Do the Dating Methods Agree?

Vardiman, Larry et al.  “RATE Responds to the Isaac Essay Review, with Isaac Replies.”  Perspectives on Science and Christian Faith 60(1):35–38.  Mar 2008.  A reply by the RATE group, the researchers who ran the project, signed collectively by all seven authors rather than by Humphreys alone, followed by Isaac's rejoinder.  The article uses it because the group challenges Isaac to document the "vast amount of concordance" he claims, and this article answers that challenge.

1. 1a.

   Page 35:  “Isaac needs to provide documentation from technical literature where vast amount of concordance is established.  Does he have examples of concordant isochrons between U-Pb, Sm-Sr, Rb-Sr and K-Ar in suites of earth rocks?  If he has such documentation of a vast amount of concordance, he could easily trivialize the RATE researcher's statements about discordant isochrons.  If Isaac could provide this documentation, he would have one of the strongest arguments in favor of the accuracy of radioisotope ages.”  This is the challenge the article answers. ↩ ↩

2. 1b.

   Page 35:  The discordances are.  “as large as factors of two or three in some cases, much larger than the 15% Isaac stated.”  RATE is right on the arithmetic.  Answers in Genesis' own published table gives Cardenas Basalt K-Ar 516 ± 30 Ma against Sm-Nd 1588 ± 170 Ma, a factor of 3.1. ↩

Beachy, Micah D. et al.  “How Often Do Radioisotope Ages Agree? A Preliminary Study of 29,000 Radioisotope Ages in the USGS National Geochronological Database.”  Proceedings of the Ninth International Conference on Creationism, 387–411.  2023.  A conference paper that scores how well the ages in the USGS National Geochronological Database agree with one another, using a "Concordance Score" from 0.00 (no pair of ages overlaps) to 1.00 (every pair overlaps).  It found an average score of 0.64 across the database, and a perfect 1.00 in only one of the 34 records with three or more methods, two of them not U-Th-Pb (2.9%).  Beachy and Kinard were sophomore undergraduates at Cedarville; Garner, a Fellow of the Geological Society of London at the Biblical Creation Trust, was one of seven Area Editors of the volume, though the acknowledgements thank Whitmore, Adam Hammett and three anonymous reviewers, so there is no sign that Garner reviewed this paper.  The volume's editor apologises in the preface that the editor's illness delayed the cycle and that "the peer-review process was rushed in too many cases."  The article calls it the best quantitative young-earth argument on this subject in print. ↩

1. 2a.

   Section Abstract, page 387:  “Only 34 records included ages using three or more different radioisotope methods where at least two were not U-Th-Pb, and only one of these (2.9%) had a concordance score of 1.00.” ↩

2. 2b.

   Section Abstract, page 387:  The sentences immediately preceding the one above:  “Overall, the average concordance score for the whole database is 0.64.  Only 4,875 of the 18,575 records (26.2%) included more than one calculated age.  Of these, 1,135 (23.3%) had a concordance score of 0.00 and 2,593 (53.2%) had a concordance score of 1.00.  There were 998 records with age determinations using three or more U-Th-Pb methods.  Concordance scores for these records ranged from 0.18 to 0.79, averaging 0.57.” ↩ ↩

3. 2c.

   Page 388:  “If two such ages had uncertainty ranges that overlapped, then they were considered concordant with one another…  For cases in which no uncertainty range was given, we assumed a range of ±10% of the reported age.  For comparison, the average uncertainty range in the database was ±6.3% of the corresponding age.”  And:  “By dividing the number of concordant comparisons by the number of comparisons in the record, the metric yields a concordance score for the record.” ↩

4. 2d.

   Page 395:  The authors' own caveat, which this article endorses: a record is not a rock unit, and the Pikes Peak Granite appears as roughly 20 separate records carrying about 90 ages.  The same discussion section says.  “there are clearly limits to what we can say with confidence about the prevalence of concordance”.  and calls the study.  “preliminary.” ↩

5. 2e.

   Section Results §D, Systematic Discordances, page 394:  “For the same rock units, radioisotope dating methods tended to yield ages from oldest to youngest in the following order:  Pb-Pb > Rb-Sr > ²³⁵U-²⁰⁷Pb > ²³⁸U-²⁰⁶Pb > ²³²Th-²⁰⁸Pb > K-Ar > FT.  This compares to the pattern of systematic discordances reported by the RATE group, in which Sm-Nd > U-Pb > Rb-Sr > K-Ar.”  Section Abstract, page 387:  “RATE found that β-decaying isotopes tended to yield younger ages than α-decaying isotopes; in our study ⁴⁰K followed this pattern but ⁸⁷Rb did not.” ↩

Isaac, Randy.  “Assessing the RATE Project.”  Perspectives on Science and Christian Faith 59(2):143–146.  Jun 2007.  An essay review of the second volume of the RATE project's book Radioisotopes and the Age of the Earth (2005), written by a physicist.  Isaac was then Executive Director of the American Scientific Affiliation, with a PhD in solid-state physics and 28 years at IBM.  The article uses it as the critic's claim that the RATE book failed to overturn the vast amount of concordance in the literature.

1. 3a.

   Page 144:  “fail to invalidate the vast amount of concordance”.  Isaac also characterises RATE's discordances as.  “ages that differ by as much as 10-15%”.  , which RATE correctly disputes in its 2008 reply. ↩

Computed for this article.  Independent implementation from the paper's published method description, not from its code:  ngdb_lib.py and ngdb_analysis.py in this article's research folder, Python 3 standard library only, re-runnable in about 95 seconds with python3 ngdb_analysis.py > ngdb_output.txt.  Results:  the same 34 record numbers, checked against the paper's Appendix C; 1 of 34 at score 1.00 (2.9%); mean score of the 34 = 0.3877 against their published 0.39; database-wide mean = 0.6404 against 0.64; mean reported uncertainty 6.38% against their 6.3% (6.31% if the one determination with a negative reported ± is left in); pool size 29,067 age determinations.  Their worked example (record 558, 33 of 36 pairs = 0.92) reproduces exactly.  All six tables parse to the row counts printed in Appendix 1 of the USGS report. ↩ ↩

Computed for this article.  One change at a time from the published treatment:  as published 2.9%; drop zero and negative ages 2.9%; give zero-width error bars the same default a blank gets 2.9%; harmonise decay constants 5.9%; count only cross-method pairs 5.9%; additionally drop determinations the USGS compilers flagged as not dating rock formation 35.7% (n = 14).  Sweeping the assumed default uncertainty from 1% to 20% moves the headline only from 0.0% to 5.9%, so the ±10% substitution is not the lever a critic would expect, and the article says so. ↩ ↩

Snelling, Andrew A.  “Radiometric Dating: Making Sense of the Patterns.”  Answers magazine 5(1 (Jan-Mar 2010)).  2010.  “Radiometric dating methods sometimes yield conflicting results, but the technique itself is scientific and reliable, and once the results are interpreted in a biblical framework, they yield clear patterns that help us better understand the earth's history since creation six thousand years ago”.  “the radiometric dating methods accurately show the top rock layer is younger than the layers beneath it”.  “And apart from the potassium-argon age for the Cardenas Basalt, all the radioactive clocks have correctly shown that these four rock units were formed earlier than Cambrian, so they are pre-Cambrian.”  Part 3 of the Radiometric Dating 101 series, and the strongest young-earth concession in the popular literature.  It reports that creationist researchers dated the same Grand Canyon rock units with all four common clocks and got different ages for the same rocks.  The article uses it for the concession that the clocks still get the order of the rock layers right. ↩

Dalrymple, G. Brent.  “Radiometric Dating Does Work!”  Reports of the National Center for Science Education 20(3).  2000.  The mainstream's most-cited short answer, printed in the Reports of the National Center for Science Education.  The article cites it as one of the mainstream's standard replies, which it calls overstated.

1. 7a.

   Table 2:  The Hell Creek Z-coal block is seven rows, not one: tektites, ⁴⁰Ar/³⁹Ar total fusion, 28 analyses, 64.8 ± 0.1 Ma; tektites, ⁴⁰Ar/³⁹Ar age spectrum, 1, 66.0 ± 0.5 Ma; tektites, ⁴⁰Ar/³⁹Ar age spectrum, 1, 64.7 ± 0.1 Ma; tektites, ⁴⁰Ar/³⁹Ar total fusion, 17, 64.8 ± 0.2 Ma; biotite and sanidine, K-Ar, 12, 64.6 ± 1.0 Ma; biotite and sanidine, Rb-Sr isochron (26 data), 1, 63.7 ± 0.6 Ma; zircon, U-Pb concordia (16 data), 1, 63.9 ± 0.8 Ma.  The cross-method gotcha is weaker than it looks: the 64.8 ± 0.1 argon age is on tektites while the 63.7 ± 0.6 rubidium-strontium isochron is on biotite and sanidine, and on the same material the K-Ar age 64.6 ± 1.0 Ma does overlap the rubidium-strontium one.  What does not overlap is 66.0 ± 0.5 against 64.7 ± 0.1, one method, one material, one locality, which is the counter-example the article uses.  Separately, the biotite and sanidine pairs in his Figure 1 are a cross-mineral check within one method, since the figure is captioned.  Figure 1:  “⁴⁰Ar/³⁹Ar ages for 8 volcanic ash beds in the Pierre Shale”.  and the piece does not say so.  His St Severin table is also looser than the text implies: the Sm-Nd isochron is 4.55 ± 0.33 Ga, a ±7.3% constraint. ↩

Wiens, Roger C.  Radiometric Dating: A Christian Perspective.  American Scientific Affiliation.  2002.  A popular introduction to radiometric dating for Christians, by Roger C. Wiens, a physicist whose PhD thesis was on isotope ratios in meteorites; the first edition appeared in 1994 and this is the 2002 revision.  It describes how the main dating methods work and how their results are checked against one another, and it is the piece most often handed to a doubting Christian.  The article uses it as that popular reference, whose central quantitative claim carries no citation.

1. 8a.

   Its three numbered tables are half-lives, extinct radionuclides and ice-core layer counts.  Page 10:  There is one unnumbered same-rock table:  “the ages, in billions of years, from twelve different studies using five different techniques on one particular rock formation in Western Greenland, the Amitsoq gneisses,”.  It is compiled from Dalrymple 1991: U-Pb 3.60 ± 0.05; Pb-Pb 3.56 ± 0.10, 3.74 ± 0.12, 3.62 ± 0.13; Rb-Sr 3.64 ± 0.06, 3.62 ± 0.14, 3.67 ± 0.09, 3.66 ± 0.10, 3.61 ± 0.22, 3.56 ± 0.14; Lu-Hf 3.55 ± 0.22; Sm-Nd 3.56 ± 0.20.  That is one rock formation, twelve studies, no uncertainties on the techniques and no sample-level pairing.  Page 19:  Its central quantitative claim carries no citation:  “The differences actually found in the scientific literature are usually close to the margin of error, usually a few percent, not orders of magnitude!”  It is a good popular explainer and thin as evidence. ↩

Computed for this article.  Of the 1,752 pairwise comparisons producing the published 2.9%, 814 (46.5%) compare two different isotopic systems and 938 (53.5%) compare one system against itself — ²⁰⁶Pb/²³⁸U against ²⁰⁷Pb/²⁰⁶Pb in one zircon, or biotite against hornblende in one K-Ar study. ↩ ↩

Computed for this article (null-model.py and null-34-exact.py, in this article's research folder).  If two independent, unbiased measurements of the same quantity carry 1σ error bars, the probability that the bars overlap is erf(1) = 0.8427.  Chaining that across the 36 pairs of a nine-age record — 0.8427³⁶ = 0.002 — is wrong, because the 36 pairwise events are all functions of the same nine measurements and are strongly dependent.  With equal 1σ bars, "all 36 pairs overlap" is exactly "the range of nine standard normals is at most 2", whose probability by Monte Carlo (200,000 draws) is 0.107.  Re-run on record 558's own nine ages with their own published ± values it is 0.106.  Applying the same simulation to all 34 records, on exactly the 1,752-pair pool the published algorithm builds, a set of perfectly accurate methods would produce 5.6 records at score 1.00 against the observed 1.  This is why the article treats the average score (0.64 / 0.57 / 0.39) as the informative statistic:  the count of perfect scores is not uninformative, it is informative in the direction opposite to the one an independent-pairs calculation suggests. ↩ ↩

Method mechanics

Snelling, Andrew A.  “The Relevance of Rb-Sr, Sm-Nd and Pb-Pb Isotope Systematics to Elucidation of the Genesis and History of Recent Andesite Flows at Mt Ngauruhoe, New Zealand.”  Proceedings of the Fifth International Conference on Creationism, 285–304.  2003.  Page 285:  “did not yield any 'age' information, although the Pb isotopic data are strongly linear.”  The "selective manipulation of the data" sentence is in RATE Volume II, p. 435.  A conference paper by Andrew Snelling of the Institute for Creation Research, also chapter 6 of RATE Volume II.  It analysed andesite lava from Mt Ngauruhoe in New Zealand, erupted in 1949 and 1954 with avalanche deposits in 1975, and found that its isotope ratios were inherited from the mantle and crust rather than giving an age.  The article uses it because RATE's own dates for these flows came out at 133 to 3,908 million years, and its author says the isotope data gave no age information. ↩

1. 11a.

   In RATE Volume II, p. 414, Table 2, for lava erupted in 1949, 1954 and 1975: Rb-Sr 133 ± 87 Ma, Sm-Nd 197 ± 160 Ma, Pb-Pb 3908 ± 390 Ma.  In RATE Volume II, p. 422, Table 3, the whole ²⁰⁶Pb/²⁰⁴Pb range across the eleven whole rocks is 18.797 to 18.842, a span of 0.045, against RATE's own stated 2σ per-point error of 0.45%, which is ±0.085 on 18.8.  The entire data spread is half the error bar on one point. ↩

Computed for this article (decay-constancy.py), and independently reproducing Uzan's ²³⁸U sensitivity at +551 from the Gamow factor plus the Q-value term.  Setting Δα/α = +3.78%, enough to make ²³⁸U decay 10⁹ times faster:  ²³⁵U runs 22 times slower, ¹⁴⁷Sm 5,200 times faster, ¹⁸⁷Re effectively stops.  Because an apparent age scales with the acceleration factor, concordance between two methods measures the ratio of their accelerations, whenever the acceleration occurred — which is why the standard "your bounds are time-averaged, the pulse lasted a year" reply does not apply.  Note:  this corrects an earlier figure in this project's own notes, which put the requirement at |Δα/α| ≥ 0.12 using the Coulomb-barrier term alone; the Q-value term dominates and flips the sign, and the corrected 3.78% makes every observational limit bite harder. ↩ ↩

Failure modes

Chaffin, Eugene F.  “Accelerated Decay: Theoretical Considerations.”  Radioisotopes and the Age of the Earth II:525–585.  Institute for Creation Research / Creation Research Society.  2005.  A theory chapter, the seventh in RATE's second volume, by Eugene Chaffin, a physicist at Bob Jones University, on whether nuclear decay rates could have been faster in the past.  It is RATE's own physics chapter and the source of several of the numbers that sink the model.  The article uses it because the RATE project's own physicist supplied the sensitivity figures and admissions that undercut the model.

1. 13a.

   Page 566, table 2:  Chaffin's Table 2, captioned.  “Sensitivity of various forbidden β-decays to changes in decay energy (after Dyson, [1972])”.  The table tabulates ∂lnλ/∂lnα, each decay rate's sensitivity to a change in the fine-structure constant, for the β-decayers: −30 for ⁴⁰K, −180 for ⁸⁷Rb and −18,000 for ¹⁸⁷Re.  These are negative, while the alpha emitters' sensitivities are large and positive, so a single change in the fine-structure constant cannot accelerate both families. ↩ ↩ ↩

2. 13b.

   Page 528:  “If the accelerated decay needed to explain the data is restricted to about one year or so, then a change in the decay constant of 10⁹ or so may be required [DeYoung, 2000]. Thus these considerations seemed to indicate that a one-year episode of accelerated decay at the time of the Flood may not be enough.”  Page 572:  His own ceiling for alpha decay is a change.  “in some cases could be as much as a factor of 10⁸”.  and those results.  “remain exploratory in nature”.  Page 573:  On the beta side:  “Whether the weak force can vary in a way consistent with observation seems doubtful”.  Page 540:  On his own Oklo samarium study:  “Half-life differences of more than one order of magnitude would seem to be ruled out, unless two or more factors may have changed concurrently, with each effect canceling the other”. ↩

3. 13c.

   Page 567:  Chaffin's explanation for Austin and Snelling's measured K-Ar-versus-Rb-Sr discordance of about 1.6 is the forbiddenness difference:  “This should help explain the differences between K-Ar results and Rb-Sr results which Austin and Snelling have been finding.”  The next paragraph adds:  “Furthermore, the mathematics shows that Re-Os decays should be the most sensitive of all to accelerated decay, by a factor of about 100 over Rb-Sr and 600 over K-Ar (see Table 2).”  Because these are logarithmic derivatives, the same mechanism run at the strength the model needs puts Rb-Sr at 10⁵⁴ and Re-Os at 10⁵⁴⁰⁰ when K-Ar is at 10⁹. ↩

Computed for this article from Austin's Volume II chapter 5 data on a single Beartooth Mountains amphibolite hand sample.  Ranked by the closure temperature of the carrier phase, the seven ages fall in strict monotonic order:  Sm-Nd (titanite) 2886, Pb-Pb (titanite) 2689, K-Ar hornblende 2620, Rb-Sr (biotite) 2515, K-Ar biotite 2403, K-Ar whole rock 2011, K-Ar quartz-plagioclase 1520 Ma.  Spearman ρ = +1.000; exact one-sided permutation probability 1/5040 = 2.0 × 10⁻⁴.  The closure temperatures used are compilation figures — the primaries (Cherniak, Harrison, Dodson, Reiners & Brandon, Kohn) are paywalled and were not obtained — but the test depends only on the undisputed rank order of those temperatures, not on their absolute values. ↩ ↩

Oard, Michael J.  “Radiometric dating and old ages in disarray.”  Journal of Creation 20(2):36–41.  Aug 2006.  The web version was posted on 9 May 2008.  A book review, in the creationist Journal of Creation, of Volume II of RATE's book Radioisotopes and the Age of the Earth.  RATE was an eight-year, $1.25 million research project by young-earth researchers, and the review argues that the assumptions behind radiometric dating are violated.  The article uses it as a young-earth statement of those assumptions, and it is cited here as the book review it is.

1. 15a.

   “There are many assumptions behind radiometric dating, but there are three main ones.  Uniformitarian scientists assume (1) the initial isotope amounts are known, (2) the decay rate has remained constant at today's rate, and (3) the sample has remained in a closed system for millions and billions of years.  Evidence is presented that all three assumptions are violated in various contexts, but the RATE project concludes that the assumption of constant decay at today's rates is the most significant wrong assumption.” ↩

2. 15b.

   Oard's summary of why the isochron method is considered superior, verbatim:  “Austin used the isochron technique that employs different minerals from the same rock.  The isochron method is considered superior because a straight line on the isochron plot informs us that two of the three main assumptions of radioactive dating (the closed system and initial conditions assumptions) are supposed to be validated.”  Oard's review also concedes the heat problem ("would raise the crust to 22,000 K, more than enough to melt the whole crust and boil away the oceans") and the radiation problem ("just the radioactive potassium-40 in the bodies of those animals and the people on the ark would kill them when decay accelerated"). ↩

Snelling, Andrew A.  “Radiometric Dating: Problems with the Assumptions.”  Answers magazine 4(4 (Oct-Dec 2009)).  2009.  Part 2 of a three-part Answers magazine series called Radiometric Dating 101, by Dr. Andrew A. Snelling.  It names the three assumptions behind radiometric clocks, that the starting amounts are known, that the decay rate has stayed constant and that the rock has stayed closed to outside material, and it explains how each can lead to a wrong date.  The article uses it as the source of the three-assumptions list and of the author's own concession on isochrons.

1. 16a.

   Snelling's own concession, in the article that names the three assumptions: an isochron is.  “still based on unproven assumptions 2 and 3”.  that is, the initial-conditions assumption is not among them. ↩

Clarey, Tim, and Vernon Cupps.  “Does Radioisotope Dating Prove an Old Earth?”  Acts & Facts.  Institute for Creation Research.  May 28, 2021.  A question-and-answer article in the Institute for Creation Research's magazine Acts & Facts, by Tim Clarey and Vernon Cupps, both with PhDs.  It lists the common radioisotope methods, states the assumptions each one makes, and argues that a constant decay rate cannot be taken for granted.  The article uses it as a current popular statement of the three assumptions.

1. 17a.

   Clarey and Cupps state the three assumptions for the methods generally:  “each method has to assume a starting amount of both parent and daughter isotopes”.  and list Rb-Sr, Sm-Nd, U-Pb and Pb-Pb among those methods, which in practice are isochron methods.  As applied to isochron work that is wrong: the isochron measures the initial ratio rather than assuming it, and Snelling concedes as much.  Their separate criticism of isochrons is the mixing argument, which this article agrees with:  “the raw data appear to be better explained by isotope mixing.”  See also Cupps, "Revisiting the Isochron Age Model, Part 1," Acts & Facts 49(6):10–13, 2020. ↩

Wendt, I., and C. Carl.  “The statistical distribution of the mean squared weighted deviation.”  Chemical Geology: Isotope Geoscience Section 86:275–285.  1991.  The source of the standard acceptance window for the MSWD test, which is MSWD < 1 + 2√(2/f), where f is the number of degrees of freedom.  The article uses it for that formula and tests how well the formula holds at small sample sizes. ↩

Computed for this article (mswd-window.py).  Wendt & Carl's 1 + 2√(2/f) is a normal approximation that runs slightly generous at small degrees of freedom:  for f = 8 it gives 2.000 where the exact χ² 95th percentile is 1.938; for f = 48, 1.408 against 1.358.  The script also confirms that Powell et al. 2020's Table 1 is tabulated in √MSWD rather than MSWD — squaring their one-sided column reproduces the exact χ² quantiles to three decimals — so anyone quoting that table as MSWD values is off by a square. ↩

Vermeesch, Pieter.  “High MSWDs are not the problem, low ones are.”  National Science Review 12:nwaf036.  2025.  A short perspective by a geochronologist, Pieter Vermeesch, who uses six synthetic datasets rather than new measurements to argue that a low MSWD is not a sign of good data.  The article uses it for a publication-bias argument made by a mainstream geochronologist, which cuts for the sceptic.

1. 20a.

   “High MSWD datasets are virtually absent from high-impact publications,”.  having been deemed.  “unpublishable”.  by reviewers, which.  “may tempt some scientists to 'cherry pick' data.” ↩

2. 20b.

   “High MSWD datasets are the rule, not the exception,”.  and low ones.  “should not be celebrated but approached with a healthy dose of scepticism.” ↩

Field, D., and A. Råheim.  “A geologically meaningless Rb-Sr total rock isochron.”  Nature 282:497–499.  1979.  A paper in Nature describing a partially reset rubidium-strontium system that gave a clean straight line, dating an event that never happened, at 1259 ± 26 million years with an MSWD (a measure of scatter around the line) of 1.58.  The article uses it because the general point, that a clean line can be an age of nothing, was made in Nature in 1979, though by partial resetting rather than mixing.  The mixing objection is not a creationist invention. ↩

Zheng, Yong-Fei.  “Influences of the nature of the initial Rb-Sr system on isochron validity.”  Chemical Geology: Isotope Geoscience Section 80:1–16.  1989.  A paper on whether the starting rubidium-strontium composition of a rock can make an isochron misleading.  Its conclusion, that Rb-Sr data alone cannot tell a valid isochron from an apparent one, is the sentence the article quotes, and the article uses it because it argues for cross-checking between dating systems, the subject of this article.

1. 22a.

   Section Abstract (last sentence):  “As it is impossible to distinguish a valid isochron from an apparent isochron in the light of Rb-Sr isotopic data alone, caution must be taken in explaining the Rb-Sr isochron age of any geological system.”  Snelling (Volume II, published 2005, p. 429) gives Zheng's conclusion in his own words, without quotation marks, and the young-earth literature has circulated his wording as though it were Zheng's. ↩

2. 22b.

   Zheng is a geochemist of the Chinese Academy of Sciences who co-authored the characterisation of the 91500 zircon U-Pb reference material used in laboratories worldwide, and who publishes on Triassic ultrahigh-pressure metamorphism.  Zheng accepts deep time without qualification.  The paper's conclusion is an argument for cross-checking between systems, and the qualifier.  “in the light of Rb-Sr isotopic data alone”.  is inside the sentence the young-earth literature quotes. ↩

Wendt, I.  “Isochron or mixing line?”  Chemical Geology 104:301–305.  1993.  A paper that establishes a diagnostic for telling a mixing line from a true isochron.  Mixtures lie on a plane in three dimensions (⁸⁷Rb/⁸⁶Sr, ⁸⁷Sr/⁸⁶Sr and 1/⁸⁶Sr), and that plane's parameters recover both the true age and the mixing slope.  The article uses the test to show how a mixing line can imitate an isochron, and it reports the failure mode the paper itself gives. ↩

1. 23a.

   Wendt's own statement of the test's failure mode, verbatim:  “If the Rb and Sr concentrations of the samples are well correlated, however, the data points are arranged on a straight line in the x,y,z-space at any time and an appearantly 'perfect' isochron with a possibly wrong age is obtained.”  The spelling is the author's. ↩

Computed for this article (mixing-vs-isochron.py).  A two-component mixture of a Sr-poor radiogenic crustal endmember with a Sr-rich mantle one gives, on the ordinary isochron plot, R² = 0.99999918, an entirely plausible initial ratio of 0.70331, and a fake "age" of 261 Ma — while giving R² = 1.00000000 on ⁸⁷Sr/⁸⁶Sr against 1/Sr, which is the diagnostic.  The script also confirms a second blind spot Wendt does not state:  if the two endmembers happen to have equal Sr concentrations, 1/Sr is constant across the series and the diagnostic has nothing to work with. ↩

The best cross-method measurements

Jaffey, A. H. et al.  “Precision measurement of half-lives and specific activities of ²³⁵U and ²³⁸U.”  Physical Review C 4:1889–1906.  1971.  A precise 1971 measurement of the half-lives of the two uranium isotopes, and still the source of both uranium decay constants, which set how fast each isotope decays.  The measured half-lives are 4.4683 and 0.70381 billion years (Gyr), with 2σ uncertainties (a measure of how well each value is pinned down) of 0.107% on λ²³⁸ and 0.137% on λ²³⁵.  The article uses the ratio of the two half-lives, 6.349, which is what makes the uranium-lead pair a genuine double clock. ↩

Condon, Daniel et al.  “Recommendations for the reporting and interpretation of isotope dilution U-Pb geochronological information.”  GSA Bulletin 136(9/10):4233–4251.  2024.  Recommendations written by 37 geochronologists on how laboratories should report and interpret the most precise uranium-lead dates, from the ID-TIMS method (isotope dilution thermal ionization mass spectrometry).  The article calls it the field's own community-standards paper and the most candid mainstream source in the article.

1. 26a.

   Lead loss.  “may be undetectable at low Pb*/Pbc levels due to large uncertainties resulting from the Pbc correction, and in rocks less than ca. 300 Ma, where Pb-loss trajectories parallel the concordia curve.”  This is the strongest honest concession in the article: in Phanerozoic rocks, the younger part of the rock record, the concordia test is close to blind to small Pb loss. ↩

2. 26b.

   On chemical abrasion: it.  “should nearly always be used,”.  but.  “it should be kept in mind that this procedure is imperfect, and there are plenty of examples where individual zircon grains have been interpreted to carry residual Pb-loss following chemical abrasion.”  Seven citations follow. ↩

3. 26c.

   “Decay constants used in other radioisotopic dating methods (e.g., Rb-Sr, ⁴⁰Ar/³⁹Ar, and Re-Os) have been intercalibrated with U-Pb dates from the same geological materials.”  The circularity charge is real at the sub-percent level, and the article states it in its own voice rather than waiting to be caught.  Schoene's 2014 Treatise chapter puts a number on the residual:  “systematic discordance in the U-Pb system of about 0.3% exists within analyses spanning >3 Ga, suggesting that one or both of the mean values of the uranium decay constants are inaccurate… it is impossible to determine whether the inaccuracy exists in one or both of the decay constants.” ↩

Mattinson, James M.  “Zircon U-Pb chemical abrasion (CA-TIMS) method: combined annealing and multi-step partial dissolution analysis for improved precision and accuracy of zircon ages.”  Chemical Geology 220:47–66.  2005.  A methods paper on the zircon chemical abrasion technique, which changed the discordance statistics by removing the damaged domains that leak lead, not by diagnosing them.  The article cites it for the point that chemical abrasion largely removes the lead-loss problem rather than diagnosing it. ↩

Hiess, Joe et al.  “²³⁸U/²³⁵U systematics in terrestrial uranium-bearing minerals.”  Science 335(6076):1610–1614.  2012.  A measurement study of the uranium-238 to uranium-235 ratio in 58 uranium-bearing minerals, from zircon to monazite, with 141 separate measurements.  The ratio assumed invariant at 137.88 for thirty years measures 137.818 ± 0.045 in zircon, and is genuinely variable at the several-per-mil level, with more than 5 per mil between samples.  The revision is the self-check working; the article uses it to show that the pre-2012 literature is uncorrected. ↩

Schaen, Allen J. et al.  “Interpreting and reporting ⁴⁰Ar/³⁹Ar geochronologic data.”  GSA Bulletin 133(3–4):461–487.  2021.  A community paper, written by a large group of argon-dating specialists, on how to interpret and report argon-argon ages.  It asks for stricter tests before a step-heating spectrum counts as a plateau, because modern experiments run 15 to 40 steps where the old rules were written for 5 to 10.  The article uses it for that stricter plateau rule, which replaces the older 1977 rule.

1. 29a.

   The old rule traces to Fleck et al. 1977.  The wording that circulates is.  “3 or more contiguous heating steps comprising 50% or more of the ³⁹Ar_K released and overlapping at the two sigma confidence level”.  , and it comes from the New Mexico laboratory's data page, not Fleck's own.  That page adds:  “The NMGRL no longer relies on the plateau definition of Fleck et al (1977), but rather tests the statistical precision of a number of contiguous heating steps using the criteria of Mahon (1996).”  Page 475:  Schaen et al. 2021 raise the bar for the community, in the printed paper:  “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…; 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).”  Their reason is that modern experiments run 15-40 steps rather than the 5-10 the old rule was written for. ↩

New Mexico Geochronology Research Laboratory.  K/Ar and ⁴⁰Ar/³⁹Ar Methods.  New Mexico Bureau of Geology and Mineral Resources.  n.d.  A laboratory's own web page explaining the potassium-argon and argon-argon dating methods, published with the New Mexico Bureau of Geology and Mineral Resources.  The article uses it for a worked example of an inverse isochron catching a 28% error, and for its statement that a primary standard must first be dated by another method.

1. 30a.

   “In order for an age to be calculated by the ⁴⁰Ar/³⁹Ar technique, the J parameter must be known…  Because this (primary) standard ultimately cannot be determined by ⁴⁰Ar/³⁹Ar, it must be first determined by another isotopic dating method…  However, while it is often easy to determine the age of the primary standard by the K/Ar method, it is difficult for different dating laboratories to agree on the final age.” ↩

2. 30b.

   Section Worked example on the data page:  The laboratory's own worked example, on its data page at https://geoinfo.nmt.edu/labs/argon/data/home.html: a groundmass concentrate whose age spectrum ranged over 4.5-6.0 Ma with no plateau; the inverse isochron gave a trapped ⁴⁰Ar/³⁶Ar of 312 ± 3 rather than the atmospheric 295.5, and a corrected age of 4.05 ± 0.01 Ma against roughly 5.2 Ma from the spectrum; a ~28% error found and fixed by the check. ↩

Dodson, M. H.  “Closure temperature in cooling geochronological and petrological systems.”  Contributions to Mineralogy and Petrology 40(3):259–274.  1973.  A theory paper by Martin H. Dodson of the University of Leeds on how a mineral's radiometric clock is set as the rock cools.  It derives a formula for closure temperature, the temperature of a system at the time its apparent age refers to, and works out 280 to 300 °C for rubidium-strontium dates on Alpine biotite, assuming grains about 0.5 mm across.  The article uses it for the exact definition of closure temperature, which marks when the clock starts, not when the rock formed.

1. 31a.

   Page 259:  The first sentence of the abstract:  “Closure temperature (T_c) of a geochronological system may be defined as its temperature at the time corresponding to its apparent age.”  Page 260:  Restated:  “The effective closure temperature, Tc, can therefore be defined as the temperature of the system at the time represented by its apparent age.”  The clock starts when the mineral stops leaking, not when the rock forms. ↩

Nebel, Oliver.  “Rb-Sr Dating.”  Encyclopedia of Scientific Dating Methods, 687–694.  Springer.  2014.  An encyclopedia entry on the rubidium-strontium dating method by Oliver Nebel of the Australian National University, covering how an isochron, a straight line through several minerals from one rock, turns measurements into an age.  It notes that uranium-lead dating of zircon can date similar events to higher precision, and that the Great Dyke of Zimbabwe, which cooled very slowly, gives rubidium-strontium ages about 23 million years younger than its uranium-lead ages.  The article uses the Great Dyke example to show that slowly cooled rocks give ages that differ in a specific, predictable order.

1. 32a.

   Page 691:  The Great Dyke of Zimbabwe cooled slowly enough that its Rb-Sr ages come out about 23 Myr younger than U-Pb ages on the same rocks, read by the field as a cooling history rather than as method failure.  This is the honest version of the closure-temperature defence, and it is testable precisely because it predicts an ordering. ↩

Mitchell, William S. III.  High-resolution U-Pb Geochronology of Terrestrial Cretaceous-Paleogene and Permo-Triassic Boundary Sequences in North America.  University of California, Berkeley.  2014.  Mitchell's own caveats are quoted in the article's discussion of selection:  “the majority of samples analyzed show clear and pervasive inheritance, even among the most pristine zircon grains,”.  “Choosing the right tephra can make a big difference, and often there is no way to tell which tephra is the right one until the mass spectrometry is underway.”  The X-coal outlier rests on two zircon grains and he flags it himself.  A PhD dissertation giving high-resolution uranium-lead dates for volcanic ash beds (tephras) at the Cretaceous-Paleogene and Permo-Triassic boundaries in North America.  It was written to be compared against the Berkeley argon laboratory's ages on the same beds, and it also allows comparison with argon-argon geochronology.  The article uses it for the uranium-lead half of the Hell Creek comparison, which is set against Sprain's argon half. ↩

Sprain, Courtney J. et al.  “Calibration of chron C29r: New high-precision geochronologic and paleomagnetic constraints from the Hell Creek region, Montana.”  GSA Bulletin 130(9/10):1615–1644.  2018.  Page 1622:  On where the analyses were made:  “The ⁴⁰Ar/³⁹Ar analyses were performed at the BGC”.  Page 1633:  Its Table 2 supplies the X-coal age used in the article's table, 65.704 ± 0.036 Ma on sample LG13-1, and withdraws the earlier figure on the same page:  “Based on the placement of the reversal in this section, we believe the MC11-3 age to be in error.”  The other six sanidine ages in the article's table are not tabulated in this paper.  They come from its companion, Sprain, Renne, Wilson & Clemens, "High-resolution chronostratigraphy of the terrestrial Cretaceous-Paleogene transition and recovery interval in the Hell Creek region, Montana," GSA Bulletin 127(3-4):393–409, 2015, doi:10.1130/B31076.1.  Its Table 1 gives U 64.866 ± 0.023, W 65.118 ± 0.024, Y 65.741 ± 0.022 and Null 66.289 ± 0.051 Ma, and its Table 2 gives the pooled HFZ 65.973 ± 0.020 and pooled Z 66.013 ± 0.015 Ma, with uncertainties excluding systematic sources.  The X coal in that 2015 Table 1, 65.491 ± 0.032 Ma, is the figure the 2018 paper withdraws.  The same two tables are reprinted as Tables 2.1 and 2.2 of Sprain's 2016 Berkeley dissertation.  One wrinkle the article's table does not show: the 2018 paper's own pooled Z-coal age is 66.024 Ma, 11 kyr older than the 2015 value used here.  The argon half of the Hell Creek comparison, reporting 18 new high-precision argon-argon dates from the Hell Creek region of Montana.  It is the source of the argon ages the article uses, and it states where its argon analyses were made. ↩ ↩

Kuiper, K. F. et al.  “Synchronizing Rock Clocks of Earth History.”  Science 320(5875):500–504.  2008.  A research report in Science that calibrated the argon-argon standard, Fish Canyon sanidine, against astronomically dated volcanic ash layers (tephras) in marine rocks from Morocco.  The astronomical calibration gives Fish Canyon sanidine 28.201 ± 0.046 Ma (2σ), and it reduced the argon-argon method's absolute uncertainty from about 2.5 to 0.25 percent.  The article lists it as one of the competing calibrations of that standard. ↩

1. 35a.

   Page 501:  “Comparison of conventional ⁴⁰Ar/³⁹Ar and U/Pb ages for diverse rock types over more than 3 billion years of geological time demonstrates a systematic offset, in which the U/Pb ages are older by 0 to 1% than the ⁴⁰Ar/³⁹Ar ages for the same rocks, although scatter in the offset suggests that some of the differences may result from interlaboratory biases or geological complexities.”  A systematic that everyone publishes is not a concealed problem. ↩

Renne, Paul R. et al.  “Response to the comment by W. H. Schwarz et al. on 'Joint determination of ⁴⁰K decay constants and ⁴⁰Ar*/⁴⁰K for the Fish Canyon sanidine standard.'.”  Geochimica et Cosmochimica Acta 75(17):5097–5100.  2011.  A reply by the Berkeley Geochronology Center group that made the calibration, answering a published comment on its 2010 work on the potassium decay constants.  It drops a disputed liquid-scintillation measurement from the calculation, and says Phanerozoic ages change by less than 0.041 percent.  The original determination is Renne et al., GCA 74(18):5349–5367, 2010, doi:10.1016/j.gca.2010.06.017.  The article cites it as the competing calibration giving 28.294 ± 0.036 Ma (1σ). ↩

Computed for this article (cross-method-diffs.py).  The two candidate Fish Canyon sanidine ages differ by 0.329%.  Propagating that choice through the same argon measurements moves the K-Pg boundary from 66.052 ± 0.008/0.043 Ma (Renne 2011 calibration) to 65.845 ± 0.053 Ma (GTS2012 calibration), a shift of 0.314%.  Historical series of assigned values for the same standard:  27.84 (used 1992-93), 28.02 ± 0.56 (Renne 1998), 28.05 (Min 2000 preferred), 28.201 ± 0.046 (Kuiper 2008), 28.294 ± 0.036 (Renne 2011) — extremes spanning 1.6%, the last two spanning 0.33%. ↩

Renne, Paul R. et al.  “⁴⁰Ar/³⁹Ar Dating into the Historical Realm: Calibration Against Pliny the Younger.”  Science 277(5330):1279–1280.  1997.  A two-page report in which laser heating of sanidine from the white pumice of the AD 79 eruption of Vesuvius was dated by argon-argon methods, using 46 analyses, and the result was compared with the known date.  The isochron age of 1925 ± 94 years is 0.36 percent high against a true age of 1918, with the true value inside the error bar, while the total-gas age of the same sanidine, 3300 ± 500 years, is 72 percent too old.  The article uses it because the eruption's date is fixed by Pliny the Younger's letters, which gives a known age to test the method against. ↩

Burgess, Seth D. et al.  “High-precision timeline for Earth's most severe extinction.”  PNAS 111(9):3316–3321.  2014.  A PNAS paper that dates five volcanic ash beds at the Permian-Triassic boundary stratotype at Meishan, China, by uranium-lead zircon dating, to build an age model for the end-Permian mass extinction.  The boundary is dated at 251.902 ± 0.024 Ma, and the extinction interval runs from 251.941 ± 0.037 to 251.880 ± 0.031 Ma, which is 60 ± 48 thousand years in the abstract (0.061 ± 0.048 Ma in Table 1).  Table 1 reports each bed with both an analytical uncertainty (0.015%) and a full external one (0.11%), and the article uses the paper for the rule on which uncertainty to quote.

1. 39a.

   “If calculated dates are to be compared with other U-Pb laboratories not using the EARTHTIME tracer, then ±y should be used for each laboratory.  If compared with other chronometers such as Ar-Ar or astrochronology, then ±z should be used.”  This is the rule most participants in the age-of-the-earth debate, on both sides, do not apply. ↩

Min, Kyoungwon et al.  “A test for systematic errors in ⁴⁰Ar/³⁹Ar geochronology through comparison with U/Pb analysis of a 1.1-Ga rhyolite.”  Geochimica et Cosmochimica Acta 64(1):73–98.  2000.  A laboratory study that dated the 1.1-billion-year-old Palisade Rhyolite by argon-argon on alkali feldspar and by uranium-lead on zircon, then asked whether the two agree once systematic errors are counted.  It is the cleanest single argon-argon against uranium-lead test: Palisade Rhyolite alkali feldspar gives 1088.4 ± 4.0 Ma against zircon ²⁰⁷Pb/²⁰⁶Pb 1097.6 ± 2.1 Ma, a 0.84% difference.  The article cites it for the point that the two methods agree once their systematic errors are included.

1. 40a.

   Section 5, page 93:  “On the basis of conventional decay constants and ignoring systematic errors, the ⁴⁰Ar/³⁹Ar age (1088.4 ± 4.0 Ma) of the F239 alkali feldspar is distinct from the ²⁰⁷Pb/²⁰⁶Pb zircon age (1097.6 ± 2.1 Ma).  However, including systematic errors in both U/Pb (1097.6 ± 5.3 Ma) and ⁴⁰Ar/³⁹Ar (1088.4 ± 23.3 Ma) ages shows that they are indistinguishable.”  The root cause: ²³⁸U's decay constant is known to 0.107% at 2σ — Jaffey et al. 1971's figure, restated at Schoene 2014 p. 359, not Min et al.'s — while ⁴⁰K's is known to about 2% at 2σ. ↩

Computed for this article (cross-method-diffs.py), from fifty same-material method-versus-method pairs assembled from the primary papers, with each pair's source recorded.  Median relative difference 0.76%; 31 of 50 under 1%.  The seven Hell Creek tephra pairs (Ar-Ar sanidine against CA-ID-TIMS U-Pb zircon) run 0.020% to 0.237% with a mean of 0.148%; six of the seven are under 0.21%.  Every one of the ten worst pairs is a lunar or meteoritic sample with a young argon age, a 1970s measurement, or a conventional total-gas K-Ar age on a historic flow.  Relative difference throughout is |a − b| divided by the mean of the pair.  On that convention the Vesuvius total-gas age scores 53.0%;  quoted against the true age of 1,918 years it is 72% too old.  The Mt Etna 122 BC flow's whole-rock potassium-argon age scores 197% on the mean and 11,857% against the true age. ↩

Computed for this article from the NASA Lunar Sample Compendium entries and the primary chronology papers.  On well-studied Apollo rocks, Rb-Sr, Sm-Nd, Pb-Pb and U-Pb agree within a few tenths of a percent to a few percent, while ⁴⁰Ar/³⁹Ar and K-Ar run systematically younger — 5.3% (60025), 10.6-10.9% (10017, 62236, 77215), and 41.8% for the 10017 whole-rock K-Ar (2.35 against 3.59 Ga), all as |difference| over the mean of the pair, the same convention as the fifty-pair table.  The direction is always the same.  Dalrymple concedes the 76535 Rb-Sr/Ar-Ar discrepancy in Ancient Earth, Ancient Skies (2004), p. 128, and writes that "the cause of this discrepancy is not fully understood."  His own explanation runs the other way from the one a reader might expect:  the younger argon ages "are due to elevated temperatures within the lunar crust between 4.5 and 4.2 Ga," and what needs explaining is why the rubidium-strontium clock survived that.  His answer:  "the Rb-Sr system in the mineral olivine actually resides in microscopic inclusions of other minerals within the olivine grains," and the olivine "acts as an impenetrable barrier to rubidium and strontium migration even during intense metamorphism or reheating."  Apollo 15555 has a crystallisation age of about 3.3 Ga and a ⁸¹Kr cosmic-ray exposure age of 81 Myr — different clocks measuring different events, and conflating them is the standard error in popular treatments. ↩

The young-earth position

RATE

Pogge, R. David.  “All Published Moon Rock Ages.”  Science Against Evolution.  2008.  All 116 rows are Apollo 11 samples, eighteen of them.  The workbook's second sheet heads its source list "References from Science, 30 January, 1970 'The Moon Issue' dedicated to the proceedings of the Apollo 11 Lunar Science Conference" and numbers nine chronology papers running pp. 461–480, printing no volume number, though the volume is Science 167.  Minimum row: sample 10069, 0.04 Ga, "Cosmic ray exposure."  Maximum row: sample 10084, 8.2 Ga, "²⁰⁸Pb/²³²Th."  Seven rows read "⁴⁰Ar/³⁹Ar low temp" and all seven are printed as upper limits: <1.0, <0.25, <0.75, <0.2, <0.8, <1.0 and <0.6 Ga.  The sheet gives no rock types: calling 10084 a soil and 10060 a breccia is this article's identification from the sample numbers, not the page's.  A website called Science Against Evolution has a spreadsheet of 116 published ages for Apollo 11 moon samples, drawn from nine chronology papers that Science printed in 1970.  It mixes ages of several kinds in one list, including cosmic-ray exposure ages, argon release steps and lead model ages.  The article uses it to show that the list combines quantities that are not comparable. ↩

Zartman, R. E. et al.  “National Geochronological Database.”  U.S. Geological Survey Open-File Report 03-236.  2003.  A database report from the U.S. Geological Survey.  Zartman, Bush and Abston wrote it, and Sloan, Henry, Hopkins and Ludington revised it.  The article's own audit ran an independent implementation of the method on this database. ↩

1. 44a.

   Report front matter, verbatim:  “This report is preliminary and has not been reviewed for conformity with U.S. Geological Survey (USGS) editorial standards or with the North American Stratigraphic Code,”.  and, in the documentation for the Decay Constant fields (KarDecCon, RbSrDecCon, UPbDecCon, FTDecCon, SmNdDecCon):  “We did not recalculate radiometric ages using more recently accepted decay constants…  To avoid misinterpretation, check for consistency between records; some recalculation may be needed where direct comparison is intended.” ↩

Computed for this article from the NGDB's per-determination interpretation field, which USGS compilers wrote from each source paper.  Across the whole database 22.1% of usable determinations are flagged as not dating rock formation; among the 34 headline records the figure is 39.6% of determinations, and 30 of the 34 records contain at least one such age.  Typical flag texts:  "minimum age of intrusion", "initial age disturbed by later metamorphism", "age of uplift and cooling", "sample contains excess radiogenic argon, age is too old". ↩

Computed for this article.  Restricting to NGDB samples carrying two or more genuinely independent parent systems, n = 744:  51.2% agree completely once decay constants are harmonised and each system is reduced to one age, 59.3% once compiler-flagged ages are dropped (n falls to 364).  Cross-method relative difference |t₁−t₂|/mean, best-controlled pooled set (n = 374 pairs):  lower quartile 3.8%, median 8.7%, upper quartile 23.0%; 11.0% of pairs inside 1%; 7.0% differ by more than a factor of two.  K-Ar versus Rb-Sr, the commonest comparison (n = 511), median 9.6%. ↩

Computed for this article by parsing the reference field of all 18,576 location records.  Median publication year of the underlying study 1976; 61.3% published in the 1960s or 1970s; 99.7% in or before 1990; latest 1991.  Of 14,187 usable K-Ar ages, 13,472 (95.0%) are conventional K-Ar; 305 (2.2%) are ⁴⁰Ar/³⁹Ar incremental-heating plateau ages and 574 (4.0%) are ⁴⁰Ar/³⁹Ar of any kind.  Of 6,161 Rb-Sr ages, 1,321 are model ages using an assumed initial ⁸⁷Sr/⁸⁶Sr and 71 use a measured one.  Only 38.7% of K-Ar and 40.2% of Rb-Sr ages use the modern Steiger & Jäger (1977) constants, against 83.1% of U-Th-Pb — so a raw K-Ar-versus-U-Pb comparison in this file carries a 1-3% built-in offset before any geology.  Also:  647 of the 2,621 non-blank ²⁰⁷Pb/²⁰⁶Pb ages are exactly zero, which the USGS documentation warns are converted blanks. ↩

Computed for this article.  5,509 K-Ar rows record the ⁴⁰K/K abundance as 1.1167E-04; no published value equals that, and the IUGS value is 1.167E-04.  Recomputing 4,126 of those ages from the database's own K₂O% and ⁴⁰Ar* gives a median residual of −0.009% against the published age using 1.167e-4, and +4.42% using the literal field value.  This recomputation is also the pipeline's strongest self-check, and it independently shows that 40ArRad is stored in 10⁻¹⁰ mol/g rather than the documented mol/g. ↩

Wise, Kurt P.  Faith, Form, and Time: What the Bible Teaches and Science Confirms about Creation and the Age of the Universe.  Broadman & Holman, Nashville.  2002.  The endnote says the database holds thousands of examples of rocks dated with multiple methods, and that a careful examination of these records shows that.  “the methods rarely yield the same ages”.  A book-length young-earth argument for general readers by Kurt P. Wise, published in Nashville in 2002.  The footnote claiming that in the National Geochronological Database the methods rarely yield the same ages is offered without analysis, and the article uses it as the claim that Beachy, Kinard and Garner set out to test; at the whole-database level they found the opposite. ↩

Snelling, Andrew A.  “Isochron Discordances and the Role of Inheritance and Mixing of Radioisotopes in the Mantle and Crust.”  Radioisotopes and the Age of the Earth II:393–524.  Institute for Creation Research / Creation Research Society.  2005.  Steven Austin's companion chapter 5, pages 325–392, covers the Bass Rapids sill and the Beartooth amphibolite.  The sixth chapter of RATE's second volume, by Andrew Snelling of the Institute for Creation Research.  It sets out the project's isochron case in technical detail, and explains that isochron discordances, where rock dates from different minerals or samples disagree, can come from inherited isotopes and from mixing.  The article uses it as the main technical statement of RATE's isochron argument.

1. 50a.

   Chapter 6, section 3–4, page 406:  “A total of one hundred and thirty-six samples were collected”.  Across ten rock units.  The methods run to pages 406–410.  Whole-rock powders went to XRAL Laboratories and Amdel (Adelaide).  K-Ar went to Geochron Laboratories, Massachusetts (Dr Richard Reesman), and Activation Laboratories, Ontario (Dr Yakov Kapusta).  Rb-Sr, Sm-Nd and Pb-Pb went to the University of Colorado at Boulder (Prof. G. Lang Farmer) and the PRISE laboratory at the Australian National University (Dr Richard Armstrong).  Isochrons were fitted with Isoplot/Ex v2.49 and York two-error regression, with MSWD reported for nearly every plot.  Snelling's acknowledgment says that the four named analysts.  “may not agree with some aspects of the interpretation of the data and the conclusions of this study.” ↩

2. 50b.

   Page 435:  Volume II, on the Mt Ngauruhoe isochrons, says the ages.  “were derived by selective manipulation of the data.”  Snelling's own conclusion for that rock is inheritance of mantle Pb and Nd plus about 5% crustal sediment (pp. 435–436). ↩

3. 50c.

   Austin, Volume II, pp. 352–353, on K-Ar ages of 1520, 2011, 2403 and 2620 Ma within a single Beartooth hand sample.  His full sentence is:  “Because it is extremely unlikely that the minerals of the amphibolite within a few centimeters of each other record a billion year cooling history, differential Ar loss can be supposed.”  That is a direct objection to this article's own Beartooth reading, and it deserves an answer:  minerals a few centimetres apart in a slowly exhumed Archean terrane do cool together, and it is the terrane that takes hundreds of millions of years to pass through 700 °C to 200 °C, not the hand sample relative to itself.  Either way the spread is an 1,100 Ma spread inside one decay system, three times the 371 Ma α-β gap attributed to nuclear physics in the same rock.  Page 428:  The Brahma amphibolites are worse: K-Ar model ages from 405.1 ± 10 to 2574.2 ± 73 Ma, with two samples 0.84 m apart in one outcrop giving 1205.3 and 2574.2 Ma. ↩

4. 50d.

   Austin, Volume II, p. 371: Isoplot.  “shows no significant Pb-Pb age information can be derived”.  from the DI-13 and DI-15 mineral suites; and on p. 366.  “No Sm-Nd age information can be derived from the eleven whole-rock samples.”  In the best-sampled rock, ⁸⁷Rb/⁸⁶Sr spans a factor of 66 across mineral separates while ¹⁴⁷Sm/¹⁴⁴Nd spans 1.9 and ²⁰⁶Pb/²⁰⁴Pb spans 1.06. ↩

5. 50e.

   Page 465:  Volume II, in the Future Work section:  “It remains a puzzle as to why the meteorite isochron 'ages' are apparently concordant, when those for the Precambrian rock units tested in this study are discordant.” ↩

Mohr, Michael T. et al.  “High-precision U-Pb geochronology links magmatism in the Southwestern Laurentia large igneous province and Midcontinent Rift.”  Geology 52(3):193–198.  2024.  A 2024 dating study by a team from several universities that measured zircon crystals from Southwestern Laurentia rocks with a high-precision method called chemical abrasion isotope dilution thermal ionization mass spectrometry.  It found an age of 1098.16 ± 0.59 Ma (million years), from the ²⁰⁶Pb/²³⁸U ratio, on a granophyre, a coarse-grained silica-rich rock, in a sill intruding the Bass Formation and Hakatai Shale at Colorado River mile 107, in the Bass Rapids outcrop belt that RATE sampled, and a companion age of 1082.18 ± 1.25 Ma on the Cardenas Basalt.  The article uses it as the modern date for the same granophyre that RATE dated. ↩

Computed for this article (rate-isochron-arithmetic.py), by ordinary least squares on RATE's own published isotope ratios.  The eleven Bass Rapids whole rocks split into three granophyre samples at ²⁰⁶Pb/²⁰⁴Pb 21.9-25.4 and eight diabase at 17.3-19.4, with 31% of the x-axis empty between them; refitting separately gives 1705 Ma for the diabase alone and 1062 Ma for the granophyre alone against 1246 Ma pooled.  The script reproduces RATE's own Isoplot fits to a fraction of a percent wherever the data are well conditioned, which validates the method before it is used to disagree.  Snelling then reports (Volume II, p. 449) that adding two samples of the contact hornfels — baked Hakatai Shale, a sedimentary wall rock not part of the intrusion — improves the statistics and tightens the age to 1280 ± 100 Ma.  Adding a foreign rock should wreck an isochron; it improves a mixing line. ↩

Computed for this article (rate-isochron-arithmetic.py) across all 25 α/β age pairs in RATE Volume II Table 2.  Median implied acceleration ratio 1.31, range 0.55 to 29.4, four pairs with the wrong sign (α younger than β), including both Apache Group units and the Uinkaret Plateau basalts.  Snelling's own single numeric statement of the differential, ICC 2008 p. 422 — "the U-Pb system decayed through 1,883 million years … while the Rb-Sr system decayed through 1,240 million years … over the identical actual time interval" — is a ratio of 1.52.  The script also documents six cases in which RATE printed two different answers for the same rock and the same system, swinging 24-48%, with Volume II Table 2 carrying the number that maximises the α-β gap in each case. ↩

Accelerated decay

Radioisotopes and the Age of the Earth: A Young-Earth Creationist Research Initiative.  Volume 1.  Institute for Creation Research / Creation Research Society.  Edited by Larry Vardiman, Andrew A. Snelling and Eugene F. Chaffin.  2000.  The first volume of a young-earth creationist research initiative on radioisotope dating, edited by Larry Vardiman, Andrew A. Snelling and Eugene F. Chaffin.  The article uses it because its chapter on meteorites says one thing and the same author's later work on terrestrial rocks says the opposite.  ISBN 0-932766-62-5.

1. 54a.

   Austin, on chondrites:  Pages 114–115:  “Can a conclusion be offered for the near concordance between α-emitters (U and Sm) and β-emitters (Rb and K) within chondrites?  The data seem to be consistent with the notion, if radioisotope decay was accelerated at sometime in the past, that α-emitters (U and Sm) and β-emitters (Rb and K) within chondrites were both uniformly accelerated.”  Compare Austin, p. 95:  Page 95:  “Mineral and whole-rock ages for chondritic meteorites are approximately concordant,”.  and Baumgardner, p. 88, on.  Page 88:  “the concordant 4.5 billion year radioisotope ages for meteorites.”  Five years later the same authors argue that terrestrial rocks show α accelerated more than β. ↩

Snelling, Andrew A.  “Radioisotope Dating of Meteorites: V. Isochron Ages of Groups of Meteorites.”  Answers Research Journal 8:449–478.  2015.  The fifth in a series in which Snelling tested RATE's central claim against meteorite data and reported that it failed.  The article cites it because Snelling checked meteorite isochron ages, found no RATE-style pattern, and offers the rescue that accelerated decay happened on Earth only.

1. 55a.

   Section Abstract:  A young-earth author testing his own project's central claim and reporting that it failed.  “No pattern was found in the isochron ages for these groups of meteorites similar to the systematic patterns of isochron ages found in Precambrian rock units during the RATE project, so there is no evidence of past accelerated radioisotope decay having occurred in these meteorites, and therefore on their parent asteroids.  This is not as expected…  Thus it is argued that accelerated radioisotope decay must have only occurred on the earth.” ↩

Computed for this article.  Across 130 NGDB records holding both, α-emitter ages read older than β-emitter ages in 57.6% of 540 comparisons, exact two-sided sign test p = 0.00048, median ratio 1.027 — a real effect, reported here because it favours the young-earth side.  Under controls it falls to p = 0.141.  The signal is carried entirely by ²⁰⁷Pb/²⁰⁶Pb:  ²⁰⁶Pb/²³⁸U, also an α parent, reads younger than Rb-Sr in 65.6% of comparisons (median ratio 0.919).  Median |ratio−1| is 9.33% for α/α pairs against 7.78% for β/β pairs, though the α/α pairs come from the same crystal in the same run.  The ordering ²⁰⁶Pb/²³⁸U < ²⁰⁷Pb/²³⁵U < ²⁰⁷Pb/²⁰⁶Pb holds in 84.3% and 89.8% of records.  RATE's secondary claim that longer half-life gives older dates holds in 7 of 11 testable pairs. ↩

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.  “in close agreement with OSL age estimates … of 61,000 ± 2,000 years.”  A 1999 dating study of the Lake Mungo 3 skeleton in Australia.  It used electron spin resonance (ESR), which dates by the radiation damage built up in the material, and two uranium-series methods, which date by the decay of uranium.  These gave 62,000 ± 6,000 years.  The authors reported four concordant methods, but the uranium-series and ESR ages share one uranium-uptake model on the same bone, and the optically stimulated luminescence (OSL) age is on the sediment.  The article uses it as the Lake Mungo date that was later overturned. ↩

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 with 25 new optical (light-based) ages from the Lake Mungo sediments in Australia.  The authors date both burials to 40,000 ± 2,000 years ago, superseding the concordant 62,000 ± 6,000 years, an estimate wrong by 55%.  The paper also lists three earlier estimates for Mungo III: 30 kyr, 42–45 kyr and 62 ± 6 kyr.  The sentence "For this complex, laboratory-based dating to be successful, the data must be compatible with the external field evidence" is widely quoted from this dispute, but it is not from this paper.  It is attributed to Bowler & Magee, "Redating Australia's oldest human remains:  a sceptic's view," Journal of Human Evolution 38(5):719–726, 2000, and is therefore unverified at the primary level, so the article does not use it.  The article uses this paper as the correction to the 62,000-year date. ↩

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.”  The explanation given by one of the original authors.  The anomalous age is excess ⁴⁰Ar in sub-micrometre inclusions of mantle fluid, identified by its correlation with chlorine.  What caught it was not reproducibility or a second argon method, both of which failed, but that chemical correlation.  The article uses it as the original authors' own explanation. ↩

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.  “One experiment on a different sample of high-Ti basalt from the Qiaojia section that was analysed at the Center for Geodynamics and Geochronology … in Irkutsk, Russia gave a plateau age of 261.7 ± 3.4 Ma.”  It is not the same sample, but including it, the three-laboratory spread is 1.9 Ma on 260 Ma, or 0.73%, still inside the quoted uncertainties.  A 2021 intercalibration in which argon-argon laboratories in China, the USA and Russia dated the same rocks to test whether they agree.  Two laboratories dated splits of one plagioclase separate, sample QJ 13-1: Beijing 259.9 ± 1.2 and Madison 259.8 ± 1.2 Ma (2σ analytical), 0.04% apart.  The article uses it as the modern best case for laboratories agreeing.  A third result, on a different sample, is quoted here: ↩

Scott, E. M. et al.  “A Report on Phase 2 of the Fifth International Radiocarbon Intercomparison (VIRI).”  Radiocarbon 52(2–3):846–858.  2010.  “Sample E = 39,305 ¹⁴C yr BP (standard deviation [1σ] = 121 yr) … Sample H = 9528 yr BP (1σ = 7 yr).”  Table 3:  Table 3 gives the full reported ranges.  Mammoth bone E runs from 21,684 to 42,500 BP across 57 results, and whale bone H from 4,810 to 9,880 BP across 55.  “Up to 12% of the results for any 1 sample were identified as outliers (~10% of the total…).”  The phase 2 report of an international radiocarbon intercomparison, in which more than 40 laboratories dated five bone samples.  The paper gives the consensus age of each sample.  The article uses it for the bone results of the same trial. ↩

Hand, Eric.  “Planetary Science: The time machine.”  Nature 487:422–425.  2012.  “we really struggle to get these ages ourselves.”  Quoted without its antecedent (the paragraph before it), the line is a quote-mine.  The paragraph in full:  “Anderson will have to show not only that his chronometer is fast and light, but also that his dates make sense.  Radiometric dates are some of the trickiest, most delicate and most disputed measurements on Earth.  Anderson wants to transform what has been a laborious process of chemical extraction and analysis into a laser-based system, automate it and shrink it into a robot small and reliable enough to send to another planet.  'We're extremely sceptical of these things working,' says Lars Borg, a chemist at the Lawrence Livermore National Laboratory in Livermore, California, whose three-person lab usually produces just two dates a year.  'We really struggle to get these ages ourselves.'”  A 2012 Nature news feature on a laser-based dating instrument, small enough to send to another planet.  It is the source of a widely quoted line from Lars Borg, a chemist at Lawrence Livermore National Laboratory, about his own lab's dating.  Borg is talking about a miniaturised instrument for the Moon or Mars, and the "struggle" is his three-person laboratory's two dates a year.  The article uses it to show a famous quotation stripped of its context. ↩

Heaton, Timothy H.  “Recent Developments in Young-Earth Creationist Geology.”  Science & Education 18(10):1341–1358.  Oct 2009.  “An examination of these efforts demonstrates the anti-scientific nature of using the Bible as a non-negotiable framework for earth history”.  A 2009 review in Science & Education of recent young-earth creationist work in geology, covering the RATE group's accelerated-decay model, catastrophic plate tectonics, and creationist accounts of the fossil record and the Ice Age.  It appeared in Science & Education, not Reports of the NCSE, and some descriptions call it "far more honest than most objections."  The article cites it as a peer-reviewed survey of the field.  Its own abstract is a harsher verdict: ↩

Jeng, Monwhea.  “A selected history of expectation bias in physics.”  American Journal of Physics 74:578.  2006.  A 2006 survey of expectation bias in physics, the pull that published or expected values exert on later measurements.  It is the standard catalogue of this publication pull, and includes Millikan's oil-drop measurement of the electron charge, the speed of light, Einstein and de Haas's gyromagnetic ratio, and the |η₊₋| bandwagon.  The word "neutron" does not appear in it.  No geochronology-specific equivalent has been found; the closest analogues are the Fish Canyon reference-age drift, Vermeesch's MSWD argument, and the Lake Mungo episode.  The article uses it as the standard physics example of expectation bias, set against the absence of a geochronology case. ↩

Reimer, Paula J. et al.  “The IntCal20 Northern Hemisphere Radiocarbon Age Calibration Curve (0-55 cal kBP).”  Radiocarbon 62(4):725–757.  2020.  The Northern Hemisphere calibration curve that converts radiocarbon ages into calendar years back to 55,000 years ago, published by a team of 41 authors.  The article relies on it for its claim that Lake Suigetsu is the only truly atmospheric record that old.

1. 65a.

   Page 737:  “only the Lake Suigetsu varved sediment macrofossil data (Bronk Ramsey et al. 2012) is included in the IntCal20 curve.  It is useful because it provides the only quasi-continuous, truly atmospheric record older than ca. 14,190 cal BP, extending to the limit of the technique.”  IntCal20 also downweights the Bahamas stalagmite record and removes all coral data older than 25 cal kBP.  Both are concessions that critics of this article are entitled to point to. ↩

Computed for this article (suigetsu-varve-vs-uth.py) from the supplementary data table of Bronk Ramsey et al., "Reanalysis of the Atmospheric Radiocarbon Calibration Record from Lake Suigetsu, Japan," Radiocarbon 62(4):989-999, 2020, doi:10.1017/RDC.2020.18, over 593 dated levels.  Residual of counted laminae against the U-Th-anchored calendar scale:  −32 yr at 10-15 ka, +1,409 at 20-25 ka, +2,655 at 30-35 ka, and +2,744 at the deepest level (47,762 counted against 50,506 cal BP), a 5.4% shortfall; worst bin 6.9-7.5%.  Only 16% of levels sit inside the 2σ U-Th band.  The count runs short, which is the direction that missing laminae predict — and about half the years in the record have no preserved layer.  The 2012 Science paper itself was not obtained (Cloudflare at every repository). ↩

Computed for this article from the same table.  Faster ¹⁴C decay in the past would leave less ¹⁴C and make radiocarbon ages come out older than true ages.  In this dataset radiocarbon ages run younger than calendar ages throughout the glacial period — the ratio of ¹⁴C age to calendar age is 0.83 to 0.95 across 11-55 ka.  Accelerated decay predicts the wrong sign.  Also correcting a figure circulating in this project's own notes:  atmospheric Δ¹⁴C peaks at about +500 to +535 per mil in this binning, which is 1.5× the modern ¹⁴C/¹²C ratio, not 8×;  Δ¹⁴C is a per-mil quantity, so +700‰ means 1.70×. ↩

Svensson, A. et al.  “A 60 000 year Greenland stratigraphic ice core chronology.”  Climate of the Past 4:47–57.  2008.  Table 2, page 50:  The Laschamp event at 41,250 ± 814 counted years against 40,400 ± 1,000 ⁴⁰Ar/³⁹Ar years.  North Atlantic Ash Zone II / Z2 tephra at 55,380 ± 1,184 counted years against 54,500 ± 1,000 argon.  “The MCE for the 41.8-60 ka section is 950 years or 5%.”  Page 54:  “GICC05 agrees with the Hulu Cave time scale within 800 yrs throughout the 60 ka period,”.  “The maximum counting error of the time scale is on average 1% in the Holocene and 5% in the Glacial.”  A 2008 paper, with 14 authors, that builds a 60,000-year timescale for Greenland ice cores (GICC05) by counting annual layers.  It is checked against independently dated markers: the Laschamp geomagnetic excursion, a volcanic ash layer, and the Hulu Cave speleothem timescale.  The article uses it as a counted chronology whose layer counts are tested against argon-argon dates and the Hulu Cave timescale. ↩

Sigl, M. et al.  “Timing and climate forcing of volcanic eruptions for the past 2,500 years.”  Nature 523:543–549.  2015.  One named cause was a tephra layer wrongly attributed to the AD 79 eruption of Vesuvius and used as a fixed tie point.  The attribution is retracted here, and the layer is 87/88 CE.  The other, in the authors' own words, was.  “use of manual-layer interpretation techniques that may favor interpretations consistent with a priori knowledge or existing chronologies”.  The paper tags this cause "(WDC)", the WAIS Divide core in Antarctica.  The revised scales are accurate to.  “within less than ±5 years during the past 2,500 years.”  Seven years on 1,240 is 0.56%, admitted in Nature in the language of confirmation bias.  A 2015 study that dates volcanic eruptions over 2,500 years by matching Greenland and Antarctic ice cores against one another and against dated markers.  The Greenland and Antarctic chronologies were seven years wrong at 775 CE, detected by a beryllium-10 (¹⁰Be) spike matched to tree-ring radiocarbon.  The article uses it as its main example of a dating error that an independent clock caught. ↩

Kent, Dennis V. et al.  “Empirical evidence for stability of the 405-kiloyear Jupiter-Venus eccentricity cycle over hundreds of millions of years.”  PNAS 115(24):6153–6158.  2018.  “within an insignificant 1 ky or 0.25%”.  of the hypothesised 405-kyr period.  Orbital mechanics and nuclear decay share no physics.  A 2018 test of the 405-kyr orbital cycle, driven by the gravity of Jupiter and Venus, using a drill core from Petrified Forest National Park in Arizona.  Counting slightly fewer than 21 eccentricity cycles across the Late Triassic gives 8.50 Myr between two horizons.  Chemically abraded uranium-lead zircon dates on the same two horizons give 8.48 ± 0.3 Myr.  The article uses it to show that the most stable orbital clock agrees with radiometric dates over 8.5 million years.  The authors say the two results agree. ↩

Kent, Dennis V. et al.  “Magnetostratigraphic confirmation of a much faster tempo for sea-level change for the Middle Triassic Latemar platform carbonates.”  Earth and Planetary Science Letters 228:369–377.  2004.  A 2004 magnetostratigraphy study of the Latemar, a Middle Triassic carbonate platform in the Dolomites of Italy with about 600 metre-scale cycles.  Its polarity record, with uranium-lead zircon dates and fossil zones, puts the platform at about 1 million years.  Counting the cycles as precession had given 9–12 million years, an order of magnitude more.  The beds are now read as millennial rhythms of roughly 1.7 kyr.  Cycle counting is falsifiable, it was falsified, and the radiometric clock is what falsified it.  The article uses it as an example of cycle counting that a radiometric clock overturned. ↩

Wells, John W.  “Coral Growth and Geochronometry.”  Nature 197(4871):948–950.  1963.  “usually about 400, ranging between extremes of 385 and 410”.  growth lines per year, in Middle Devonian corals about 380 million years old.  A 1963 paper by a Cornell University geologist that counts growth lines in fossil corals to estimate how many days there were in a year long ago.  The article uses it as the classic coral argument for a longer past day, and says it is weaker than it looks. ↩

Williams, George E.  “Geological constraints on the Precambrian history of Earth's rotation and the Moon's orbit.”  Reviews of Geophysics 38(1):37–59.  2000.  A 2000 review of the rock record of tidal rhythmites (layered sediments laid down by ocean tides) before 540 million years ago, used to reconstruct how Earth's rotation and the Moon's orbit have changed.  The article uses it for two things: the caveat that fossil clocks are unreliable, and the strong case from tidal rhythmites.

1. 73a.

   “The great hopes held for paleontological 'clocks' have not, however, been fully realized,”.  The paper calls the Palaeozoic coral values.  “of dubious reliability.”  Do not lean on Wells 1963. ↩

2. 73b.

   This is the strong case.  The Elatina and Reynella rhythmites of South Australia, about 620 million years old, give 400 ± 7 solar days per year, 13.1 ± 0.1 synodic months per year, and a day of 21.9 ± 0.4 hours.  These three quantities must satisfy one another by celestial mechanics, and they do.  The implied mean lunar recession rate since then is 2.17 ± 0.31 cm/yr,  “little more than half the present rate of lunar recession of 3.82 ± 0.07 cm/yr obtained by lunar laser ranging”.  which is what defeats the "extrapolate today's recession backwards and the Moon touches the Earth at 1.5 Ga" argument. ↩

Friedrich, Michael et al.  “The 12,460-Year Hohenheim Oak and Pine Tree-Ring Chronology from Central Europe.”  Radiocarbon 46(3):1111–1122.  2004.  When the floating Preboreal Pine Chronology was finally cross-matched dendrochronologically, the match.  “revealed a difference of only 8 yr”.  from the radiocarbon wiggle-match position: two clocks 8 years apart over 11,000 years, or 0.07%.  A 2004 paper on a German tree-ring chronology, the backbone of central European radiocarbon calibration, covering 12,460 years.  It is built from more than 7,000 individual oaks, with a mean replication of 108 trees per year, and 96% of its length is covered by more than 20 crossdated trees.  It was independently confirmed against the Irish and Göttingen chronologies back to 9741 BP.  The article uses it for the agreement of the floating pine chronology with radiocarbon. ↩

Compiled for this article from Reiners, Peter W., Terry L. Spell, Stefan Nicolescu & Kathleen A. Zanetti, "Zircon (U-Th)/He thermochronometry:  He diffusion and comparisons with ⁴⁰Ar/³⁹Ar dating," Geochimica et Cosmochimica Acta 68(8):1857-1887, 2004, doi:10.1016/j.gca.2003.10.021, Table 3 and §3.4-3.5.  The 60 Ma zircon U-Pb age is Reiners' report of Gehrels et al. 1991, 60.4 +1.3/−3.0 Ma "for the Speel River pluton" rather than for sample 8500-15 itself.  Southwest Arm pluton, Stewart Island, New Zealand:  titanite (U-Th)/He (T_{c c}0 °C) 92.7 ± 7.4 and 85.4 ± 6.8 Ma; zircon (U-Th)/He (T_c 168-176 °C) 83.1 ± 8.3 Ma; apatite (U-Th)/He (T_c ~70 °C) 54.5-60.4 Ma.  Coast Plutonic Complex sample 8500-15:  zircon U-Pb crystallisation 60 Ma; K-feldspar ⁴⁰Ar/³⁹Ar model cooling 350→160 °C between 54 and 50 Ma; zircon He 49.3 ± 3.9 Ma; apatite He 7.49 Ma.  Perfect closure-temperature order in both.  The point that matters for this article:  the sequence alternates between α-based clocks (U-Pb, (U-Th)/He) and an electron-capture clock (⁴⁰Ar/³⁹Ar), so uniform acceleration would preserve the ordering while RATE's differential acceleration would invert it.  See also Willigers, Krogstad & Wijbrans, "Comparison of Thermochronometers in a Slowly Cooled Granulite Terrain," Journal of Petrology 42(9):1729-1749, 2001. ↩

Also consulted

Uzan, Jean-Philippe.  “Varying Constants, Gravitation and Cosmology.”  Living Reviews in Relativity 14(2).  2011.  Table 9:  The α-decay sensitivities s = ∂lnλ/∂lnα are +548 for ²³⁸U, +466 for ²³⁵U and +774 for ¹⁴⁷Sm.  The review is also the source of the published Re-Os-versus-U-Pb bound, Δα/α = (−8 ± 16) × 10⁻⁷.  A review article by Jean-Philippe Uzan, a physicist at the Institut d'Astrophysique de Paris, that surveys the measured limits on whether fundamental constants have changed.  It covers atomic clocks, the Oklo reactor, meteorite dating and quasar spectra, and it was also posted as arXiv:1009.5514.  The article uses its table of sensitivities for the alpha-decay values. ↩

Ohtsuki, T. et al.  “Enhanced Electron-Capture Decay Rate of ⁷Be Encapsulated in C₆₀ Cages.”  Physical Review Letters 93:112501.  2004.  The largest measured effect on any decay rate short of full ionisation is 1.5%, reported in the 2007 companion (PRL 98:252501): ⁷Be in C₆₀ at 5 K, half-life 52.47 ± 0.04 d, against ⁷Be in beryllium metal at 293 K.  Widening the comparison to the most slowly decaying host in the literature, gold, 53.311 ± 0.042 d, tabulated in Ohtsuki et al. 2004, Table I, from Norman et al., gives 1.58%, but that is a difference computed across two experiments, not a measured one.  A 2004 laboratory experiment by a group at Tohoku University in Japan.  It measured how long beryllium-7 takes to decay when it sits inside a C60 cage and when it sits in beryllium metal, and found a clear difference between the two.  The article cites it for the largest measured environmental effect on a decay rate, but the 1.5% figure comes from the 2007 companion paper below. ↩

Stone, N. J. et al.  “On the absence of appreciable half-life changes in alpha emitters cooled in metals to 1 Kelvin and below.”  Physical Review C 76:054315.  2007.  A 2007 review by N. J. Stone and colleagues at Oxford and two US universities of published experiments that cooled alpha emitters implanted in iron foil to as low as 25 mK.  It found no change in half-life on cooling larger than about 1%, against predicted changes of several orders of magnitude, and the paper is also on arXiv as nucl-ex/0611041.  The article uses it for that 1% ceiling on alpha decay, since the review rests on the few alpha emitters it tested and on forty years of low-temperature nuclear-orientation data. ↩

Pommé, S. et al.  “Evidence against solar influence on nuclear decay constants.”  Physics Letters B 761:281–286.  2016.  See also Kossert & Nähle, and Norman et al. 2009.  A 2016 study by a team of 44 authors from 14 metrology laboratories, which measured radioactive sources repeatedly for periods from 200 days to four decades.  It set an upper limit of 0.0006% to 0.008% on any annual oscillation in decay rates, found none in step with Earth's distance from the Sun at the 10⁻⁶ to 10⁻⁵ level, and attributes the deviations it did see to instrument instability.  The article treats it as the definitive refutation of the claimed yearly cycle in decay rates that tracks the Earth's distance from the Sun. ↩

Meshik, A. P. et al.  “Record of Cycling Operation of the Natural Nuclear Reactor in the Oklo/Okelobondo Area in Gabon.”  Physical Review Letters 93:182302.  2004.  The α-variation bound from Oklo is Damour & Dyson 1996: −0.9 × 10⁻⁷ < Δα/α < 1.2 × 10⁻⁷.  A 2004 study from Washington University in St Louis that measured the fission gases in a uranium-free aluminium phosphate mineral from the Oklo reactor in Gabon, which is about 2 billion years old.  The xenon in the mineral holds the highest fission xenon concentration ever found in natural material, and its isotopes show a reactor that ran in pulses of about 30 minutes, each followed by a dormant period of 2.5 hours.  The article uses it for that duty cycle. ↩

Simakin, A. V., and G. A. Shafeev.  “Accelerated alpha-decay of ²³²U isotope achieved by exposure of its aqueous solution with gold nanoparticles to laser radiation.”  arXiv.  2011.  A 2011 preprint (arXiv:1112.6276), published in 2013 as "Accelerated alpha-decay of uranium isotopes induced by exposure of aqueous solution of uranium salt with gold nanoparticles to laser radiation" in Physics of Wave Phenomena 21:31–37.  It reports that laser light on a water solution of uranium salt with gold nanoparticles speeds up the alpha decay of uranium.  By the claim's own stated pulse-energy budget it is internally inconsistent by a factor of about 10⁶, and the electric field involved is roughly 3.6 × 10⁻⁹ of the field at a uranium nucleus's surface, which is worth Δλ/λ ≈ 5 × 10⁻⁶, not the claimed 10⁸.  The article uses it as the laser example that fails its own arithmetic. ↩

Adamenko, S. V. et al.  “Exploring new frontiers in the pulsed power laboratory: Recent progress.”  Results in Physics 5:62–68.  2015.  The paper's own concession, that the transmutation process is likely to need an outside energy input of at least 100 J, contradicts the laboratory's earlier over-unity framing:  “is likely to demand an external energy input”.  Reifenschweiler's tritium result (Physics Letters A 184:149-153, 1994) is a 40% slow-down from unrepeatable 1958-68 experiments, i.e. the wrong sign for the young-earth argument.  A 2015 report from the Proton-21 Electrodynamics Laboratory in Kiev, which fired a high-power electron beam pulse at a small metal target and found anomalously high amounts of new chemical elements in what was left of the target.  No independent group has repeated the results.  The article uses it because the paper itself concedes that the process needs outside energy, which removes the heat sink that Walt Brown's hydroplate model invokes it for. ↩

Vardiman, Larry et al.  “Summary of Evidence for a Young Earth from the RATE Project.”  Radioisotopes and the Age of the Earth II:735–772.  Institute for Creation Research / Creation Research Society.  2005.  A summary chapter, chapter 10 of RATE's second volume, from the RATE project, a team of creationist scientists who began studying radioisotopes in 1997, five of them at the Institute for Creation Research.  It reports that nuclear decay appears to have sped up for brief periods in Earth's past, and it lists the objections critics would raise.  The article uses it for the authors' own admissions about heat, an unproven mechanism and radiation.

1. 83a.

   Page 761:  “The heat would have melted the crustal rocks many times over unless there was some mechanism for simultaneously removing it quickly.”  Page 763:  On Humphreys' volumetric-cooling proposal, the authors write that it.  “represents only a beginning attempt to understanding the complicated events of the Genesis Flood and may not, in fact, be the correct mechanism.”  Page 765:  “the radiation dose from similar levels of ⁴⁰K within Noah's body likely would have been lethal.”  Page 743:  The radiohalo annealing figure is on this page. ↩

Computed for this article from Sprain et al. 2018's own published used/run column (GSA Bulletin 130:1615-1644, Table 2), which reports per sample how many single-crystal fusions were retained out of how many were run:  1,131 of 1,315 retained, i.e. 86.0% retention and 14.0% rejection, with one sample (PH13-3, Purgatory Hill, Y coal) retaining only 13 of 69 (19%).  The rejection rules and the raw analyses are both published, which is the correct answer to a file-drawer charge at the level of individual analyses.  At the level of which samples and which papers reach print, nobody has quantified it and we found no one who has tried. ↩

Clocks that do not use decay

U.S. Geological Survey.  USGS Geochron: A Database of Geochronological and Thermochronological Dates and Data (ver. 4.0).  Jun 2025.  The successor to the NGDB, the USGS's earlier database of geochronological dates, released as version 4.0 in June 2025.  It holds 76,426 usable ages, and 97.1% carry an uncertainty, against 71.0% in the NGDB.  It also has structured interpretation fields, and covers Ar-Ar, (U-Th)/He, Re-Os and Lu-Hf dates, none of which the NGDB has.  Joined on field sample name, it yields 968 samples with two or more methods and 43 with three or more.  The article recommends a modern replication of the Beachy, Kinard & Garner analysis on this database, and calls it feasible. ↩

Schoene, Blair.  “U–Th–Pb Geochronology.”  Treatise on Geochemistry 4:341–378.  Elsevier.  2nd edition.  2014.  Chapter 4.10 of the Treatise on Geochemistry, the standard reference chapter on the system, written by a Princeton geochronologist.  It covers how uranium and thorium decay to lead, the causes of discordance, and the measurement techniques.  The article quotes its figure of about 0.3% systematic discordance in its note on Condon et al.

Dalrymple, G. Brent, and Marvin A. Lanphere.  Potassium-Argon Dating: Principles, Techniques and Applications to Geochronology.  W. H. Freeman, San Francisco.  1969.  A 1969 textbook on the potassium-argon method, covering its principles, its techniques and its uses in geochronology.  The article cites it for the table of historic lava flows in which excess argon was found.

1. 87a.

   The 1969 paper by G. Brent Dalrymple, "⁴⁰Ar/³⁶Ar analyses of historic lava flows," in Earth and Planetary Science Letters 6(1):47-55, DOI 10.1016/0012-821X(69)90160-5.  The book's Table 8-5 reproduces its data on historic lava flows.

2. 87b.

   Page 133, tables 8-5:  “Historic lava flows in which excess Ar⁴⁰ was found”.  The table is headed "Source: Data from Dalrymple (1969)." Its columns are lava flow, year of eruption, excess Ar⁴⁰ in 10⁻¹² mole/gram, and apparent age in 10⁶ years.  Hualalei, Hawaii, 1801: 1.5 and 1.1.  Sunset Crater, Arizona, 1065: 0.26 and 0.22.  Mt. Etna, Sicily, 1792: 0.35 and 0.15.

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 1986 Geological Survey report by G. Brent Dalrymple, 76 pages long, that replies to "scientific" creationism, the literal reading of Genesis that puts the Earth at about 6,000 to 10,000 years old.  It tests the creationists' geology against the clocks it covers, including the potassium-argon, rubidium-strontium, uranium-lead and radiocarbon methods.  The article cites it as the Geological Survey reply to scientific creationism.

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 2003 web essay by Kevin R. Henke, Ph.D.  It argues that Austin, Snelling and Swenson miscopied the numbers for historic lava flows from Dalrymple's 1969 study, which dated 26 of them by potassium-argon.  The article uses it as the reply to the creationist use of those flows.

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.  On page 5 of the ICR reprint, Austin discloses that no information about the rock's age was given to the laboratory, which is why the "labs tune the answer" charge cannot be run in its strong form on this experiment.  A 1996 paper on excess argon, which is argon that a rock traps when it forms rather than argon from its own radioactive decay, in mineral grains from the new dacite lava dome at Mount St Helens.  The article uses it because it is the young-earth movement's own showcase experiment on this kind of dating.

Snelling, Andrew A.  “'Excess Argon': The 'Archilles' Heel' of Potassium-Argon and Argon-Argon 'Dating' of Volcanic Rocks.”  Impact.  Institute for Creation Research.  Jan 1999.  A January 1999 article by Andrew Snelling in Impact, the Institute for Creation Research's publication, on excess argon in potassium-argon and argon-argon dating of volcanic rocks.  It reproduces Austin's seven numbers verbatim under the heading "The historically dated flows and their 'ages'," and misdates the 1792 Etna flow to AD 1972.  An undated addendum on the page acknowledges that the numbers are argon amounts, not ages.

Henke, Kevin R.  “Young-Earth Creationist 'Dating' of a Mt. St. Helens Dacite: The Failure of Austin and Swenson to Recognize Obviously Ancient Minerals.”  No Answers in Genesis.  2000.  A 2000 web essay by Kevin R. Henke, Ph.D., on the No Answers in Genesis site, about the potassium-argon dates that Austin and Swenson obtained for a dacite from Mt. St. Helens, which probably erupted in 1986.  It argues that the laboratory used could not date rocks younger than about 2 million years, that four of the five results fall below that limit, and that Austin ran the test on the wrong kind of sample.  The article uses it as the reply to Austin's Mt. St. Helens dating.

Zashu, S. et al.  “K-Ar isochron dating of Zaire cubic diamonds.”  Nature 323:710–712.  1986.  A 1986 Nature paper that fits a K-Ar isochron, a line through the data of several samples, to ten diamonds from Zaire, and gives 6.0 Gyr (6.0 billion years), an age older than the Earth.  The article cites it for the 6.0 billion year isochron on the Zaire diamonds.

Podosek, F. A. et al.  “Normal potassium, inherited argon in Zaire cubic diamonds.”  Nature 334:607–609.  1988.  A 1988 Nature paper that records argon-40/argon-39 analysis of four more diamonds, which reproduced the same 6.0 Gyr (6.0 billion years): two argon methods, concordantly wrong.  The article cites it for the four further diamonds that gave the same 6.0 billion years by a second argon method.

Akagi, T., and A. Masuda.  “⁴⁰K-³⁹K ratio of cubic diamond from Zaire.”  Geochemical Journal 22:81–82.  1988.  A short paper on a proposed rescue of the diamond ages: that the potassium in these cubic diamonds was isotopically unusual, enriched in ⁴⁰K.  Two laboratories tested it and found the ⁴⁰K/³⁹K ratio normal to about 1%, so the rescue failed.  The other test, Podosek et al. 1988, carries three of Zashu's co-authors, so only this paper is fully independent of the original team.  The article uses it to show that the proposed rescue failed.

Lewin, Roger.  Bones of Contention: Controversies in the Search for Human Origins.  Simon & Schuster, New York.  1987.  A book-length account of the controversies over human origins, written by a journalist who interviewed the participants and quotes their letters.  It is the fullest published account of the KBS Tuff dispute, and the article draws two of its results from it that appear nowhere else.

Curtis, G. H. et al.  “Age of KBS Tuff in Koobi Fora Formation, East Lake Turkana, Kenya.”  Geological Background to Fossil Man, 463–470.  Scottish Academic Press, Edinburgh.  Edited by W. W. Bishop.  1978.  Page 465:  “The first dates … showed no agreement (see Table 3).  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.”  Chapter 28 of a 1978 book on the geology behind early human fossils, dealing with the age of the KBS Tuff, a volcanic ash layer at Koobi Fora in Kenya.  The authors' first dates disagreed with one another.  Their microscope check then found rounded, altered feldspar grains, which they judged detrital, meaning older rock fragments, probably derived from ancient bedrock.  The shorter Nature report of the same work, Nature 258(5534):395–398, 1975, has an abstract giving 1.60 ± 0.05 Myr for areas 10 and 105 and 1.82 ± 0.04 Myr for area 131.  The article uses it for the first dates that disagreed and for the detrital grains.

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.  “the sedimentary sequence between the KBS Tuff (1.869 ± 0.021 Ma) and the Chari Tuff (1.383 ± 0.028 Ma)”.  A 2006 argon-argon dating study of the upper Koobi Fora Formation in northern Kenya.  Its abstract gives ages for the KBS Tuff and the Chari Tuff, which bound the sediments between them.  There is no Hurford 1987 paper on this tuff.  Hurford, Gleadow & Naeser 1976 gave the 2.44 ± 0.08 Myr fission-track result discussed above, and a 1987 test on the original crystals was never published.  The article uses it for the ages of the tuffs that bound the sediment sequence.

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 fission-track dating of the KBS Tuff, a volcanic ash layer in northern Kenya that lies with associated hominid remains.  Fission-track dating counts the damage tracks that uranium atoms leave in mineral crystals as they split.  The article cites it for the fission-track age of the KBS Tuff.

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 northern Australia.  Earlier thermoluminescence ages of 116,000–176,000 years for the artefact-bearing deposit were reduced to under 10,000 years by single-grain optical dating, which measures light-sensitive sand grains one at a time.  That result agrees with the radiocarbon dates, which had disagreed with the thermoluminescence ages all along.  The disagreement between the methods was the alarm, and it worked.  The article uses it as its Jinmium case, where the older thermoluminescence dates were overturned.

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.”  Some beryls are more than 99.9% excess argon.  Excess argon is not a creationist discovery; it is why nobody dates beryl.  A 1958 laboratory study that measured helium and argon in beryl and other minerals by isotope dilution, with their potassium content and alpha activity.  It found that cordierite and tourmaline also carry excess gas, not only beryl.  The article uses it to explain why beryl is never dated by argon.

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.  The paper's own summary of the dispute:  “previously reported astronomically tuned ages range from 28.126 ± 0.019 to 28.21 ± 0.18 Ma (2σ), a range of ~0.3%.”  Against Renne et al. 2011's 28.294 ± 0.036, that is a 0.42% systematic floor under every argon age.  A 2022 recalibration of the Fish Canyon Tuff sanidine, the most widely used reference mineral for argon-argon dating.  The authors calibrate its age against astronomically tuned tephra in Crete and Morocco.  Their recommended value is 28.176 ± 0.010 (±0.023) Ma.  The article uses it as the current recalibrated reference age for the argon-argon standard.

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.  It puts that limit for the highest-volume dating technique in the world at.  “ca. 2% 2s”.  (two standard deviations), and cites Košler et al. 2013 for documented systematic offsets between laboratories.  A 2016 paper by the community of laser-ablation dating laboratories that sets new reference values for the zircon, monazite and titanite standards used to calibrate the method, and new rules for propagating uncertainty.  The article uses it for the current inter-laboratory uncertainty limit in laser-ablation dating.

Scott, E. M. et al.  “The Fifth International Radiocarbon Intercomparison (VIRI): An Assessment of Laboratory Performance in Stage 3.”  Radiocarbon 52(2–3):859–865.  2010.  Table 1, page 861:  Between-laboratory 1σ runs 0.16–2.46 pMC (percent modern carbon) for accelerator laboratories and 0.87–10.63 pMC for counting laboratories.  That is roughly ±20–180 and ±75–780 radiocarbon years, a conversion the paper itself never makes.  This leaves out charcoal sample P, whose four accelerator results scatter by 4.70 pMC.  Excluding also the modern barley-mash sample S, the worst of the rest, the ranges are 0.16–0.77 and 0.87–2.74 pMC.  The bone stage is worse still, and is reported in a separate paper on phase 2 of the trial.  VIRI is a proficiency trial.  In Stage 3, more than 50 laboratories measured identical material while the true age was withheld.  That is the blind test the "labs tune the answer" charge calls for.  The article uses it for how far radiocarbon laboratories disagree when measuring identical material.