Folded Rock Layers and the Age of the Earth

"When solid, hard rock is bent (or folded) it invariably fractures and breaks because it is brittle.  Rock will bend only if it is still soft and pliable—'plastic' like modeling clay or children's Playdough." — Andrew Snelling, Answers in Genesis, 2009¹

"Ductile deformation preserves continuity of originally continuous structures and layers, and describes a scale-dependent deformation style that can form by a range of deformation mechanisms." — Haakon Fossen, Structural Geology, 2010²

"In conclusion, therefore, the consensus from all these estimation methods is that the Tapeats Sandstone prior to the Laramide deformation responsible for the Carbon Canyon fold would have been subjected to a burial temperature of 110–130°C." — Andrew Snelling, Answers Research Journal, 2023³

Summary

Stand in Carbon Canyon, a side canyon in the eastern Grand Canyon, and you can see a bed of sandstone turned through a right angle over about thirty metres without shattering.  The layer above it does the same.  So does the limestone higher up.  Photographs of that fold have been the second item on Answers in Genesis's list of the ten best evidences for a young earth since 2012.⁴

The argument is short.  Hard rock is brittle.  Bend it and it breaks.  These layers bent and did not break, so they were still soft when they bent.  Conventional geology says the lowest of them, the Tapeats Sandstone, was deposited about 507 million years ago and folded about 60–70 million years ago.  Sand does not stay soft for 440 million years.  Therefore the 440 million years never happened, and the whole sequence was laid down and folded within months of each other — in the Flood.

The first premise is false, and it is the whole argument.  Rock is brittle when you hit it with a hammer at the surface.  Under a few kilometres of overburden, over long times, the same rock deforms without breaking, by mechanisms that have nothing to do with melting or metamorphism: slip between layers, slip along cross-bedding surfaces, dissolution of quartz where grains press hardest and re-precipitation where they do not, and millimetre-scale cracking distributed through the rock.  This is not a rescue device invented to answer creationists.  It is chapter 6 of the standard undergraduate structural geology textbook, and it is in the same book Snelling cites.

Since 2017 the argument has had a serious technical version behind it.  Andrew Snelling fought the National Park Service for a research permit, won it, collected fifty-three samples from four folds in the Grand Canyon, and published seven long papers on them between 2021 and 2026.⁵³⁶⁷⁸  Nobody had made a thin section of these folds before.  That work is real, it was needed, and this article gives it credit before disagreeing with it.

The disagreement comes down to five things.

1. The test Snelling chose cannot come back positive.  He says the deciding evidence is whether the quartz grains show undulose extinction and deformation lamellae — and in 2026 he calls this "the only definitive method."⁸  Those are crystal-plastic features.  Quartz does not deform that way below roughly 300–350 °C.²  In the same body of work Snelling establishes that the Tapeats never got hotter than 110–130 °C.³  So his own thermometer guarantees his own microscope test will find nothing, whether the fold formed in a month or in twenty million years.  A test that must return "absent" tells you nothing when it returns "absent."

2. The mineral evidence points the other way.  The 110–130 °C figure rests partly on clay minerals: the proportion of illite in the mixed-layer illite/smectite of two thin volcanic ash beds in the Tonto Group.⁹  That conversion is a chemical reaction with measured kinetics.  Run the published rate law with Snelling's own numbers and the reaction he measured needs 370,000 years at 130 °C and 2.3 million years at 110 °C for the less extreme of his two samples.  In one year at 130 °C, less than one thousandth of one per cent of the smectite converts.  To finish the job in a Flood year you would need about 360–410 °C — greenschist conditions, and exactly the metamorphism whose absence is his headline result.¹⁰

3. Bending a stack of layers takes very little strain.  The famous 90° photograph looks like it demands the impossible.  Work the geometry and it does not.  A 0.3-metre bed bent through a right angle over 30 metres has to stretch on its outer face by under one per cent.  In sandstone that is strongly cross-laminated — which Snelling says the Tapeats is³ — even that vanishes, because the millimetre-spaced cross-bed surfaces can slip past one another.  Cooke and colleagues documented exactly this mechanism in the same East Kaibab Monocline, in the cross-bedded Navajo Sandstone: "Slip along dune boundaries, as evidenced by joint clusters oblique to bedding."¹¹

4. The fold is a flexural-slip fold, and Snelling says so.  He classifies the Carbon Canyon fold as a Ramsay Class 1B parallel-concentric flexural-slip fold, records that "bedding plane or flexural slip must have occurred to accommodate their folding," and reports that bed thickness is constant from limb to hinge.³³  Constant bed thickness plus bedding-plane slip are the two textbook diagnostics of flexural slip in lithified, layered rock.  He has documented the mechanism and then denied its precondition.

5. The rock had to get hard afterwards, and that clock is worse.  Snelling's model needs the quartz cement to precipitate after the folding, inside a few hundred years.  Measured quartz cementation rates from North Sea reservoirs give six to ninety-four million years to precipitate the cement in a sandstone of this grain size at 110–130 °C.  Three hundred and fifty years buys about three-thousandths of one per cent of it.¹²

What is genuinely right in the argument

Soft-sediment deformation is real, common, and not a creationist invention.  Sediment does slump, fold, and inject while wet, and geologists have a whole vocabulary for it.  Steven Austin's 1986 study of the Split Mountain fault in California describes a genuine case, and his description of it is good.¹³  It simply does not need a global Flood: the deformation there is confined to "the 30 feet nearest the fault," in Pliocene sand a few kilometres deep in one of the fastest-subsiding basins in North America.

Snelling is also right that nobody had done the microscope work before he did, right that his critics have not been into the canyon, and right that a line drawing traced over someone else's photograph is a weak thing to argue from.⁸  His scanning-electron images of undisturbed cement crystals are the strongest single piece of evidence he has produced, and this article says where that evidence really points rather than waving it away.

And the popular version of the mainstream answer is often careless.  "Rocks fold slowly under pressure" is true but empty unless you say by what mechanism, at what temperature, and at what rate.  Most published rebuttals do not.

What this article does

It states the argument in its own words, at both the popular and the technical level, then tests it against measurements.  Where a number is computed rather than quoted, the computation is printed so a hostile reader can redo it; two short scripts are included beside this article.  Every source is catalogued, with what it measured and which way it cuts, in a companion page: the literature.

Where the argument comes from

The claim is old.  The version the TalkOrigins archive answered in 2003 came from Russell Humphreys's tract Evidence for a Young World.¹⁴  Its modern form is Andrew Snelling's, and it has three layers.

The popular layer is the 2009 Answers magazine article "Rock Layers Folded, Not Fractured," republished many times since.¹  Its core sentence is the one at the head of this page: bent rock "invariably fractures and breaks because it is brittle."  It uses the East Kaibab Monocline, and specifically the Carbon Canyon fold in the Tapeats Sandstone and the folded Muav and Redwall limestones in Kwagunt Creek.  It adds a second step that matters:

"The conventional explanation is that under the pressure and heat of burial, the hardened sandstone and limestone layers were bent so slowly they behaved as though they were plastic and thus did not break.  However, pressure and heat would have caused detectable changes in the minerals of these rocks, tell-tale signs of metamorphism.  But such metamorphic minerals or recrystallization due to such plastic behavior is not observed in these rocks."¹

That is the argument this article is really about.  It is not "rock breaks."  It is "slow folding requires heat, heat leaves fingerprints, and the fingerprints are missing."

The list layer is the 2012 article "Bent Rock Layers," number two of Answers in Genesis's 10 Best Evidences from Science That Confirm a Young Earth.  It supplies the analogy everyone remembers: "Hardened rock layers are brittle.  Try bending a slab of concrete sometime to see what happens!"⁴  Creation Ministries International's 2021 magazine version, by Gavin Cox,¹⁵ uses a dry twig instead: "Bend an old dry twig, and it will snap—suddenly!"¹⁵

The technical layer is the research programme.  On 25 November 2013 Snelling applied for a National Park Service permit to collect fifty to sixty fist-sized samples from four folds inside Grand Canyon National Park.  The permit was refused.  Alliance Defending Freedom lawyers filed suit in May 2017 alleging worldview discrimination;  the Park Service issued the permit on 23 June 2017 and the case was dismissed five days later.¹⁶  He collected fifty-three samples and published:

Two retired geologists, Stephen Mitchell and Kennen Tillman, published a three-part critique on the Peaceful Science site in August 2024.¹⁷  Snelling answered in three parts in Answers in Depth in early 2026; part three, on deformation, is the most recent word in the exchange and is quoted here throughout.⁸  Lorence Collins, a retired petrologist at Cal State Northridge, posted a shorter critique in 2022.¹⁸

What the folds are

The Grand Canyon region is cut by monoclines — step-like folds where flat-lying strata bend down (or up) across a zone and then flatten out again.  Everyone, creationist and conventional, agrees on their cause.  Faults in the Precambrian basement, formed as normal faults roughly a billion years ago, were squeezed and reversed during the Laramide mountain-building event; the Palaeozoic layers draped over the moving basement blocks and bent.  Snelling reproduces the standard cross-section for the Butte Fault and the East Kaibab Monocline and says the tectonic mechanism "is not in dispute," only the timing.³

This is the class of fold that structural geology calls a forced fold or drape fold, and it is a textbook case.  Haakon Fossen's Structural Geology, the textbook Snelling leans on throughout his own papers:²

"Classic geologic results of bending are the forced folds created in sedimentary layers blanketing faulted rigid basement blocks.  The displacement is forced upon the sediments by fault movement on a preexisting basement fault, and the sediments are soft enough to respond by monoclinal folding until at some critical point they rupture and the fault starts propagating up-section.  Such structures are particularly well exposed in the Colorado Plateau–Rocky Mountains area, where numerous Laramide-related uplifts have created such structures."²

Note the word "soft" there.  Fossen means mechanically weaker than crystalline basement, not uncemented.  That distinction runs through this whole dispute and is worth holding onto.

The four folds Snelling sampled are:

• the Carbon Canyon fold, Tapeats Sandstone, on the Butte Fault in the eastern canyon — the one in the photographs;
• the Monument fold, Tapeats Sandstone, on the Monument Monocline in the central canyon;
• the Whitmore Helipad fold, Bright Angel Formation, in the western canyon;
• the Matkatamiba fold, Muav Formation limestone, in the central canyon.

A fifth structure, the Palisades Creek branch of the East Kaibab Monocline, matters because it is the one place in the region where a full structural study exists — Reches and Matthews mapped it in 1978, and James Orofino re-studied it for a Texas A&M master's thesis in 2005.¹⁹  Snelling argues that Palisades work is irrelevant to Carbon Canyon because it concerns a different branch of the fault: "That is akin to applying features of apples to oranges!"⁸  He has a point about extrapolating detail.  He does not have a point about extrapolating mechanism, since the two structures are the same kind of fold on the same fault system in the same rocks.

The claim in its strongest form

Set out fairly, the technical argument is this.

1. If lithified sandstone were folded ductilely, the strain would show up inside the grains: sub-grains, undulose extinction, deformation lamellae, deformation kink bands. It would also show as grain-boundary sliding, grain rotation, and grain fracturing.
2. If lithified sandstone were folded ductilely, the folding would deform or break the cement holding the grains together.
3. None of that is present. The grains in the fold are indistinguishable from the grains in samples taken miles away where the same bed is flat. Detrital muscovite flakes are still bent around quartz grains with their ends still frayed. Scanning-electron images show quartz cement crystals with intact terminal faces sitting in open pores, undisturbed.
4. Therefore the cement formed after the folding; therefore the sand was uncemented when it bent; therefore the folding happened before lithification, which cannot have been 440 million years after deposition.

Points 1 through 3 are observations, and there is no reason to doubt them.  Point 4 is the inference, and it does not follow.  Here is why, in five parts.

1. The test could not have passed

In February 2026 Snelling stated the decision rule as plainly as anyone could want:

"Only if thin section examination of the sandstone reveals deformation lamellae and undulose extinction in the quartz grains under cross-polarized light due to deformation stress will it have been demonstrated that the mechanical crowding of the sandstone beds in these folds has been due to slow [ductile] deformation under stress of solid [cemented] sandstone."⁸

And:

"Thus, the only definitive method for distinguishing between slow ductile deformation of lithified sandstone layers and soft-sediment deformation of unlithified sandstone layers is microscopic examination of the minerals and textures within those sandstone layers."⁸

Undulose extinction and deformation lamellae are produced by dislocation creep — atoms in the quartz lattice shuffling under stress.  Quartz does not do this at low temperature.  Fossen, in the textbook Snelling cites throughout:

"Quartz deforms by brittle mechanisms up to about 300–350 °C, which for typical continental temperature gradients corresponds to crustal depths of around 10–12 km.  At greater depths, crystal-plastic mechanisms (creep mechanisms) and diffusion dominate."²

Snelling knows this.  His own paper says the Tonto Group "most likely were not buried deep enough to experience ductile deformation as they were well above the brittle-ductile transition zone."³  He then establishes, from three independent methods, that the maximum burial temperature was 110–130 °C and the confining pressure 0.3–0.4 kbar, and concludes that "these pressure-temperature conditions are well within the pressure-temperature field of sedimentary diagenesis."³

Put those together.  The rock never reached the temperature at which quartz can produce the microstructures.  So the microstructures were never going to be there.  Their absence is a fact about the burial temperature, not a fact about the folding rate or the folding date.

This is not a quibble about wording.  It is the load-bearing beam of the whole argument.  Snelling is looking for a signature that only a hot fold could leave, in a rock he has proved was never hot, and reading the null result as evidence about when the fold happened.  Mitchell and Tillman make the same point more gently — "Under these conditions, we do not expect extensive formation of sub-grains, deformation lamellae, kink bands or undulose extinction"¹⁷ — and Snelling reads their agreement as evasion.

What would record folding of hard rock at 110–130 °C is a different list: bedding-plane slip and its traces, fractures whose orientation is tied to the fold, cataclastic bands, indented and sutured grain contacts, stylolites, and — in limestone — mechanical twins in calcite.  Some of those Snelling reports.  Some he mentions and never looks for.  We come to them below.

2. The clay minerals run the clock the wrong way

Here is where the argument turns over.

To establish the burial temperature, Snelling uses three lines: apatite fission-track data from published studies (110–130 °C), the illite crystallinity or Kübler Index, and the smectite-to-illite geothermometer.³  The last two are clay-mineral methods, and they come from X-ray diffraction analyses of two thin volcanic ash beds — one in the Muav Formation, one in the Tapeats Sandstone — that he reported in the RATE project's second volume in 2005:⁹⁹

Volcanic ash alters first to smectite.  Smectite then converts to illite as it is buried and heated.  Because the conversion tracks depth and temperature, the proportion of illite in the mixed-layer clay is used as a geothermometer, which is exactly what Snelling does with it: he plots his two samples on John Hower's curve from a Gulf of Mexico oil well and reads off 4,800–5,600 m and 110–130 °C.

There is a warning attached to that procedure, and it comes from a paper Snelling cites as authority for it.  He introduces the method by writing that "experimental studies of the conversion of smectite to illite have demonstrated its potential use as a geothermometer (Essene and Peacor 1995; Huang, Longo, and Pevear 1993)."³  Eric Essene and Donald Peacor's 1995 review, in the same journal, concludes the opposite:²⁰

"Currently used clay mineral thermometers (illite crystallinity, smectite/illite reaction, chlorite composition) are not based on equilibrium reactions.  Such systems are not accurate thermometers and therefore have questionable utility."²⁰

Their reason is this article's argument in a sentence: "Chemical reactions involving most clay minerals are best understood with kinetic models.  These involve a great variety of parameters such as time, fluid/rock ratio, deformation history, nature of starting materials and transformation mechanisms, as well as the variables, such as temperature, pressure and composition, that are commonly used to define equilibrium."²⁰

Time is first on their list.  To Snelling's credit he flags a related worry himself — "the smectite/illite ratio relationship to temperature appears to be neither simple nor unequivocal"³ — but he flags it and proceeds, and he does not draw the consequence.  If the illite proportion is a kinetic quantity rather than an equilibrium one, then it is not only a thermometer.  It is a clock.

And the rate of that clock has been measured.

Wuu-Liang Huang, John Longo and David Pevear of Exxon ran the conversion in gold capsules at 250–325 °C and published the rate law in Clays and Clay Minerals in 1993:²¹

"−dS/dt = A·exp(−Ea/RT)·[K⁺]·S²"

with S the fraction of smectite layers, A = 8.08 × 10⁴ s⁻¹ and Ea = 28 kcal/mole.²¹  The printed paper says 8.08 × 10⁻⁴; an erratum in the same volume corrects it: "P. 162, line 8: 'A = frequency factor = 8.08 • 10 _4' should read 'A = frequency factor = 8.08 x 104'."²²  Elliott and Matisoff's 1996 comparison of the three published kinetic models²³ tabulates it independently as A = 80,800 s⁻¹, Ea = 117.15 kJ/mole, confirming the corrected value.²³

Huang and colleagues recommend a potassium concentration of 0.0026–0.0052 mol/L (100–200 ppm) when the pore fluid is unknown, "similar to the value typically reported in oil field brines."²¹  Integrate the second-order law from pure smectite to the illite fractions Snelling measured, at the temperatures Snelling derives:¹⁰

Run the same numbers forward instead.  In one year at 130 °C, 0.0009 % of the smectite converts.  At 200 °C, still only 0.15 %.  To reach 76 % illite within a single Flood year you would need about 364 °C; for 94 %, about 413 °C.  Give the reaction the whole 350-year post-Flood window that Snelling's chronology allows and you still need 230 °C and 261 °C.

Two other published kinetic models give the same verdict.  Pytte's fifth-order model requires 1.6 million years at 130 °C for the Muav sample and 106 million for the Tapeats sample; Velde and Vasseur's two-step model is slower still.¹⁰

There is also a natural benchmark, and it is in Huang's own paper.  The fastest smectite-to-illite conversion known on earth is the Salton Trough Geothermal Area in southern California — an active geothermal field whose brines carry 3,000 ppm of potassium, about seven times as much as seawater.  Huang and colleagues describe it as "the opposite extreme":

"The field observations of shales from the Salton Trough Geothermal Area show rapid and complete conversion of smectite to illite at a depth of less than 3050 m (Jennings and Thompson, 1986) due to thermal events over the last 20,000 years."²¹

Twenty thousand years, in an active geothermal field, in brine seven times richer in potassium than the sea.  Snelling needs the same conversion in one year, at a quarter of the temperature, in ordinary pore water.

The mineral evidence that is supposed to prove the folding was fast is the strongest single piece of evidence in the Grand Canyon that it was not.  The illite could not have formed in a year; it took hundreds of thousands to millions of years at the temperature it records; and on Snelling's own reading of the clay curve it formed while the rock sat under 4,800–5,600 m of overburden — which is to say, before the Laramide uplift stripped that overburden away.

There are two ways out, and both cost more than they save.

The illite is detrital, not diagenetic.  Snelling half-says this: he justifies applying the geothermometer on the ground that "the clay minerals appear to have accumulated at the time of deposition, rather than having formed in situ after deposition," proposing that the eruption's own steam converted the ash to clay before it settled.⁹  Read charitably, he means the smectite was made at deposition and the illite grew from it afterwards — which is what makes the thermometer legitimate, and which is exactly the reading the calculation above assumes.  Read the other way, if the illite was already there when the ash landed, the thermometer measures nothing about burial and the 110–130 °C figure evaporates, taking with it the "well within the field of sedimentary diagenesis" conclusion and the argument that the rock could not have deformed ductilely.

The Flood's fluids were hot.  Then the rock reached the temperature where quartz does creep, and the microstructures Snelling calls definitive should be present.  He cannot have the fluids hot enough to run the clay reaction and cool enough to keep the quartz brittle.

3. The arithmetic of a 90° bend

The photographs do the persuading, and they are misleading.  A right-angle turn looks like something only toffee could survive.  It is worth working out how much the rock actually has to stretch.

Snelling gives both numbers needed.  The Tapeats cliff unit "consists of 0.3–1.0 m thick non-bioturbated beds,"³ and the fold turns the beds through roughly 90° "within a distance of about 100 feet (30 meters)."²⁴

A bend of 90° (π/2 radians) over a 30 m arc means a radius of curvature of about 30 ÷ (π/2) = 19 m.  For a bed of thickness t bent to radius R with its neutral surface in the middle, the outer face must lengthen by (t/2)/R:

• a 0.3 m bed: 0.15 ÷ 19 = 0.8 %
• a 1.0 m bed: 0.50 ÷ 19 = 2.6 %

That is the whole demand the famous photograph makes on the rock.  Under one per cent for the thinner beds.  It is more than sandstone can take elastically, so the outer arc must crack — and cracks are exactly what is there.  Snelling describes them: fractures that "penetrate totally across, but are confined to, individual sandstone beds, including in the hinge zones," and "regularly spaced," with spacing that scales with bed thickness.³  He reads them as unloading joints formed later, when the canyon was carved.  They may be.  But they are also precisely what a lithified bed bent to 19 m radius has to produce, and there is no way to tell which from a photograph.  Deciding it needs fracture orientations measured against the fold axis, and those measurements were not made — a gap Mitchell and Tillman flag and Snelling does not fill.¹⁷

The cross-bedding makes the demand smaller still.  Snelling reports that the Tapeats beds are "strongly cross-laminated."³  Cross-bed foresets are surfaces, millimetres to centimetres apart, along which a sandstone can slip.  If the effective mechanical layer is a 5 mm cross-lamina rather than a 0.5 m bed, the outer-arc stretch drops to 0.013 % — nothing at all — and the slip needed on each surface drops to about 8 mm, far too little to grind a visible slickenside.

This is not a hypothetical.  It is the mechanism Michele Cooke and colleagues documented in 1999 in the same East Kaibab Monocline, in the cross-bedded Jurassic Navajo Sandstone:

"Dune boundaries (bedding planes) were mapped through the thickness of the aeolian Navajo Formation, and the occurrence of joints related to dune boundary slip and fold curvature was documented.  Slip along dune boundaries, as evidenced by joint clusters oblique to bedding, occurs along the steep limb of the fold and in the middle of the Navajo.  Joints perpendicular to bedding and parallel to the fold axis occur near the synclinal hinge."¹¹

Two things follow.  Interlayer slip in a cross-bedded sandstone leaves joint clusters, not necessarily striated surfaces — which answers Snelling's objection that slickensides were found on only one bedding plane in Carbon Canyon and none at all in the Matkatamiba fold.⁷  And the Navajo Sandstone, which lies above the whole Grand Canyon Palaeozoic section and above the lower Mesozoic on top of that, and is some 300 million conventional years younger than the Tapeats, was hard rock when the same monocline bent it.

Snelling's reply is to attack the interpretation of the Navajo as desert dunes: "'dune boundary slip' is an interpretation of the observed folds, so the fold curvature could still be due to damp, soft-sediment deformation."⁸  Whether those cross-beds were built by wind or by water does not matter here.  What matters is that they are slip surfaces, and Cooke's team mapped slip on them.

4. Ductile does not mean metamorphic

The 2009 article's second step — heat would leave metamorphic minerals, and there are none — assumes that a fold in hard rock requires the rock to have flowed like warm plastic.  That is not what structural geologists mean by ductile, and Snelling's own textbook says so on the page he cites for other purposes.

"Ductile structures are well represented in metamorphic rocks, i.e. rocks that have been deformed in the middle and lower part of the crust.  Ductile deformation also occurs in soils and unconsolidated to poorly consolidated sediments where distributed deformation rather than discrete fracturing occurs, even though the deformation mechanisms responsible for the ductile deformation in these cases are quite different.  Hence, ductile deformation is … a scale-dependent structural style and not related to microscale deformation processes."²

And, two pages on: "Hence, a deformed area or volume can be ductile at the seismic or mesoscopic scale and brittle at the subseismic or microscopic scale."²

So a bed can turn a smooth right angle while every grain in it is behaving brittlely.  The mechanisms available at 110–130 °C are:

Cataclastic flow.  Distributed micro-fracturing, grain crushing and frictional sliding, which lets a porous rock change shape continuously at the outcrop scale.  Snelling reproduces Fossen's own diagram of it.³

Frictional grain-boundary sliding and grain rotation — granular flow — which needs no cement to be broken because the cement is not yet everywhere.

Flexural slip on bedding and cross-bedding surfaces, discussed above.

Pressure solution, or stress-induced solution transfer.  Quartz dissolves where grains press hardest and re-precipitates where the stress is lower.  The rock changes shape without any grain ever bending.  Jean-Pierre Gratier, Dag Dysthe and François Renard's 133-page review²⁵ is unambiguous:

"Pressure solution creep is a major mechanism of ductile deformation of the upper crust, accommodating basin compaction, folding, shear zone development, and fault creep and interseismic healing."²⁵

Snelling reviews pressure solution properly in his own introduction, and even reports finding it: "The pressure solution of a few quartz grains from compaction is also evident in the SEM images in this study."³  Having found the low-temperature ductile mechanism in his own samples, he attributes it to compaction and does not consider it as a folding mechanism.  His only quantitative statement about its rate comes from a 1977 paper: "McClay (1977) estimated that stress-induced solution transfer in fine-grained quartz and calcite rocks can produce geologically reasonable strain rates at 200–300°C."³  Forty years of work since then, summarised by Gratier and colleagues, puts pressure solution to work through the entire upper crust from the surface down.

There is a further reason temperature is not the only variable.  Rocks are weaker when you deform them slowly:

"Just think about glaciers that consist of slowly flowing ice that can be shattered by a swift hammer stroke.  Parts of the Earth behave in a viscous manner because of low strain rates.  Increasing the strain rate also makes the rock stronger.  Conversely, rocks are weaker at lower strain rates because crystal-plastic processes can then more easily keep up with the applied stress."²

And confining pressure raises ductility independently of temperature: "Increasing the confining pressure allows for larger finite strain to accumulate before failure."²  A slab of concrete on a workbench is at zero confining pressure, at a strain rate of about one per second.  The Tapeats was under 0.3–0.4 kbar for the whole of the Laramide.  The analogy in the 2012 article compares the wrong two things.

5. Flexural slip is a hard-rock mechanism, and Snelling documents it

The clearest internal difficulty in the technical papers is this one.

Snelling classifies the Carbon Canyon fold, correctly, as a Ramsay Class 1B parallel-concentric fold and a Donath–Parker flexural-slip fold.  He then explains what that means, citing Hatcher and Bailey:

"Thus, in flexural-slip folds, layer thicknesses remain constant, and folding is accomplished by slip along and between layers.  They are easily recognized by slickensides, fibers, or other movement indicators such as slip lines or lineations on layer surfaces, and by constant layer thickness."³

He then reports both diagnostics.  Bed thickness is constant: "there is no mechanical crowding or thinning of the sandstone beds in either the hinge zones or the limbs."³  And slip occurred: "From field observations of the outcrop of the Tapeats Sandstone beds in the Carbon Canyon fold it is quite clear that bedding plane or flexural slip must have occurred to accommodate their folding."³

Flexural slip requires discrete, coherent layers that can slide past one another along mechanical discontinuities.  It is what a book of pages does.  Wet, uncemented sand does not have pages; it has grains, and grains flow rather than slip along surfaces.  Fossen puts flexural slip in the upper-crustal brittle regime and adds the caveat that answers Snelling's slickenside count: "In the latter case, bedding surfaces act like faults, and slickenlines … will sometimes develop on slipping surfaces."²  Sometimes.  Not always.

Snelling's own account therefore has the sand behaving as coherent beds that slip along their contacts while remaining too weak to fracture — a middle state that he asserts rather than demonstrates.  He appeals to laboratory squeeze-box experiments to make it plausible, and those experiments deserve a section of their own, because they cut the other way.

The experiments cited do not discriminate

Snelling's standing answer to every macroscopic feature that looks like hard-rock deformation is that squeeze-box experiments on damp sand reproduce it.  He lists five: Rettger 1935; Handin and colleagues 1976; Friedman and colleagues 1976; Weinberg 1979; and Friedman, Hugman and Handin 1980.³  In 2026 he restates the point: these experiments show that "fractures, flexural slip, and changes in bedding thickness" are all producible in soft sediment, so none of them can be diagnostic.⁸

Four of those five are consecutive parts of one series run at Texas A&M, "Experimental folding of rocks under confining pressure."  The series folded rock.  The title of Part VIII, verified through Crossref, is: "Forced folding of unconsolidated sand and of lubricated layers of limestone and sandstone."²⁶  Part II is "Buckling of multi-layered rock beams"; Part III is "Faulted drape folds in multilithologic layered specimens."²⁷²⁸

So the experimental programme Snelling cites made drape folds both in loose sand and in lubricated layers of solid limestone and sandstone, under confining pressure, at room temperature.  If the same features come out of both, then the features cannot tell you which one you are looking at — which is Snelling's own argument, aimed at his own conclusion.

There is a sharper problem, and Snelling states it himself.  Two pages after listing the squeeze-box work, he summarises the one experiment in his list that examined loose sand under the microscope:

"Borg et al. (1960) conducted a quantitative study of experimental deformation of sand grains washed from the St. Peter Sandstone of Illinois by applying a uniform confining pressure to simulate overburden pressure and differential load to simulate tectonic pressures.  They found that purely cataclastic brittle deformation occurred resulting in fracturing of the sand grains as the most conspicuous feature."³

Deform loose sand under simulated overburden and tectonic load, and the most conspicuous result is fractured grains.  Snelling's argument is that the near-absence of fractured grains in the Tapeats proves it was loose sand.  His own summary of the experiment says loose sand is exactly where you should expect them.

The abstract of Friedman, Hugman and Handin 1980 has not been read in the original for this article — the publisher's site is behind a bot wall and no open copy was found — so nothing is claimed from it beyond its title, which Crossref supplies.  That is enough for the point being made.

The evidence for hard rock that was not collected

Three tests would settle the question at 110–130 °C.  One was partly done, one was mentioned and skipped, and one nobody has done.

Calcite twins

Limestone records low-temperature deformation in a way sandstone cannot.  Calcite crystals twin mechanically under stress, and the twins are a thermometer and a strain gauge at the same time.  Snelling knows this and sets it out, citing Martin Burkhard's 1993 review:

"Burkhard (1993) reviewed deformation twinning in calcite, inferring that micro-twins and straight, narrow twins (< 1 µm thick) are characteristic of very low temperature deformation, whereas above about 100°C wider (> 1–5 µm), and fewer twins occur."⁷

That passage is on page 506 of a 133-page paper about a fold in limestone.  Search the rest of the paper and the word "twin" appears in the results only where he describes the primary growth twins of feldspar and plagioclase grains.  The abstract's list of "microscopic features that should be present if the Muav Formation beds in the fold had been bent via ductile deformation" does not include calcite deformation twins at all.⁷  The single diagnostic that works at the temperature his own thermometer specifies, in the one rock type that carries it, was described in the introduction and never applied.

This matters because someone else did apply it, on the same monocline.  Reches and Matthews mapped the Palisades Creek branch in 1978 and, as Mitchell and Tillman summarise, "documented abundant twinning of calcite grains in the Muav and Redwall Limestones at Palisades Creek," and showed "that these calcite twins indicated shortening perpendicular to the fold axis."¹⁷  Twins that record shortening perpendicular to a fold axis formed during that fold, in calcite that had already crystallised.  (Reches and Matthews 1978 is a Geological Society of America memoir chapter and was not obtained for this article; the finding is reported here at second hand, through Mitchell and Tillman and through Orofino's thesis, which cites the same result.¹⁹)

Slip recorded by things that could not exist in wet sand

Orofino's thesis reports something at Palisades Creek that no soft-sediment model can absorb:

"Flexural-slip folding rotated bedding and existing joints and veins, while offsetting the joints and veins across bedding."¹⁹

A vein is a mineral-filled crack.  For a vein to exist and then be offset by bedding-plane slip, the rock must have been hard enough to crack, wet enough to fill the crack with cement, and hard enough again for its layers to shear past each other.  Wet sand does none of the three.  The fault zone itself carries breccia — angular broken rock — which is the same argument in coarser form.¹⁷

Fracture orientations

The one measurement that would settle the joints-versus-fold-fractures question is the orientation of the fractures relative to the fold axis.  Fold-related extension fractures should run parallel to the hinge; unloading joints should follow the regional stress field and should be no denser inside the fold than outside it.  Mitchell and Tillman say so and note the data do not exist.¹⁷  Snelling replies that this "admit[s] the deficiency in their own armchair study" and that no structural analysis "can 'trump' such definitive microscopic-scale evidence."⁸  But the microscopic evidence is, as shown above, a null result at a temperature where a null result was guaranteed.  The orientation data would be real evidence, and both sides agree nobody has collected it.

Snelling does report one relevant field observation, and dismisses it on grounds that undercut him.  A small fault in a hinge-zone bed "appears to coincide with a cross-bedding parting surface, as it does parallel the cross-bedding, and it is confined to just part of the thickness of this one sandstone bed.  This thus suggests that movement was only within part of the bed and facilitated by the cross-bedding plane surface, which means it was not related to brittle deformation fracturing."³  Slip localised on a cross-bedding surface inside a fold hinge is not a reason to discount the observation.  It is the mechanism Cooke and colleagues mapped in the Navajo, seen once in outcrop.

The rest of the column bent too

The Carbon Canyon fold is not an isolated structure.  It is one small feature on a monocline that runs for hundreds of kilometres and deforms everything from the Precambrian basement up to the Cretaceous.  Whatever bent the Tapeats bent all of it at the same time, and the argument has to work for the whole stack.

Three things in that stack are hard to fit into a soft-sediment model.

The faults.  The monoclines exist because basement faults moved.  Those faults cut up through the sedimentary section, and they carry the marks of brittle failure in rock: breccia, gouge, and polished slip surfaces.  Reches and Matthews describe the Palisades fault zone as a 5–60 m wide zone of breccia and altered rock.¹⁷  Carol Hill and Stephen Moshier put the difficulty plainly in 2009:²⁹

"With regard to faulting, it is extremely puzzling to visualize how thousands of feet of offset along Grand Canyon faults could have been achieved in sediments that were still soft!  In addition, how could slickensides (polished and smoothed striations made in hard rock by fault action), fault gouge and breccias (pieces of angular rock and earthy material along faults), and the sharply offset rock layers along faults (rather than layers slumping into faults) have formed in rock that was still soft?"²⁹

Snelling's answer is that the faults are in the basement and "those faults are not within these folds."⁸  For the deep Precambrian faults that is fair, and he grants that the basement was hard rock long before the Flood.  But the Butte Fault does not stop at the Great Unconformity.  Snelling's own account has the Tapeats beds "dragged up to the Butte Fault as they were upturned through the ~90° fold, and at the Butte Fault they are chopped off and end."³  Wet sand dragged against a moving fault plane smears; it does not get chopped off and end.

The Jurassic Navajo Sandstone.  Cooke and colleagues mapped interlayer slip and curvature joints in the Navajo along the same East Kaibab Monocline in Utah.¹¹  The Navajo sits above the entire Palaeozoic section and the lower Mesozoic besides.  On a Flood chronology it was deposited months later than the Tapeats and, on Snelling's own logic, would have been even wetter and softer when the monocline formed.  It behaved as rock.

Everything in between.  Sarah Tindall and George Davis traced fold-related brittle faults along a 50 km stretch of the East Kaibab Monocline and found them "exposed in all stratigraphic intervals from Proterozoic through Cretaceous."³⁰  Snelling's reply is that "these workers have never studied samples from those folds under microscopes to ascertain the timing of those faults in relation to the folding of the layers."⁸  That is a request for the microscope test discussed above — the one that cannot come back positive at 110–130 °C.  A brittle fault cutting a bed is direct evidence that the bed could break, which is direct evidence that it was rock; it does not need a thin section to say so.

Getting the rock hard afterwards

The soft-sediment model owes an account of when the sandstone became sandstone.  Snelling's answer is that the cement grew after the folding — his scanning-electron images of pristine quartz cement crystals are offered as proof of exactly that — and that it grew fast.  In Earth's Catastrophic Past²⁴ he writes that "under normal conditions sediments lithify relatively quickly, often in a matter of years, but at most, perhaps hundreds of years.  Given ideal conditions, lithification can happen within days."²⁴

Quartz cementation rates have been measured.  Olav Walderhaug³¹ determined them for Jurassic sandstones on the Norwegian shelf by combining thin-section petrography with fluid-inclusion thermometry and burial-history modelling.  His rates "increase systematically with temperature from approximately 1 × 10⁻²⁰ moles/cm².s at 80 degrees C to approximately 5 × 10⁻¹⁹ moles/cm².s at 140 degrees C," and the silica comes "dominantly from dissolution of quartz at stylolites and to a smaller degree at grain contacts."³¹

Apply that to the Tapeats.  Its grains are mostly 0.5–1.0 mm across.⁵  Take 10–20 % of the rock volume as quartz cement and a spherical-grain packing, which maximises the surface area available and therefore minimises the time.  At 110–130 °C the answer is 6 to 94 million years.  In one year you precipitate roughly one part in ten million of it; in Snelling's whole 350-year post-Flood window,⁸ a few thousandths of one per cent.¹²

Two things follow.

First, the rock could not have been cemented in the time the model allows, whether before or after the fold.  Snelling's answer is that Flood waters were chemically exotic — a hypothesis with no measurements attached, applied to an area of hundreds of square miles.

Second, the premise of the whole argument is wrong in the other direction too.  "After ~450 million years it should have been fully cemented and lithified" is the opening sentence of the Carbon Canyon abstract.³  It does not follow.  Quartz cementation is temperature-driven and barely runs below about 80 °C; sandstone that never gets hot stays friable indefinitely.  This is not obscure — it is why deep petroleum reservoirs hold 25–30 % porosity in Jurassic and Cretaceous sandstones, and why sand control is an engineering discipline.  Snelling quotes McKee's own description of the Tapeats: the cement "varies so greatly in amount that some beds are crumbly, whereas others are quartzitic and hard."⁵  Some beds of a 507-million-year-old sandstone are crumbly.  Age alone does not make rock hard.

Two places the story has changed

Arguments should be judged on their best version, but it is fair to note when the best version contradicts the last one.

Bed thickness.  In 2009 Snelling made bed thinning in the hinge the visible proof of softness:

"yet at places within the hinge zone of the monocline, it can be seen that entire layers within the sandstone have thinned as they were stretched during bending.  This is visible confirmation that the sandstone must have still been relatively soft and plastic under the stress of the deformation event."²⁴

In 2023, with the fieldwork done, he reports the opposite and makes constant thickness the proof:

"there is no mechanical crowding or thinning of the sandstone beds in either the hinge zones or the limbs, contrary to the claims of Huntoon (2003) and Hill and Moshier (2009).  Indeed, the thicknesses of the individual sandstone beds are very consistent along their lengths when traced from one limb through the hinge zone and out onto the opposite limb."³

Both observations cannot be right, and the later one is the measured one, so the 2009 claim should be withdrawn.  It has not been.  Creation Ministries International was still running the 2009 version in its magazine in 2021: the rock "becomes thinner where the fold has been in tension (not showing tensional fractures), and thicker where it has been in compression (but not showing signs of being crushed).  This is powerful evidence that the sediment was still wet and plastic."¹⁵

The reversal also matters mechanically, because constant bed thickness through a fold hinge is the signature of flexural slip in coherent layers — the thing the 2023 paper says happened.

Microscope results announced before the microscope.  In 2009 Snelling wrote that "close examination of the sandstone does not reveal any evidence of elongated sand grains, or of broken and recrystallized cement."²⁴  In 2023 he writes that "it appears that no previous investigators have done any thin section investigations of the Tapeats Sandstone, Bright Angel Formation, and Muav Formation to substantiate their claims."³  The thin sections that back the 2009 sentence were cut in 2018.  The claim was published nine years before the evidence for it was collected.  That the evidence, when collected, came out the way he expected does not retroactively make the 2009 statement warranted.

The other folded-rock arguments

Grand Canyon is the flagship, but two other cases are cited often enough to deserve an answer.

Split Mountain, California

Steven Austin's paper at the First International Conference on Creationism in 1986 described tight folds in the Split Mountain Formation along the Split Mountain Fault in the Anza-Borrego desert.¹³  The description is careful and, as far as it goes, convincing.  A concave-upward fault opened a gap; a boulder breccia collapsed into it; the underlying marine sandstone was "tightly folded, overturned, inverted and injected, but rarely broken."¹³

The trouble is Austin's own list of supporting observations.  The first is: "The beds were insufficiently rigid to transfer the stress of impact any great distance away from the fault.  The major deformation is restricted to the 30 feet nearest the fault."¹³  Deformation confined to nine metres beside a collapsing fault gap, in Pliocene sand, in the Salton Trough — one of the fastest-subsiding basins in North America, where several kilometres of sediment accumulated in a few million years — is a textbook syn-depositional slump.  Mainstream geologists describe such features constantly and give them the same name Austin does.  Nothing about it requires a global Flood, and nothing about it generalises to a 507-million-year-old, quartz-cemented Cambrian sandstone.

Austin's conclusion asks: "How could the sediments escape lithification after deep burial over a duration assumed to be 6 million years?"¹³  The answer is the one above.  Cementation is driven by temperature; Pliocene sand a couple of kilometres deep is cool; cool sand does not cement.

The Ute Pass clastic dikes

The same 1986 paper described sandstone dikes along the Ute Pass Fault west of Colorado Springs, where Cambrian sandstone has been injected into Precambrian granite, and argued that the sand must have been unlithified when the Laramide faulting injected it, hundreds of millions of years after deposition.

The same argument is made about the Coconino Sandstone, where John Whitmore and Paul Garner³² attribute sand injectites at the base of the formation to movement on the Bright Angel Fault during the Laramide, and note that this requires the basal Coconino to have been "water-saturated and only partially lithified (at most) during the faulting" — 225 million conventional years after it was deposited.³²  They state the objection in exactly the form Austin does: "The problem for the conventional view is how the Coconino remained unlithified for such a long period of time."³²

The answer is in their own next paragraph, and they supply it themselves: "It is not unusual for clastic dikes, injectites, and sand volcanoes to occur coincident with faulting; fine-grained water-saturated sands are especially mobile."³²  Sand injectites are a large and well-studied subject with a standard reference work behind them, and they form where a porous, water-saturated, under-cemented sand body is over-pressured and breaks its seal.  Their existence across the geological record is evidence that sand does stay mobile at depth, which is the opposite of the premise the argument needs.  Once again, "old therefore hard" is what fails.

Where the young-earth case is strongest

Three things in this dispute genuinely favour Snelling, and they should be said plainly.

The fieldwork exists and the criticism largely does not.  Mitchell and Tillman have not been to these outcrops.  Their figures are line drawings traced over other people's photographs, one of them mirror-imaged.  Snelling's complaint that this is "armchair geology" is fair, and his complaint that annotated tracings can manufacture features is a real methodological objection, not just rhetoric.⁸  Anyone who wants to settle this should go and measure the fold.

The scanning-electron images are the best evidence he has.  Pristine euhedral quartz cement crystals with intact terminations, sitting in open pores, in samples from a fold hinge, do say that the cement has not been crushed since it grew.  The soft-sediment reading of that is not silly.  The competing reading — that a fold requiring under one per cent internal strain, accommodated by slip on bedding and cross-bedding surfaces, would not crush cement in the middle of a pore — is better, but it has to be argued rather than assumed.

Nobody has done the work he says needs doing.  There is no published measured cross-section of the Carbon Canyon fold, no fracture-orientation dataset, no oriented thin-section study.  That is a genuine hole in the conventional literature, and he found it.

Common objections

"Laboratory experiments only run for hours.  You have never watched a rock fold for a million years, so the slow mechanism is an assumption."

Partly true and mostly beside the point.  Nobody watches the folding.  But the mechanisms are observed directly in the rocks: pressure-solution seams and stylolites, indented grain contacts, deformation bands, calcite twins, offset veins.  Those are products, and they are visible in thin section today.  Rates come from experiments extrapolated by well-tested rate laws — the same kind of extrapolation the smectite-to-illite geothermometer depends on, which Snelling accepts when it gives him a temperature.  Rejecting the extrapolation would cost him the 110–130 °C figure that the rest of his argument stands on.

"Both sides interpret the same data through their assumptions.  This is a worldview conflict, not a data conflict."

Sometimes.  Not here.  The illite calculation and the cementation calculation take Snelling's own measurements and his own cited sources, apply published rate laws, and return numbers six to eight orders of magnitude away from what his model needs.  Six orders of magnitude is not an interpretive difference.  If the rate laws are wrong, that is a specific, checkable claim someone can make; but it has not been made, and making it would remove the temperature estimate too.

"The absence of metamorphic minerals still has to be explained."

It is explained, and nobody disputes it.  The Tapeats reached 110–130 °C under about 3.3–4.5 km of rock, on Snelling's own reckoning.  Metamorphic minerals require several hundred degrees.  No conventional geologist has ever claimed the Grand Canyon Palaeozoic strata were metamorphosed.  The argument refutes a position nobody holds.

"Snelling says he found no difference at all between samples in the fold and samples miles away.  How can the fold have deformed the rock if there is no difference?"

This is his best argument and it deserves a straight answer.  The strain the fold requires — under one per cent per bed, less if cross-bed surfaces slipped — is at or below the level a thin section can resolve without oriented samples and quantitative fabric analysis.  Mitchell and Tillman put it well: "The fact that the beds are sinuously folded demonstrates that penetrative strains were involved, although the amount of penetrative strain required is rather small."¹⁷  A comparison that finds nothing at a resolution coarser than the effect is not evidence of no effect.  This is why the fracture-orientation data matter: they detect a small, systematic, geometrically diagnostic signal that a grain-by-grain description cannot.

"What about folds in rocks that everyone agrees were hard — metamorphic rocks, gneisses, slates?  Doesn't the same argument apply?"

Creationists agree those folded while hot.  Snelling: "young-earth creationists agree that at those temperatures the rocks become plastic and are able to be readily deformed by ductile processes, which is also the case for the deformation observed in metamorphic rocks all around the globe."⁸  Creation Ministries International's article carries the same caveat.¹⁵  This is an honest concession, and it costs something: it means "bent rock proves soft sediment" was never a general rule, only a claim about particular rocks.  It also creates a much larger problem elsewhere, since Phanerozoic rocks that were buried, heated to 300–500 °C, folded, and exhumed have to do all of that inside the Flood year.  That is a heat-budget problem, and it belongs to a different article.

"Doesn't the Salton Trough show that clay conversion can be fast?"

It shows the fastest it has ever been observed to be: complete conversion in 20,000 years, at less than 3 km depth, in an active geothermal field, in brine carrying 3,000 ppm potassium.²¹  That is 20,000 times longer than a Flood year, in conditions far more favourable than any the Grand Canyon strata are proposed to have seen.

What would settle it

This dispute is unusually tractable.  Five measurements would close it, and four of them are field work anyone with a permit could do.

1. Fracture orientations across the Carbon Canyon and Monument folds, plotted against the fold axis, and compared with fracture orientations in the same beds outside the folds. If the fracture sets inside the fold are hinge-parallel and denser than outside, they formed during folding in rock that could crack. Both sides agree this is the decisive missing dataset.
2. Calcite deformation twins in the Matkatamiba fold, measured with Burkhard's criteria on oriented thin sections. Snelling has the samples. If the Muav limestone carries thin low-temperature twins recording shortening perpendicular to the fold axis, the limestone was rock when it bent. If it carries none, that is a real result in his favour.
3. Oriented thin sections cut in the three standard planes, so that a shape or crystallographic preferred orientation could be detected if it exists. Snelling's sections were not cut this way; his own criterion of "no preferred orientation" cannot be evaluated without it.
4. A modal count of the quartz cement in the Tapeats, and a measurement of grain surface area, which would replace the estimates in the cementation calculation above with real numbers.
5. A fresh X-ray diffraction run on the two tuff beds, and preferably on more of them. The illite/smectite figures the entire temperature argument rests on are approximate, from two samples, with what Snelling himself calls "likely large errors in the XRD determinations."³ If those numbers changed a great deal, the calculation in this article would change with them — though the conversion would have to be an order of magnitude smaller before the conclusion moved at all.

Conclusion

The folded rocks of the Grand Canyon are a good argument to have chosen.  They are dramatic, they are accessible, and they raise a genuine question about how a stack of layers turns a right angle without falling apart.

The answer is that it does not take much.  A fold like Carbon Canyon asks each bed to stretch by around one per cent, and asks a strongly cross-laminated sandstone for less than that, because it can slip on its own internal surfaces.  Rock a few kilometres down, squeezed slowly, accommodates that by cracking a little, sliding a lot, and dissolving quartz where the grains press hardest.  None of it requires heat, none of it makes metamorphic minerals, and none of it would show up as a bent quartz crystal.

The new technical work does not change that, because it went looking for bent quartz crystals.  Snelling picked as his decisive test the one microstructure that only high-temperature folding produces, in a rock he had already shown never got hot.  The test could not pass, it did not pass, and the failure was reported as a discovery.

And the mineral evidence that was supposed to clinch the case turns out to be the strongest clock in the canyon running the other way.  The illite in those two ash beds took hundreds of thousands to millions of years to grow at the temperature it records.  The quartz cement holding the sandstone together took millions more.  Both numbers come from Snelling's own measurements, his own temperature, and rate laws from the literature he cites.  Both say the same thing: whatever happened in Carbon Canyon, it was not over in a year.

Every source behind this article, with what it measured and which way it cuts, is catalogued in the literature.

Sources