1–2):1–22. 2002. A paper in the journal Chemical Geology on excess argon, meaning argon in a rock beyond what its potassium has produced. Its abstract says that the methods sometimes give anomalously old ages, and that fluid inclusions, tiny pockets of fluid trapped in a mineral, dominate the argon in low-potassium minerals. The article uses it to predict where the worst cases will be: young, low-potassium rock full of inclusions.
1. 14a.
  Section Abstract: “The K–Ar and Ar–Ar dating techniques occasionally produce anomalously old ages attributed to excess argon, and such data is often rejected as not offering any insight into the age, thermal history or geochemistry of the rock.” “Partition coefficients between minerals and hydrous fluids as low as 10⁻⁶ lead fluid inclusions to dominate the radiogenic argon budget, particularly in low potassium minerals. Melt inclusions are less dominant but become critical in dating younger samples.” ↩
15.
Ozawa, A. et al. “Argon isotopic composition of some Hawaiian historical lavas.” Chemical Geology 226(1–2):66–72. 2006. A research paper in Chemical Geology on the argon isotopes in historical lava flows from Hawaii. It compared the whole rock with its groundmass, the fine material around the crystals, and found that the whole rock carried extra argon from outside the rock, which could add up to 0.5 million years to its apparent age. The article uses it as a measured case of young lava carrying extra argon.
1. 15a.
  Section Abstract: “Comparison between groundmass and whole rock samples revealed that the latter has higher 40Ar/36Ar ratios than the former, indicating extraneous 40Ar contamination. The amount of extraneous 40Ar contained in whole rock samples corresponds to an apparent age increase of up to 0.5 my.” “the argon isotopic composition of a 180–250 μm fraction agreed with the mass fractionation line, while the 250–500 μm fraction still contained extraneous 40Ar.” ↩
16.
Damon, P. E., and J. L. Kulp. “Excess helium and argon in beryl and other minerals.” American Mineralogist 43(5–6):433–459. 1958. Page 433: “All beryl crystals appear to contain a quantity of helium and argon in great excess over that which can be accounted for from radioactive decay.” A measurement study from the Lamont Geological Observatory of Columbia University, published in a mineralogy journal. The authors measured the helium, argon, potassium and alpha activity of minerals of known geological age, and found that all beryl crystals appeared to hold more helium and argon than radioactive decay accounts for, with cordierite and tourmaline showing an excess too. They read the excess as a sample of the magmatic gases around the forming crystal, and found a hundred-fold difference between crystals from the early Precambrian and those from the Paleozoic. The article uses it as the first entry in its list of materials that nobody dates. ↩
17.
Chang, Su-Chin et al. “Argon-based geochronology: advances, limitations and perspectives.” National Science Review 12(9):nwaf277. 2025. “Argon-based geochronology: advances and limitations.” A 2025 review of the two argon dating methods, potassium-argon and argon-argon, published in the journal National Science Review. It covers recent advances and the known limitations of the methods, including the choice of potassium decay constants and the reliability of ages from clay minerals. The article uses it for two admissions: that ages from glauconite and illite run mainly younger than those from other methods, and that the choice of decay constants moves a standard mineral's age by about 0.4 per cent.
1. 17a.
  “recent findings indicate that the ages obtained from these minerals are mainly younger than those obtained from other methods such as lithostratigraphy, paleomagnetism and biostratigraphy, leading to skepticism about the reliability of 40Ar/39Ar and K–Ar ages.” ↩
2. 17b.
  Table 2: Fish Canyon sanidine at 28.201 ± 0.0383, 28.294 ± 0.0474 or 28.183 ± 0.035 Ma depending on the decay-constant set, about 0.4 %. ↩
18.
Schoene, Blair. “U–Th–Pb geochronology.” Treatise on Geochemistry 4:341–378. Elsevier. 2nd edition. 2014. A chapter meant to give readers a basic understanding of the uranium-lead method, number 4.10 in volume 4 of the second edition of the Treatise on Geochemistry, a multi-volume reference. Its author is at Princeton University. It explains how the method works and lists the causes of discordance, the cases where the uranium and lead ages of one mineral disagree. The article uses it for the field's own admissions that samples can show open-system behavior and still look concordant, and that "recrystallization" is poorly defined.
1. 18a.
  Page 348: “many samples may have experienced open-system behavior but are still statistically concordant (i.e., overlap the concordia curve within uncertainty).” ↩
2. 18b.
  Page 348: “The concept of 'recrystallization,' however, is a nebulous and poorly defined process.” ↩
19.
Mattinson, James M. “Zircon U–Pb chemical abrasion ('CA-TIMS') method.” Chemical Geology 220(1–2):47–66. 2005. A 2005 methods paper in Chemical Geology on the chemical abrasion method of uranium-lead dating of zircon crystals, abbreviated CA-TIMS. The article uses it for the point that once the damaged, lead-losing zones are removed, the ages from successive steps reveal inherited older material.
1. 19a.
  Page 64: “once effects of Pb loss have been removed by early CA-TIMS steps, an age progression from step to step is an extremely sensitive detector of inheritance.” ↩
20.
Zashu, S. et al. “K-Ar isochron dating of Zaire cubic diamonds.” Nature 323:710–712. 1986. A 1986 paper that dates cubic diamonds from Zaire by a potassium-argon isochron, a line through several measurements whose slope gives the age. The article takes from it the 6.0 billion-year age, which it calls older than the earth. ↩
21.
Podosek, F. A. et al. “Normal potassium, inherited argon in Zaire cubic diamonds.” Nature 334:607–609. 1988. A 1988 Nature paper by Podosek and colleagues, three of whom also co-wrote the 1986 diamond study. Its title names the finding that the potassium is normal and the argon was inherited, and it records that argon-argon dating of four more diamonds reproduced the 6.0 Gyr (six billion year) figure. The article uses it as the second argon method that gave the same wrong age. ↩
22.
Akagi, T., and A. Masuda. “⁴⁰K-³⁹K ratio of cubic diamond from Zaire.” Geochemical Journal 22:81–82. 1988. “the hypothetical interpretation advanced by Zashu et al. (1986) … is dismissed.” A two-page paper in which Akagi and Masuda measured the potassium isotope ratio of a cubic diamond from Zaire, and found the potassium isotopically normal to about 1 %. The authors dismiss the interpretation advanced by Zashu et al. (1986). The article uses it to show that the proposed rescue of the 1986 age failed. ↩
23.
Ozima, M. et al. “Origin of the anomalous ⁴⁰Ar-³⁹Ar age of Zaire cubic diamonds: excess ⁴⁰Ar in pristine mantle fluids.” Nature 337:226–229. 1989. “no age significance”. A 1989 paper by Ozima and colleagues, Zashu among them, that explains the anomalous argon age as excess argon in pristine mantle fluids. The paper argues that the excess argon sits in sub-micrometre fluid inclusions and correlates with chlorine. The article uses it as the authors' own answer to the 1986 result, three years later. ↩
24.
Thorne, A. et al. “Australia's oldest human remains: age of the Lake Mungo 3 skeleton.” Journal of Human Evolution 36(6):591–612. 1999. Section Abstract: “in close agreement with OSL age estimates on the sediment into which the skeleton was buried of 61,000 ± 2,000 years.” A 1999 paper that dates the Lake Mungo 3 skeleton, Australia's oldest human remains, by uranium-series dating (which measures the decay of uranium) and electron-spin-resonance dating of the bone, giving 62,000 ± 6,000 years. The article uses it as a concordant estimate that was wrong, because its uranium-series and resonance ages all read the same uranium in the same bone. ↩
25.
Bowler, J. M. et al. “New ages for human occupation and climatic change at Lake Mungo, Australia.” Nature 421:837–840. 2003. A 2003 Nature letter that reports a new series of twenty-five optical ages, which date when sand grains were last exposed to light, from the Lake Mungo site. Twenty-five optical ages give 40,000 ± 2,000 years for the two burials, and the paper says the new chronology corrects earlier estimates. The article uses it to replace the 1999 skeleton age of 62 ± 6 thousand years with a younger one. ↩
26.
Lewin, Roger. Bones of Contention: Controversies in the Search for Human Origins. New York: Simon & Schuster. 1987. A 1987 book about the controversies in the search for human origins. The article uses it for the 41 dates on the KBS Tuff that it attributes to Fitch and Miller, and for how widely those dates scattered.
1. 26a.
  Pages 194–195: “at a conference in Nairobi held in September 1973 they presented 41 separate age determinations on the KBS Tuff, which varied between 223 million and 0.91 million years. Only seven of those 41 measurements came within a quarter of a million years either way of the original 2.61 number, while eight were as close to 1.9.” ↩
27.
Curtis, G. H. et al. “Age of KBS Tuff in Koobi Fora Formation, East Lake Turkana, Kenya.” Geological Background to Fossil Man, 463–470. Edinburgh: Scottish Academic Press. Edited by W. W. Bishop. 1978. Page 465: “Microscopic examination of the remaining concentrated crystals revealed numerous yellow grains of rounded and altered K-spar, clearly detrital in origin and probably derived from ancient bedrock.” A 1978 chapter in a book on the geological background to fossil man, dating the KBS Tuff at Lake Turkana in Kenya. The article uses it for the contamination it found in the crystals it concentrated for dating, and for the lower ages that clean crystals gave. ↩
28.
McDougall, Ian, and Francis H. Brown. “Precise ⁴⁰Ar/³⁹Ar geochronology for the upper Koobi Fora Formation, Turkana Basin, northern Kenya.” Journal of the Geological Society 163(1):205–220. 2006. Section Abstract: “the KBS Tuff (1.869 ± 0.021 Ma)”. A 2006 paper that dates the upper Koobi Fora Formation in northern Kenya by argon-argon dating. The article uses its figure for the KBS Tuff, a volcanic ash layer, as the modern age of that tuff. ↩
29.
Gleadow, A. J. W. “Fission track age of the KBS Tuff and associated hominid remains in northern Kenya.” Nature 284(5753):225–230. 1980. A 1980 Nature paper that dates the KBS Tuff in northern Kenya by fission-track dating, which counts the damage tracks that uranium decay leaves in zircon crystals. The article uses it as the fission-track result that seemed to confirm the old age, and notes that its author admitted in print that the confirmation was not independent.
1. 29a.
  Quoted in Lewin 1987, p. 246, on the earlier zircon fission-track ages. “apparently agree closely with each other but this is mainly due to the close communication between these authors on track identification and discrimination in these samples.” Section Abstract: Gleadow's own 1980 words, which give the mean age: “Primary igneous zircons from five pumice samples give a mean age of 1.87 ± 0.04 Myr”. ↩
30.
Roberts, R. G. et al. “Optical and radiocarbon dating at Jinmium rock shelter in northern Australia.” Nature 393:358–362. 1998. A 1998 Nature letter on the Jinmium rock shelter in Australia's Kimberley region. Thermoluminescence dating had given 116,000–176,000 years for the artefact-bearing deposit, but single-grain optical dating, which measures when individual sand grains were last exposed to light, reduced the age to under 10,000 years, in agreement with radiocarbon on charcoal. The article uses it as the case where a disagreement between methods raised the alarm. ↩
31.
Wendt, I., and C. Carl. “The statistical distribution of the mean squared weighted deviation.” Chemical Geology: Isotope Geoscience Section 86(4):275–285. 1991. A statistics paper on the mean squared weighted deviation (MSWD), the measure of scatter about an isochron, which is a straight line through dated samples whose slope gives an age. The article uses its published threshold to judge when the scatter is too large for the points to define an isochron.
1. 31a.
  Section Abstract: “If MSWD [>] 1+2(2/f)^1/2, there is only <5% probability that the data define an isochron.” ↩
32.
Vermeesch, Pieter. “On the treatment of discordant detrital zircon U–Pb data.” Geochronology 3:247–257. 2021. A 2021 methods paper by Pieter Vermeesch of University College London. It applies six ways of rejecting discordant zircon analyses (grains whose uranium-lead ages disagree, in detrital zircon meaning grains eroded from older rocks and redeposited in sediment) to a set of 70,869 grains, and finds that the log-ratio filter gives age spectra closest to the unfiltered data while the popular relative-age filter does worst. The article uses it to show that detrital zircon studies have no field-wide rejection rate.
1. 32a.
  Page 253: “As mentioned in the introduction, most current detrital zircon studies apply a 10 %–30 % relative age cutoff.” ↩
33.
Lanphere, M. A., and G. B. Dalrymple. “Identification of excess ⁴⁰Ar by the ⁴⁰Ar/³⁹Ar age spectrum technique.” Earth and Planetary Science Letters 32(2):141–148. 1976. A 1976 paper on how to recognize excess argon, which is argon already in a mineral when it formed, from the shape of an argon-argon age spectrum. The article uses it for a real limit of the method: the spectrum reveals excess argon but does not correct for it, and the article notes that the authors are the people who built the method.
1. 33a.
  Section Abstract: “saddle-shaped age spectra are diagnostic of excess 40Ar in igneous minerals as well as in igneous rocks. The minima in the age spectra approach but do not reach the crystallization age. Neither the age spectrum diagram nor the 40Ar/36Ar versus 39Ar/36Ar isochron diagram reliably reveal the crystallization age in such samples.” ↩
34.
Jicha, Brian R. et al. “Intercalibration of ⁴⁰Ar/³⁹Ar laboratories in China, the USA and Russia for Emeishan volcanism and the Guadalupian-Lopingian boundary.” National Science Review. 2021. A 2021 paper that compares argon-argon dating laboratories in China, the USA and Russia. The laboratories dated splits of one plagioclase (feldspar) separate, and Beijing gave 259.9 ± 1.2 Ma while Madison gave 259.8 ± 1.2 Ma (million years ago). The article uses it as the blind comparison between laboratories that tests the charge that labs tune their answers. ↩
35.
Phillips, David et al. “Revised astronomically calibrated ⁴⁰Ar/³⁹Ar ages for the Fish Canyon Tuff sanidine: closing the interlaboratory gap.” Chemical Geology 597:120815. 2022. A 2022 paper that recalculates the age of Fish Canyon Tuff sanidine, the most widely used argon-argon reference mineral, against astronomically tuned sediments. Its inter-laboratory mean age is 28.176 ± 0.010 Ma, and it says the method's precision is better than 0.05 % but is limited by uncertainties in the potassium decay constants and in the ages of reference minerals. The article uses it for the systematic floor that these uncertainties leave in argon-argon ages. ↩
36.
Horstwood, Matthew S. A. et al. “Community-Derived Standards for LA-ICP-MS U-(Th-)Pb Geochronology.” Geostandards and Geoanalytical Research 40(3):311–332. 2016. “ca. 2% 2s”. A 2016 paper by a community of geochronology laboratories that sets how uncertainties should be propagated in laser-ablation uranium-lead dating, and defines new reference values for zircon, monazite and titanite standards. Its stated inter-laboratory uncertainty limit is the quoted figure, in which 2s means two standard deviations. The article uses it as the community's limit for laser-ablation uranium-lead work. ↩
37.
Scott, E. M. et al. “A Report on Phase 2 of the Fifth International Radiocarbon Intercomparison (VIRI).” Radiocarbon 52(2–3):846–858. 2010. Table 3: Whale bone H, consensus 9,528 BP, reported from 4,810 to 9,880 BP across 55 results. “Up to 12% of the results for any 1 sample were identified as outliers (~10% of the total…).” A 2010 report on the second phase of an international test in which 42 laboratories dated the same five bone samples. The article uses its figures for the spread of results: about 10% of the returned results were outliers, and one laboratory reported 4,810 BP for a sample whose consensus was 9,528 BP (before present). ↩
38.
Woodmorappe, John. “Radiometric Geochronology Reappraised.” Creation Research Society Quarterly 16(2):102–129, 147. Sep 1979. Table 1, page 103: “a compilation of over 350 different radiometric dates that are very anomalous”. A review by a creationist author, who holds a bachelor's degree and has studied biology and geology, that tabulates over 300 radiometric dates which disagree with the fossil order of their rocks. It argues that most such discrepant results are not published, and the abstract says over 300 while the table caption says over 350. The article treats it as the founding document of the young-earth case against radiometric dating.
1. 38a.
  Page 113: “With few exceptions, Table 1 is confined to datings on material that is considered to be reliable.” ↩
2. 38b.
  Page 114: “Usually, however, glauconite K-Ar and Rb-Sr values are 'too young' for their biostratigraphic position. Many such anomalously young dates are entered into Table 1.” ↩
3. 38c.
  Page 113: “The uniformitarians may contend that there are many more dates in agreement with accepted values than there are anomalies such as all of Table 1. Even if this were true, it would not appear to be an overwhelming majority, and a significant minority of discrepant dates would probably be sufficient to discredit all of radiometric dating.” The "somewhat less than half" estimate is on the same page. ↩
39.
Faure, Gunter. Principles of Isotope Geology. New York: Wiley. 2nd edition. 1986. A university textbook on isotope geology, the methods that date rocks, including potassium-argon and argon-argon dating. The article cites it as the standard textbook of its day, which said that glauconite argon dates are not geologically meaningful.
1. 39a.
  Page 103: The conclusion, in the book's words: “that 40Ar/39Ar dates of glauconites are not geologically meaningful.” ↩
40.
Brown, Robert H. “Are Radioisotope Dating Methods Reliable?” Origins (Geoscience Research Institute) 25(2):104–108. 1998. A review, in the journal Origins, of Woodmorappe's book The Mythology of Modern Dating Methods. The reviewer argues that a long list of discordant dates does not prove that isotope ratios are never time significant, and that radioisotope ages can be explained on either uniformitarian or young-earth terms. The article cites it as a reply to Woodmorappe from inside the young-earth movement.
1. 40a.
  Page 104: “The reader should keep in mind that such a list, however long, is not proof that daughter/parent isotope ratios are never time significant.” ↩
2. 40b.
  Page 105: “Radioisotope ages may be reasonably explained on the basis of either uniformitarian geology or young-earth biblical geology. … A choice between an 'old-earth' concept and a 'young-earth' concept must be made on a basis other than success in the development of a reasonable explanation for radioisotope daughter/parent ratios.” ↩
41.
Snelling, Andrew A. “Radiometric Dating: Problems with the Assumptions.” Answers magazine. Answers in Genesis. Oct 2009. Part 2 of a three-part popular series in Answers magazine by Andrew Snelling. It argues that three assumptions behind radiometric dating cannot be tested: conditions at time zero, no contamination, and a constant decay rate. The article cites it as the popular version of the excess-argon point, set against Snelling's own technical papers.
1. 41a.
  Section Assumption 1: Conditions at Time Zero: “Yet lava flows that have occurred in the present have been tested soon after they erupted, and they invariably contained much more argon-40 than expected.” ↩
42.
Snelling, Andrew A. “Conflicting 'Ages' of Tertiary Basalt and Contained Fossilised Wood, Crinum, Central Queensland, Australia.” CEN Technical Journal 14(2):99–122. 2000. A technical paper on a basalt flow at Crinum, Queensland, with fossil wood inside it, dated by potassium-argon at 37 to 47 million years and by radiocarbon at 29,544 to 44,700 years (0.4 to 2.5 per cent modern carbon, after acid-base-acid pretreatment). The 30 million year 'expected' age is itself a potassium-argon age from Webb and McDougall 1967, on basalts elsewhere in the district. The article treats it as Snelling's most-cited result, and the one he draws his strongest conclusion from. ↩
1. 42a.
  Page 120: “The presence of the radiocarbon in the fossilised wood demonstrates that the enclosing basalt cannot be millions of years old and that the radioisotopic 'dating' is grossly in error.” ↩
2. 42b.
  Page 120: “While not providing the true age, the excessive radiocarbon 'age' is consistent with a stronger magnetic field and changes in atmospheric 12C levels around the Earth during and immediately after the Flood.” ↩
3. 42c.
  Page 110: “it could be that the graphite is not '14C-dead', and the 14C in it is there because the graphite is in fact not all that old.” ↩
43.
Bird, M. I. et al. “Radiocarbon Dating of 'Old' Charcoal Using a Wet Oxidation, Stepped-Combustion Procedure.” Radiocarbon 41(2):127–140. 1999. A laboratory study that dated charcoal known to be older than 50,000 years, cleaned in three ways before its carbon was counted: not at all, by acid-base-acid (ABA) pretreatment, and by the wet oxidation, stepped-combustion procedure named in the title (ABOX-SC). Its finding is that no single method removes all contaminants. The article uses it to show that acid-base-acid cleaning, the method Snelling's laboratories used, can leave enough modern carbon to make very old charcoal look younger than it is.
1. 43a.
  Page 137: “Neither the stepped combustion of an unpretreated sample nor an ABA pretreatment with stepped combustion is sufficient to remove all contaminants.” Table 1, page 134: Table 1 (p. 134) and p. 137 give the weighted means, for charcoal independently older than 50 ka (50,000 years), of 41.6, 46.3 and 51.3 ka for the untreated, ABA and ABOX-SC samples, at 0.56, 0.32 and 0.17 pMC. ↩
44.
Snelling, Andrew A. “Determination of the Radioisotope Decay Constants and Half-lives: Potassium-40 (⁴⁰K).” Answers Research Journal 9:171–196. 2016. A technical paper on the decay constant and half-life of potassium-40, the radioactive form of potassium used in potassium-argon dating. It compares two published half-life values for potassium-40, from Min et al. (2000a) and Renne et al. (2011). The article cites it for the paper's conclusion that both values were fitted to other dating methods.
1. 44a.
  Section Conclusions: “both the Min et al. (2000a) and Renne et al. (2011) total 40K decay half-life values ultimately depend on adjusting (that is, massaging) K-Ar and Ar-Ar ages to conform to U-Pb and Pb-Pb ages, often obtained from different minerals respectively in the same rocks.” ↩
45.
Walker, Tas. “The way it really is: little-known facts about radiometric dating.” Creation 24(4):20–23. Creation Ministries International. Sep 2002. A popular magazine article by Dr Tas Walker, of Creation Ministries International, on what radiometric dates can and cannot show. Walker argues that a geologist can always interpret a date to fit what he expects. The article quotes him for a concession: wrong dates from one area may still show a pattern from the Flood's lava flows.
1. 45a.
  Section The error is not the real error: “No matter what the radiometric date turned out to be, our geologist would always be able to 'interpret' it. He would simply change his assumptions about the history of the rock to explain the result in a plausible way.” ↩
2. 45b.
  Section Why use it?: “rocks in the same area which give similar 'dates' are likely to have formed from the same lava at about the same time during the Flood. So, although the assumptions behind the calculation are wrong and the dates are incorrect, there may be a pattern in the results that can help geologists understand the relationships”. ↩
46.
Wiens, Roger C. Radiometric Dating: A Christian Perspective. American Scientific Affiliation. 1994 (edition of 2002). A paper by a physicist with a minor in geology, then at Los Alamos National Laboratory, that explains in plain terms how the main dating methods work and how well their half-lives and dates are known. It argues that the methods agree with one another and that Christian doubts about them rest on misconceptions. The article cites it as a fair summary of the limit of potassium-argon dating on young rocks.
1. 46a.
  Section Misconception 14: “Anyone can move the hands on a clock and get the wrong time. Likewise, people actively looking for incorrect radiometric dates can in fact get them. Geologists have known for over forty years that the potassium-argon method cannot be used on rocks only twenty to thirty years old. Publicizing this incorrect age as a completely new finding was inappropriate.” ↩