Orbital Cycles in Rock

"this immediately refutes suggestions that significant time is missing in the Newark–Hartford sequence used to build the APTS." — Dennis Kent, Paul Olsen and Giovanni Muttoni, PNAS, 2018¹

"The application of two sophisticated modern techniques to measure geologic time—astrochronology and U–Pb single-crystal zircon dating—thus leads to age estimates of Latemar carbonate platform deposition that differ by about one order of magnitude." — Dennis Kent, Giovanni Muttoni and Peter Brack, 2004²

"the dating of the ice and seafloor sediment cores is a gigantic exercise in circular reasoning." — Jake Hebert, Institute for Creation Research, 2014³

Summary

The Earth's orbit is not fixed.  Jupiter and Venus and the other planets pull on it, so its eccentricity, the tilt of its axis and the direction its axis points all drift on periods of tens to hundreds of thousands of years.  Those drifts change how sunlight is distributed over the seasons and the latitudes.  That changes climate.  And climate changes what a lake or a seafloor deposits.

So a rhythmically bedded sedimentary section can, in principle, be read as a stack of orbital cycles, and counting the cycles gives elapsed time with no nuclear physics in it at all.  This is cyclostratigraphy, or astrochronology, and it is the only non-radiometric clock that reaches deep time.

It is also the counting clock with the sharpest documented failure and the most serious circularity problem, and both should be on the table before the good result is presented.

Ten conclusions, stated first.

1. The metronome is the 405,000-year eccentricity cycle, driven by the g₂–g₅ frequency — the interaction of Venus's and Jupiter's orbits. It is the most stable term in the solar system's orbital solution, and it is the one term astronomers are confident survives back hundreds of millions of years.
2. The best agreement between a counting clock and a decay clock ever measured is 0.25%, and it is here. Between two horizons in the Late Triassic Newark–Hartford basins, counting slightly less than 21 of the 405-kyr cycles gives 8.50 million years. Uranium-lead on zircons from the same two horizons gives 8.48 ± 0.3 million years. The implied cycle length is 404 kyr against the predicted 405.¹
3. That test was designed to answer the obvious objection. If cycles were missing in a hidden unconformity, the count would come out short. So the team drilled a different basin 3,000 km away in Arizona, obtained its magnetostratigraphy, correlated it to the Newark scale, and dated its ash beds independently. The uranium-lead date landed inside the range the cycle count had predicted in advance.¹ Their conclusion: "this immediately refutes suggestions that significant time is missing."
4. The worst published failure of cycle counting is the Latemar, and it is an order of magnitude. About 600 metre-scale shallowing-upward cycles in a 470-metre Middle Triassic atoll were read as orbital precession and gave 9 to 12 million years. Uranium-lead zircon, ammonoid biostratigraphy and magnetostratigraphy together give about one million.² The beds are now interpreted as millennial, roughly 1.7 thousand years each.²
5. The circularity charge against orbital tuning is largely correct for the Neogene and Quaternary, and it should be conceded. Jake Hebert documents that the potassium-argon age of the Matuyama–Brunhes magnetic reversal was revised from 730,000 to 780,000 years because the astronomical theory required it; that the Fish Canyon sanidine standard, on which most argon-argon dating rests, was itself calibrated against astrochronology; and that ice-core flow models are tuned to marine chronologies which are themselves tuned.³ Every step of that is documented in the mainstream literature. Anyone presenting Neogene argon-argon and orbital tuning as independent confirmations of each other is misdescribing how they were built.
6. The fullest young-earth treatment is a four-part series, and its sharpest point is a quotation from Laskar. Michael Oard and John Reed's Cyclostratigraphy series ends by asking whether the pre-Pleistocene record is orbitally forced at all, and quotes the authors of the standard orbital solution: "it will never be possible to recover the precise evolution of the Earth's eccentricity beyond 60 Myr."⁴ If the orbital solution cannot be computed back to 210 Ma, how can a cycle counted there be identified?
7. That objection is answered by the design of the Newark test, and the answer is in its title. Kent's paper is called "Empirical evidence for stability of the 405-kiloyear Jupiter–Venus eccentricity cycle." It does not assume the orbital solution at 210 Ma; it counts cycles between two independently dated horizons and asks what period they imply. The answer comes back 404 kyr. That is a measurement of the thing Laskar says cannot be computed. Oard and Reed's paper appeared in 2020 and does not cite it.
8. The charge does not reach the Triassic test, for four reasons. That test uses uranium-lead, not argon-argon, so it does not pass through the Fish Canyon standard. It is in the Triassic, outside the astronomically tuned interval. It uses a counted number of cycles rather than a fitted match to an orbital solution. And the 405-kyr period comes out of the data at 404 kyr rather than being imposed.
9. There is also an untuned chronology of the Pleistocene, and creationist sources do not engage it. Peter Huybers built marine age models using depth as the proxy for time, with sediment-accumulation uncertainty modelled explicitly and "not relying upon orbital assumptions." The result agrees with the orbitally tuned chronologies to within its stated uncertainty of about ±9,000 years, and it independently recovers obliquity pacing: 33 of 36 major deglaciations over the last two million years occur when the Earth's obliquity is anomalously large.⁵
10. A real, unresolved tension exists and should be printed. Five North Atlantic records place the Matuyama–Brunhes reversal at 773.1 ± 0.4 ka, about 7 kyr younger than the accepted astrochronological age of 780–781 ka; and the two recently proposed revisions of the Fish Canyon standard would make the argon-argon ages 8 to 24 kyr older than the astrochronological ones.⁶ The methods do not all agree to the last digit, and the disagreements are argued about in print.

The thing that is actually surprising

The most striking item in this subject is not a success.  It is the Latemar:  six hundred visually obvious, beautifully rhythmic beds, confidently read as twelve million years of orbital precession by careful people, and actually about one million.  A methodological claim that "it looks periodic, therefore it is orbital" was tested and failed by a factor of ten.  Cycle counting is falsifiable, it has been falsified, and the thing that falsified it was radiometric dating.

Part 1: What the clock is

The three motions

• Eccentricity — how elliptical the Earth's orbit is. Driven by the gravitational tug of the other planets, with dominant periods near 405,000 years and around 100,000 years. The 405-kyr term comes from the difference between the precession frequencies of Venus's and Jupiter's orbits, and it is the most stable of them.
• Obliquity — the tilt of the Earth's axis, oscillating between about 22.1° and 24.5° with a period near 41,000 years.
• Precession — the axis wobbling like a spinning top (about 26,000 years) combined with the slow rotation of the orbital ellipse itself, producing an effective climatic precession cycle of roughly 19,000 to 23,000 years.

Hebert's own description of these motions is accurate and can be adopted:  "it would take about 41,000 years for the tilt of the earth's axis to go from 22.1° to 24.5° and back again…  This variation can be described by cycles of varying periods, the most important of which have periods of approximately 100,000 and 405,000 years."³

Why 405,000 years is the one to use

Chaos in the inner solar system means the individual planetary orbits cannot be integrated backwards with confidence beyond about 60 million years.  But the frequency difference that produces the 405-kyr eccentricity term is unusually insensitive to that chaos.  The question of whether it has actually been stable is empirical, and Kent and colleagues set out to test it in rock.

Two different things called "orbital dating"

These are constantly conflated and they are not the same:

Counting.  Identify cycles in a section, count them, multiply by the period.  This is a genuine counting clock.  It has one strong assumption — that the cycles are what they look like — and the Latemar shows what happens when that assumption fails.

Tuning.  Take a proxy record (say δ¹⁸O down a marine core), and stretch and squeeze its depth scale until its wiggles line up with a computed insolation curve.  This is not a counting clock;  it is a fit.  It produces very precise ages, and it cannot then be used as independent confirmation of the orbital theory it assumed.

Almost every criticism in the young-earth literature is aimed at tuning, and almost every criticism aimed at tuning is right.  This article's positive argument uses counting.

Part 2: The Newark–Hartford test

The setting

The Newark and Hartford basins of eastern North America are rift lakes that filled during the Late Triassic and earliest Jurassic.  Their sediments record lake-level cycles driven by monsoon intensity, which tracks precession modulated by eccentricity.  Counting the 405-kyr modulation through the section builds an astrochronostratigraphic polarity time scale — the APTS — covering about 30 million years, anchored at its young end by uranium-lead zircon at 201.6 Ma just below the first flood basalt of the Central Atlantic Magmatic Province.

The obvious objection, and the design that answers it

If the section has a cryptic unconformity — a gap that looks like continuous sediment — the cycle count is short and the scale is wrong.  This is the objection a critic would raise, and it is the one the study was built to answer.

Rather than argue about it, Kent, Olsen and Muttoni went to a different basin on a different continent margin:  they drilled the Chinle Formation in Petrified Forest National Park, Arizona, obtained its magnetic polarity stratigraphy, correlated that to the Newark APTS pattern, and had its volcanic ash beds dated by uranium-lead chemical-abrasion thermal-ionisation mass spectrometry — a method with no connection to lake cycles at all.

The numbers

The Black Forest Bed in the Chinle correlates magnetostratigraphically to Chron E16r in the Newark scale.  Counting 405-kyr cycles back from the 201.6 Ma anchor, the APTS predicted an age of 209.95–210.25 Ma for that chron before the date was known.

The measured uranium-lead dates:  210.08 ± 0.22 Ma from the drill core and 209.93 ± 0.07 Ma from an outcrop 1.5 m above the base — differing from each other by an insignificant 0.15 Myr.¹

And the interval test, which is the part that matters:

"the age difference (8.50 My) estimated from slightly less than 21 long eccentricity cycles between the U–Pb-dated tie points at 210.08 Ma for the BFB within Chron E16r… and 201.6 Ma just below first CAMP basalt… is indistinguishable from the difference in the U–Pb dates (8.48 My, with an estimated uncertainty of ±0.3 My from propagation of the individual 2σ errors), which would imply a cycle duration of close to 404 ky, or within an insignificant 1 ky or 0.25% of the hypothesized 405-ky period."¹

Two supporting figures:  two entirely separate regressions of depth against age — one using only the uranium-lead dates, one using only the magnetostratigraphic correlation to the APTS — give sediment accumulation rates of 34.3 and 34.8 m/Myr, with R² of 0.9994 and 0.9845.¹

The authors' conclusion:  "this immediately refutes suggestions that significant time is missing in the Newark–Hartford sequence used to build the APTS."¹

One sample does not fit, and the authors say so.  "sample GPL has a reported U–Pb CA-TIMS zircon date of 218.02 ± 0.28 Ma that is much too old for its projected level," and they defer it pending further work.¹

Why this is the strongest single number in the series

A gravitational clock and a uranium clock, on the same two horizons, agreeing to 0.25% across 8.5 million years at 210 million years ago.  Orbital periods are set by Newtonian gravity and planetary masses;  they contain no decay constant.  Uranium-lead contains nothing else.

And the escape a young-earth reader will reach for does not work here.  Suppose decay was accelerated during the Flood by a factor f.  Both uranium-lead dates are then inflated by roughly f, and so is the difference between them, because the difference of two ages scales with a common multiplicative factor.  The cycle count does not change.  So the observable — (uranium-lead interval) ÷ (405 kyr × cycles counted) — would be f, and it is measured as 404/405 = 0.9975.  Accelerated decay operating between 210 and 201 Ma is bounded to a fraction of a per cent.

Part 3: The Latemar, and what a failure looks like

The Latemar is a Middle Triassic carbonate atoll about 2.5 km across in the Dolomites of northern Italy.  Its interior is a 470-metre lagoonal succession of roughly 600 shallowing-upward cycles, and they are visually obvious — a photograph of the Latemar looks like a stack of pages.

They were read as precession cycles.  Six hundred precession cycles at about 20,000 years each gives 9 to 12 million years, and that duration was published and defended.

Then the ash beds were dated.

"U–Pb dating of zircons from volcanoclastic layers within the Latemar succession and the correlative basinal Buchenstein Beds suggests that the ~600 Latemar cycles instead span only up to a few million years.  The application of two sophisticated modern techniques to measure geologic time—astrochronology and U–Pb single-crystal zircon dating—thus leads to age estimates of Latemar carbonate platform deposition that differ by about one order of magnitude."²

Four independent constraints all point the same way against the cycle count.

• Ammonoid biostratigraphy. The whole 470 m falls within slightly more than one ammonoid zone, and that zone "is less than 10 m thick in the Buchenstein Beds at Seceda."²
• Uranium-lead on the Buchenstein ash beds. The correlative Secedensis Zone "encompasses less than 1 m.y."
• Uranium-lead inside the Latemar itself. Three ash layers give "an overall sediment accumulation rate of ~500 m/m.y."
• Magnetostratigraphy — the new evidence in the 2004 paper. The Latemar is predominantly a single normal-polarity interval. A 9–12 Myr duration would make that chron anomalously long in a period when "polarity interval lengths average only about 0.5 m.y."²

The verdict:  "a duration about an order of magnitude less (~1 m.y.)… is compatible with all the other age constraints with the notable exception of the cycle stratigraphic analyses."  And the reinterpretation of the beds themselves:  "the visually obvious meter-scale bedding in the Latemar does not reflect precessional forcing but instead must lie in the sub-Milankovitch (millennial) band of variation…  Our estimated mean bedding duration of ~1.7 k.y."²

Mundil and colleagues drew the methodological lesson, and it is the right one:

"Our findings show that frequency spectra should only be interpreted in combination with robust age control."⁷

Print the Latemar prominently.  It answers two opposite bad arguments.  Against the claim that cyclostratigraphy is unfalsifiable:  here it was falsified, by radiometric dating, and the field changed its mind in print.  Against complacency about periodic-looking beds:  six hundred of them were read wrong by a factor of ten.

Part 4: The circularity problem, at full strength

Jake Hebert's 2014 paper in the Answers Research Journal is the fullest young-earth treatment of orbital tuning, and most of its factual claims check out.  The argument should be stated as he states it before it is answered.

The four-part series that states it most fully

Before Hebert's paper, the fuller treatment is John Reed and Michael Oard's four-part Cyclostratigraphy series in the Creation Research Society Quarterly (2015, 2016, 2020, 2020).  Parts I and II describe the method and its history;  Part III attacks the Milankovitch mechanism;  Part IV asks whether the pre-Pleistocene record is orbitally forced at all, and it is the one that bears on this article.⁴

Its six problems, and what each is worth:

1. The astronomical record cannot be computed through deep time.  This is the strongest point, and it is made with a quotation from the people who compute it.  Laskar and colleagues, on the effect of the asteroids Ceres and Vesta:  "it will never be possible to recover the precise evolution of the Earth's eccentricity beyond 60 Myr."⁴  Oard and Reed add:  "That is why, at present, geologists are cautious in extending cyclostratigraphy beyond the Cenozoic, but we predict that future studies will return the desired answers and push the dating method as far back as is needed."

2. There is no established mechanism connecting insolation to sedimentation.  Quoting de Boer and Smith:  "Processes of sedimentation clearly depend on many more variables than insolation alone, and variations in insolation must be propagated through the complexities of the system before they can be encoded in the sediment.  The mechanisms for this are not yet clear and they may be many and various."  And the related question:  if Milankovitch cycles ran through deep time, "Where are the innumerable ice ages?"⁴

3. Cycles are dated by methods two orders of magnitude less precise than the cycles themselves.  Quoting Miall and Miall:  "We suggest that attempts to develop a time scale with an accuracy and precision in the 10⁴-year range by calibrating it against conventional chronostratigraphic dates up to two orders of magnitude less precise presents a fundamentally flawed methodology."  And magnetic reversals, offered as an independent check, "have relatively little individuality, one reversal looks like another" — Torrens's "bar-code effect."⁴

4. Sedimentation rates are neither constant nor knowable.  Here they marshal Andrew Miall's own work:  the Mesaverde Group preserves "a record of only 10% of the time, while 90% is not represented by any strata at all," and Miall's own question, quoted against him — "Is the stratigraphic record fundamentally unrepresentative of the geological past?  These conclusions would appear to invalidate virtually the whole of the last two centuries of stratigraphic progress."⁴

5 and 6. Natural variability and the bandwagon effect.  Including Zeller's 1964 experiment in which geologists given logged sections to correlate produced divergent interpretations of patterns that were not there, and Bailey's judgement that "This tendency to circular reasoning is endemic in cyclostratigraphic analysis and is acknowledged by its proponents."⁴

They also give a specific, checkable case of a radiometric date being adjusted to fit the astronomical scale — the Monte dei Corvi succession, where "The discrepancy leads to an adjustment of the dating of the two ash layers — the independent data upon which the calibration is based," and "The 'astronomical' time calibration thus supersedes the radiometric data."⁴

And the newest one

The most recent young-earth engagement, published in August 2026, is Michael Thompson's critique in the Journal of Creation of Huang et al. 2024, a cyclostratigraphic reconstruction of the Earth–Moon distance from eight sections spanning 245 to 570 Ma.⁸  Two of its points are worth carrying:

• The 100,000-year eccentricity term is a composite. "astronomically, the eccentricity doesn't have a 100 ka cycle, but rather a 95 ka cycle and 125 ka cycle. Early signal processing methods led to these being unreasonably lumped together into a 100 ka cycle."⁸
• Flood hydrodynamics offers a rival source of cyclicity. A global flood would carry a spectrum of wave periods — semidiurnal and diurnal tides, infragravity waves of minutes, tsunami and storm-surge periods of hours, Coriolis-driven transtidal waves of days — and with fast sedimentation, "the undulations in rock or sediment-layer data can be fitted to these sorts of periods." Thompson's own comment on this is candid: "both the global Flood model and the old age geology models have mechanisms for cycles… both with almost ubiquitous flexibility (arguably the Flood model has more possible options). This flexibility… actually reduces the explainability of any such model."⁸

Hebert's case

Tuning is a fit, not a measurement.  Marine δ¹⁸O records are dated by lining their wiggles up with a computed insolation curve.  If the astronomical theory is assumed and then used to date the sediments, the resulting agreement between sediments and theory proves nothing.

The "independent checks" are not independent.  Hebert examines the two usually offered — biostratigraphic events and paleomagnetic reversals — and argues both are dated by radiometric methods that have themselves been adjusted to the astronomical scale.

And he documents the adjustment.  Three specific, verifiable cases:

• The potassium-argon age of the Matuyama–Brunhes reversal, 730,000 years, was revised upward to 780,000 in 1990 on the basis of the astronomical theory: "Because the K/Ar dates were a little younger than those demanded by the astronomical theory, these K/Ar dates were revised upward. Likewise, the ⁴⁰Ar/³⁹Ar dates were deemed more accurate because they agreed with the astronomical theory."³
• The Fish Canyon sanidine standard, on which most argon-argon dates depend, was itself recalibrated against astrochronology by Kuiper and colleagues in 2008.³ Kuiper's own abstract confirms this: before the recalibration, astronomical and radioisotopic scales "yield discrepancies of ∼1.0% or more," and the fix was to tie argon-argon to the astronomical scale, cutting the method's absolute uncertainty from about 2.5% to 0.25%.⁹
• Ice-core chronologies are tuned to marine chronologies which are tuned to insolation, and marine cores are then tuned to ice cores. He quotes the Vostok chronology paper directly: "we have applied the orbital tuning approach to derive an age-depth relation for the Vostok ice core, which is consistent with the SPECMAP marine time scale."³

And he points at the residual disagreements.  The Devils Hole calcite record, uranium-dated, puts the second-to-last deglaciation at about 140,000 years — roughly 10,000 years before the insolation increase supposed to have caused it.  And a 2010 recalibration of argon-argon by Renne and colleagues produced a Cretaceous–Tertiary boundary age about 279,000 years older than the orbitally tuned value, a difference significant at 95% confidence.³

What holds

On the Neogene and Quaternary, essentially all of it.  The argon-argon scale in that interval is calibrated to astrochronology, and after 2008 the two are not independent.  Marine chronologies are tuned.  Deep Antarctic ice chronologies are tuned.  Presenting these as mutually confirming is a real error, and it is made routinely in popular old-earth writing.

The residual disagreements are real too.  Channell and colleagues, working on the same reversal, find the astrochronological and argon-argon ages hard to reconcile:  five North Atlantic records put the Matuyama–Brunhes midpoint at 773.1 ka with a standard deviation of 0.4 kyr, about 7 kyr younger than the accepted astrochronological 780–781 ka, and the two proposed revisions of the Fish Canyon standard would push argon-argon ages 8 to 24 kyr older still.⁶  That is a live, published disagreement between methods.

What does not hold

The conclusion does not follow from the premises, in three places.

First, the Triassic test is outside the loop.  Kent's 2018 result uses uranium-lead, not argon-argon;  it is in the Triassic, not the tuned Neogene;  it counts cycles rather than fitting a curve;  and the 405-kyr period is a result of the comparison (404 kyr), not an input.  Hebert's mechanism of circularity has no purchase on it.

And it is the specific answer to the Laskar objection, which is why the omission matters.  Oard and Reed's first and best problem is that the orbital solution cannot be integrated past about 60 Myr, so a cycle identified at 210 Ma cannot be identified from theory.  That is right, and it is exactly why Kent and colleagues did not try.  They counted cycles between two horizons, took the uranium-lead interval between the same two horizons, and divided.  The period came out at 404 kyr against a predicted 405.  Laskar's chaos limits what can be computed;  it does not limit what can be measured, and the measurement is what the paper's title claims.  Part IV of the series was published in 2020 and does not cite the 2018 paper.

The same applies to the sedimentation-rate problem.  Oard and Reed's fourth problem — that hidden gaps make cycle counts unreliable, with Miall's 90%-missing estimate behind it — is the objection the Newark study was designed around.  Rather than argue that the Newark basin has no gaps, Kent's team correlated it magnetostratigraphically to a basin 3,000 km away and dated that one.  A gap in Newark large enough to matter would have shown as a mismatch between the predicted and measured age of the Black Forest Bed.  It did not:  the prediction was 209.95–210.25 Ma, the measurement 210.08 ± 0.22 Ma.

The bar-code objection is fair and partly answered by the same design.  Murphy and Salvador are right that one magnetic reversal looks like another.  What distinguishes a polarity pattern is the sequence of interval lengths, and the Newark correlation was tested by whether an independent radiometric date fell where that pattern predicted.  It did.  That is not proof the correlation is right;  it is a passed test that could have failed.

Second, an untuned Pleistocene chronology exists and answers the charge directly.  Peter Huybers and Carl Wunsch built marine age models using depth as the proxy for time — correcting for compaction, modelling sediment-accumulation variability as a stochastic process, and estimating the resulting age uncertainty explicitly.  Their 2004 result:  "Depth-derived ages are estimated to be accurate to within ±9000 years, and within this uncertainty are consistent with the orbitally tuned age estimates."¹⁰  Extended to two million years in 2007:  "An agemodel not relying upon orbital assumptions is estimated over the last 2 Ma using depth in marine sediment cores as a proxy for time…  Major deglaciation features are identified over the last 2 Ma and a remarkable 33 out of 36 occur when Earth's obliquity is anomalously large."⁵

That is the test Hebert says has not been done:  build the chronology without orbital assumptions, then ask whether the orbital signal is still there.  It is, and the chronology matches the tuned one.  Huybers is not cited in any creationist treatment of orbital tuning found for this article.

Third, the Matuyama–Brunhes reversal has an age that does not come from tuning at all.  Channell and colleagues report the beryllium-10 flux peak marking the reversal in the Antarctic Dome C ice core at 771.7 ka on the EDC2 age model and 766.4 ka on EDC3 — measured in a different archive by a different method — and use it to adjudicate between the two ice-core age models.⁶  The reversal is not a free parameter.

And the honest residue

Hebert's advice to his own side is worth quoting, because it is more careful than his conclusion:

"creation researchers should exercise extreme caution when attempting to use oxygen isotope seafloor sediment data to constrain post-Flood models of earth history, as these data have been manipulated to agree with the evolutionary, old earth paradigm."³

He is warning creationists off a data set that some of them use.  A reader is entitled to notice both that the warning is well founded and that it does not amount to a demonstration that the cycles are not there.

Objections and responses

"Orbital tuning is circular — the theory is used to date the sediments, and then the sediments are used to confirm the theory." Correct as a description of tuning, and this article does not use tuned chronologies as evidence.  The evidence used is counted cycles compared against uranium-lead dates, in the Triassic, outside the tuned interval.

"The argon-argon scale was calibrated against astrochronology, so any agreement between them is manufactured." Correct, and it was done deliberately and published as such — Kuiper's paper is titled "Synchronizing Rock Clocks of Earth History."  It is precisely why this article uses uranium-lead rather than argon-argon for the headline test.

"Ice cores are dated by the marine record, and the marine record is dated by ice cores." For the deep, uncounted parts, this is a fair description of some published age models, and Hebert documents specific cases.  It is not true of the counted part of the Greenland record, which is treated in Ice-Core Layer Counting and is checked against argon-argon, uranium-thorium and tree-ring markers.

"Six hundred cycles at the Latemar turned out to be one million years, not twelve.  Cycle counting is unreliable." This article agrees and says so at length.  What the Latemar shows is that cycle counting can fail badly and that radiometric dating catches it — which is the opposite of the picture in which the methods are tuned to agree.  A tuned system does not produce order-of-magnitude public disagreements.

"The Milankovitch mechanism is too weak to cause ice ages anyway." A separate and largely fair criticism, discussed in The Ice Age.  It does not touch the argument here, which does not require the orbital variations to cause anything in particular — only that a sedimentary rhythm with the right period exists and can be counted.

"Devils Hole contradicts the astronomical theory by 10,000 years." It does, at Termination II, and the disagreement is real and still argued about.  It moves one boundary;  it does not change the number of cycles.

"The orbital solution is chaotic and cannot be computed beyond about 60 million years, so a cycle at 210 Ma cannot be identified." Correct, and it is the best objection in the young-earth literature on this subject.  It is also the reason the Newark test was built the way it was:  the cycle count and the uranium-lead interval are compared against each other, and the period is the output.  Recovering 404 kyr where 405 was predicted is a measurement, not an extrapolation.

"Ninety per cent of the time is missing from the rock record, so counted cycles cannot measure duration." Miall's estimate is real and it is quoted here at full strength.  It is also the reason the Newark study cross-checked against a second basin on a different continental margin:  missing cycles would make the counted interval too short relative to the uranium-lead interval, and the two agree to 0.25%.  A record with 90% of its time missing cannot produce that agreement by accident.

"Where are the innumerable ice ages, if Milankovitch cycles ran through all of deep time?" A fair question, and the answer is that the cycles are a forcing, not a switch:  they redistribute sunlight by a few per cent, which glaciates a planet only when the background state is already cold enough — polar continents, low carbon dioxide.  For most of the Phanerozoic it was not.  This is a real weakness in popular presentations that treat orbital forcing as a sufficient cause, and the mechanism critique in Part III of Oard and Reed's series lands on that version.

"Flood hydrodynamics could produce the same layering." Thompson's proposal, and he states its problem himself:  a mechanism with "almost ubiquitous flexibility" explains less, not more.  What distinguishes the orbital reading is that it makes a numerical prediction — a specific period, in years — which can be checked against an independent clock.  A tidal or storm-wave model of the Newark cycles would have to predict that counting them between two horizons yields 8.50 Myr against a uranium-lead 8.48 Myr, and no such prediction has been offered.

"Accelerated nuclear decay during the Flood would inflate the uranium-lead dates." It would inflate both endpoint ages by the same factor and therefore inflate the interval by that factor too.  The interval is compared against a count of orbital cycles which contains no decay constant, and the ratio comes out at 0.9975.  That bounds any acceleration operating across the interval to a fraction of a per cent.

"The 405,000-year period is assumed, not measured." It is predicted by the orbital solution and then recovered from the data at 404,000 years.  That is the direction a measurement runs.

Issues and unanswered questions

• Kuiper et al. 2008 could not be obtained beyond the publisher-deposited abstract. The body would give the Moroccan tephra ages and the astronomical-versus-argon residuals, which are the numbers the circularity discussion most wants.
• Mundil et al. 2003 is abstract-only. The uranium-lead ash-bed dates for the Latemar are used here as Kent 2004 reports them.
• Meyers 2008, the statistical adjudication of both the Latemar and the Green River Formation, could not be obtained. Search summaries report that it finds an orbital signal at the Latemar at a different scale from the metre beds, consistent with Kent's reassignment, but its p-values are not printed here because the paper was not read.
• The Matuyama–Brunhes discrepancy is unresolved. Channell and colleagues find that the Fish Canyon standard age best fitting the astrochronological ages is 27.93 Ma, younger than both recently proposed values; they cannot reconcile all three constraints.
• Oard and Reed's Part IV names sources this article has not read, among them Miall & Miall 2004, Miall 2014/2015/2016, Bailey 2009, Bailey & Smith 2008/2010, Smith et al. 2015, Zeller 1964 and Westphal et al. 2004. Their quotations are internally consistent and several were checked against the published abstracts, but the primaries have not been obtained. Miall 2014 and Miall 2015 are the two that most deserve reading directly, because the missing-time argument rests on them.
• Huang et al. 2024, the Earth–Moon reconstruction Thompson critiques, has not been obtained. Nothing here rests on it.
• The Devils Hole Termination II causality problem is unresolved. A second uranium-dated Nevada record, from Lehman Caves 400 km away, times the same deglaciation consistently with orbital forcing while conceding that the early Devils Hole onset "remains robust."

Verdict

Two things are true at once here and both belong in any honest account.

Cycle counting is genuinely fallible, and the field's own worst case is a factor of ten.  Orbital tuning is genuinely circular where it is used as a dating method and then cited as confirmation of the theory it assumed, and the young-earth critique of the Neogene time scale is substantially correct — the argon-argon standard really was calibrated against astrochronology, and the two really are not independent afterwards.  Oard and Reed's four-part critique is a serious piece of work built almost entirely out of mainstream self-criticism, and its central quotation, from Laskar, is exactly the right question to ask.

And the Newark–Hartford result stands outside all of that.  Twenty-one counted eccentricity cycles against two uranium-lead dates, 8.50 million years against 8.48 ± 0.3, in the Triassic, with the cycle period recovered rather than imposed and a blind prediction confirmed from a basin 3,000 km away.

A young-earth chronology needs that comparison to be wrong by a factor of about fifty thousand.  It is wrong by 0.25%.

Further reading

• Non-Radiometric Clocks — the overview of this series
• Growth Bands and Tidal Rhythmites — the other astronomical counting clock
• Ice-Core Layer Counting
• Do the Dating Methods Agree?
• The Ice Age — the Milankovitch mechanism and its real difficulties

Sources