THE INACCURACIES OF THE DATING METHODS

Carbon-14 at the Centre — and the Flaws and Inconsistencies of the Clocks That Date the Earth

1. The Question, and the Ground We Stand On

This study examines the instruments by which the modern world dates the ancient past: carbon-14 first and in depth, then potassium-argon, uranium-lead, the isochron methods, fission track, thermoluminescence, and the counting methods — tree rings, varves, and ice layers. The question is not whether radioactive decay happens; it does, and it can be measured in a laboratory today. The question is whether measurements made in the present can be stretched across an unobserved past, and how much of the answer comes out of the sample and how much goes into it from the assumptions of the man who measures it.

Let the ground be stated plainly at the start, because this study will not be used to prove more than it proves. Scripture gives no date for the creation of the earth. The Bible opens in mid-course:

In the beginning God created the heaven and the earth. And the earth was without form, and void; and darkness was upon the face of the deep. And the Spirit of God moved upon the face of the waters. — Genesis 1:1-2

Between the first verse and the second there is room that the text itself does not measure, and the genealogies of Genesis trace the line of Adam — they were never given as a stopwatch for the rocks. Of God's own eternity the Psalm says:

Before the mountains were brought forth, or ever thou hadst formed the earth and the world, even from everlasting to everlasting, thou art God. — Psalm 90:2

For a thousand years in thy sight are but as yesterday when it is past, and as a watch in the night. — Psalm 90:4

So this study does not argue that the earth is six thousand years old, or any other number of years old. It argues something narrower, and stronger because it is narrower: that the dating methods by which men claim to know the earth's age with dogmatic certainty rest on assumptions about the unobserved past that cannot be proved, that the methods repeatedly contradict known-age tests, that they contradict each other, and that results which contradict expectation are routinely set aside while results which agree are published as proof. A clock with those habits may be a useful instrument. It cannot sit in judgment over the Word of God.

The apostle Peter named the ruling assumption of the whole enterprise nearly two thousand years before the first counter was built:

Knowing this first, that there shall come in the last days scoffers, walking after their own lusts, And saying, Where is the promise of his coming? for since the fathers fell asleep, all things continue as they were from the beginning of the creation. For this they willingly are ignorant of, that by the word of God the heavens were of old, and the earth standing out of the water and in the water: Whereby the world that then was, being overflowed with water, perished: But the heavens and the earth, which are now, by the same word are kept in store, reserved unto fire against the day of judgment and perdition of ungodly men. But, beloved, be not ignorant of this one thing, that one day is with the Lord as a thousand years, and a thousand years as one day. — 2 Peter 3:3-8

"All things continue as they were" — that is uniformitarianism in one sentence, and every dating method in this study is built on it: the assumption that the past behaved as the present behaves, at the rates the present shows, in systems that were never disturbed. Keep that sentence in view. It is the thread that runs through every chapter that follows.

2. How Carbon-14 Dating Works — and the Three Assumptions It Stands On

Carbon-14 is produced in the upper atmosphere, where cosmic rays knock neutrons loose and those neutrons strike nitrogen-14, converting a small fraction of it into carbon-14. The carbon-14 oxidises to carbon dioxide, mixes through the air, and is taken up by every living thing: plants by photosynthesis, animals by eating plants or other animals. While an organism lives, it keeps exchanging carbon with its surroundings, so the ratio of carbon-14 to ordinary carbon-12 in its tissues roughly matches the ratio in the atmosphere. When it dies, the exchange stops. From that moment the carbon-14 in the remains decays back to nitrogen with a half-life of about five and a half thousand years, while the carbon-12 stays put. Measure the ratio that remains, compare it with the ratio in living tissue, and — in principle — you can compute how long the decay has been running.

In principle. The computation is a single equation with a single unknown only if three things are granted before the measurement is ever made:

First, that the starting ratio is known. The equation needs the carbon-14/carbon-12 ratio of the atmosphere when the organism died. Nobody measured it. It is assumed to equal a modern reference value, corrected by a calibration curve — and the curve, as Chapter 5 will show, is built from other dating methods and other assumptions.

Second, that the ratio in the atmosphere has a knowable history. Libby's original method assumed the atmosphere had reached a steady state — production balancing decay — and had held that ratio constant through all the millennia the method would be used on. Chapter 3 examines the record: the atmosphere has never been in that equilibrium, even in the short window we can observe directly.

Third, that the sample has been a closed system since death. No carbon in, no carbon out, except by decay. Chapter 7 will show what a single percent of stray modern carbon does to the arithmetic; Chapters 6 and 8 will show how often real samples fail this condition in ways no laboratory washing can certify away.

Strip the three assumptions down and they are one assumption in three forms: that the past resembled the present in the respects the method requires, and that nothing since has disturbed the evidence. That is Peter's sentence again — all things continue as they were. The chapters that follow test that sentence against the documented record of the method itself.

3. The Atmosphere That Never Stood Still

Libby assumed that carbon-14 production and decay had balanced each other for ages, so that a living thing in any era started its clock from the same ratio. Within his own working lifetime the assumption was already failing in both directions.

The Suess effect. Burning coal and oil pours ancient, carbon-14-free carbon dioxide into the air, diluting the atmospheric ratio. Hans Suess showed in 1955 that wood grown in the twentieth century carried measurably less carbon-14 than wood from before the industrial age — a dilution of roughly two percent by 1950, and larger since. A dilution pushes apparent ages upward: an organism that started with a weaker ratio looks older than it is.

The bomb pulse. Atmospheric nuclear testing did the opposite. By the early 1960s the bombs had roughly doubled the carbon-14 content of the Northern Hemisphere's air; monitoring records (Levin and Kromer, and the long Jungfraujoch and Vermunt series) show the ratio peaking near twice its pre-bomb value around 1963–1964 and declining ever since as the excess soaks into oceans and forests. Dilution in one century, doubling in the next — the supposedly constant starting line moved by factors no archaeologist would tolerate in a ruler.

Sun and magnetism. Production itself is not constant. Carbon-14 is made by cosmic rays, and cosmic rays are deflected by the earth's magnetic field and modulated by the sun's activity. When the sun quiets (as in the Maunder Minimum of AD 1645–1715), more cosmic rays arrive and more carbon-14 is made. When the magnetic field weakens, the same thing happens on a grander scale: during the Laschamp geomagnetic excursion, roughly 41,000 years ago by the conventional chronology, the field fell to a small fraction of its present strength and carbon-14 production spiked — a spike the calibration scientists themselves record in their own curve. The field today is measurably weakening, and its average strength across the last several thousand years has been substantially higher than it once was, which means less carbon-14 was produced per year in some past eras, and samples from those eras start depleted and read too old.

The founder's own check. Libby tested his method against samples of known historical age — the famous "curve of knowns" (Arnold and Libby, Science, 1949). The tests agreed within their generous error bars, but the agreement frayed exactly where it mattered: with the oldest Egyptian material the radiocarbon ages drifted centuries away from the historical ages, and when bristlecone pine — a tree whose rings give a true calendar count — was measured in the 1960s, the drift proved systematic. A ring whose calendar age is 3000 BC returned a radiocarbon age centuries younger, the discrepancy growing with age to roughly 700–800 years by 3000 BC on the modern calibration curve. The plain meaning of that episode deserves emphasis: the method, left to itself, got known-age wood wrong by centuries, and it was rescued not by its own physics but by borrowing a chronology from tree rings. Where the borrowing cannot be done, the error stands uncorrected and unknown.

4. Two Half-Lives for One Clock

A smaller matter, but a telling one. Libby's hal ...[truncated 2096 chars]---

6. Reservoir Effects: Living Things That Date as Ancient

The closed-system and known-starting-ratio assumptions meet their plainest refutation in organisms that are alive and yet date as ancient, because they built their bodies from carbon that was already depleted before they took it up. The mainstream literature discovered and named this itself: the reservoir effect. The cases below are not creationist folklore; each stands in the primary scientific literature, and each mainstream explanation is given with it. What the student should weigh is not whether an explanation exists — one always does — but what the cases prove about the assumptions: the starting ratio in a real organism is a fact about its diet and its waters, not a constant of nature, and for a sample dug out of the ground that fact is usually unknowable.

The mollusks (Keith and Anderson, Science, 1963). In a paper pointedly titled "Radiocarbon Dating: Fictitious Results with Mollusk Shells," M. L. Keith and G. M. Anderson reported that shells of living freshwater mollusks returned radiocarbon ages up to about 2,300 years, and warned that errors of several thousand years were possible for river shells, because the animals incorporate "dead" carbon from dissolved limestone and humus through their food web.

The snails (Riggs, Science, 1984). A. C. Riggs measured the shells of Melanoides tuberculatus snails living in artesian springs in southern Nevada and found carbon-14 contents as low as 3.3 ± 0.2 percent modern — an apparent age of about 27,000 years, in a living animal. The springs are fed by carbonate aquifers; the water carries dissolved limestone, and the snails build from it.

The seal (Antarctic Journal, 1971). A freshly killed seal from McMurdo Sound, Antarctica, was reported in the Antarctic Journal of the United States (vol. 6, 1971, p. 211) with an apparent radiocarbon age commonly cited at about 1,300 years (some secondary sources quote a figure near 1,050 years — the discrepancy between retellings is itself noted in this study's report). Antarctic surface waters are old water, upwelled from depths where carbon has been out of contact with the atmosphere for centuries, and everything that eats from that sea inherits its age. The marine reservoir correction for Antarctic waters runs to roughly a thousand years and varies from place to place and, so far as anyone can tell, from age to age.

Now the fair mainstream answer: reservoir effects are known, they are local, and competent laboratories avoid shell and marine samples where they can, or correct for them where they must. All true — and all beside the deeper point. The correction requires knowing how much old carbon the organism ate, which for an ancient sample is precisely what cannot be known. The living cases are the ones we can catch. The same effect, at unknown strength, in a sample whose diet and waters are lost to us, cannot be caught at all. It will simply be read as age.

7. The Mathematics of Contamination

Radiocarbon's structural weakness is not that contamination happens. It is what tiny amounts of it do to the answer, because the clock reads a ratio and ratios near zero are hostage to traces. Work the arithmetic with the conventional mean life of 8,033 years, using the equation every laboratory uses: age = 8,033 × ln(1/F), where F is the fraction of modern carbon measured.

Take a sample that is truly beyond the method — "infinite age," containing no carbon-14 of its own at all. Add modern carbon, as soil water, rootlets, bacteria, mould, humic acids, handling, or laboratory processing can add it, and read off the age the laboratory will report:

• 1% modern carbon mixed in → the sample reads about 37,000 years (100 half-lives of error bought with one part in a hundred).
• 0.5% modern carbon → about 42,600 years.
• 0.1% modern carbon → about 55,500 years.
• 0.01% modern carbon → about 74,000 years.

The symmetry matters. The same trace that makes an infinitely old sample read 37,000 years makes genuinely old samples read far younger than they are, and a breath of dead carbon — groundwater through limestone, volcanic carbon dioxide, the reservoir effects of the last chapter — makes young samples read old. Contamination is not a rare accident in the published record; it is the standing explanation the literature reaches for whenever a date comes back wrong, which is itself a confession that the closed-system assumption fails in practice whenever the answer is checked against something known. For young samples the absolute error is small and the method serves archaeology well. But the deeper the claimed age, the more completely the answer is manufactured by the last trace of stray carbon — exactly where the method is asked to bear the most weight.

8. Carbon-14 in Coal and Diamonds: The Dispute Stated Honestly

If the conventional timescale is right, coal seams laid down 40–350 million years ago and diamonds formed deeper still should contain no carbon-14 whatsoever — not one atom in a mass of starting material the size of the earth would survive a million and a half years. Yet accelerator mass spectrometry keeps finding some. Here the study presents both sides at full strength, attributed, because this is a live dispute and the student should see it whole.

The creationist measurements (RATE project; Baumgardner, 2005). The RATE team took ten coal samples from the US Department of Energy's Coal Sample Bank at Penn State — samples collected from freshly exposed faces in active mines and stored sealed under argon — spanning Eocene, Cretaceous, and Pennsylvanian seams, and had them measured at a leading AMS laboratory. After subtracting the laboratory background of 0.077 percent modern carbon, all ten still measured between 0.10 and 0.46 percent modern carbon, corresponding to apparent ages of roughly 44,000–57,000 years — and, strikingly, the level showed no trend with the seam's assigned geological age. Baumgardner's chapter also compiled over ninety published AMS measurements from the mainstream radiocarbon journals themselves, on materials all held to be older than 100,000 years, reporting values from 0.014 to 0.71 percent modern carbon. RATE's diamond measurements likewise returned small but detectable carbon-14, at levels equivalent to an apparent age around 55,000 years — in the hardest, most contamination-resistant natural material known.

The mainstream answer (Taylor and Southon, 2007; Bertsche's assessment of RATE for the American Scientific Affiliation). R. E. Taylor and J. Southon measured nine natural Brazilian diamonds and did find small apparent carbon-14 signals — small enough that they proposed natural diamond precisely as a monitor of AMS instrument background: in their analysis the signal arises in the ion source and sample handling, and possibly from carbon-14 manufactured in place when neutrons from uranium and thorium decay in surrounding rock strike nitrogen impurities in the diamond itself. Kirk Bertsche's technical assessment argues the coal results are contamination: coal is porous, sits in groundwater for its whole burial, and can pick up modern carbon in the seam, in storage, and in the laboratory; he notes that with extreme enrichment techniques other workers have measured anthracite at less than 0.01 percent modern carbon — below detection. RATE replies that the uniformity of the signal across unrelated seams, and its persistence after the laboratories' own aggressive pretreatments, is not what random contamination looks like.

The student need not settle the dispute to learn from it. Both sides agree on the datum: at the bottom of the radiocarbon range, the measured signal and the plausible contamination are the same size. That means every radiocarbon date beyond roughly 40,000 years is an argument about background subtraction, not a reading of a clock — and the men who operate the machines are the first to say so when the samples are diamonds.

9. The Ceiling, and Why the Dates Cluster Beneath It

Radiocarbon laboratories rarely report finite ages beyond about 50,000–55,000 years, and this is usually presented as the natural limit of the method — the point where so little carbon-14 remains that none can be counted. Notice, though, what the limit is made of. An AMS machine has a background: stray carbon in the ion source, in the graphitisation lines, in the reagents, memory of the last sample measured. Laboratories measure machine blanks — materials treated as carbon-14-dead — and the blank defines the floor. Samples measuring near the floor are reported as "greater than" ages or not reported at all.

Two consequences follow. First, the ceiling is instrumental, not historical: it marks where the signal drowns in the apparatus, not where the past ends. Second, genuinely ancient samples of every description — coal, shell, bone, diamond — return finite apparent ages clustered in the 30,000–55,000 year band, the very band where a percent or a tenth of a percent of stray carbon would put them (Chapter 7's arithmetic), and where the machine background itself sits. The creationist reads the clustering as a ceiling on the samples' true age. The mainstream reads it as the noise floor of the instrument. Either way, this band of the timescale is not measured history; it is the sound of the method reaching its limit — and a great deal of what the public is told about the antiquity of mankind and his artefacts rests on dates from inside it.

10. When the Same Material Disagrees with Itself

A method earns trust by agreeing with itself: same sample, same answer, whoever measures it. The published record contains enough self-disagreement to keep a student humble.

The mammoths. Troy Péwé's US Geological Survey compilation (Professional Paper 862, 1975) lists radiocarbon results from the Alaskan frozen mammoths that sit awkwardly together: parts of the one Vollosovitch mammoth dated at 29,500 and at 44,000 years; the baby mammoth Dima with parts at about 40,000 and 26,000 years and wood from immediately around the carcass at 9,000–10,000; and the Fairbanks Creek mammoth with a lower leg at 15,380 years and skin and flesh at 21,300. Mainstream reviewers answer — with some justice — that early compilations mixed specimens and materials, that one famous hide had been treated with glycerin, and that later work on Dima produced a consistent cluster near 40,000 years. This study reports the case with that answer attached, and draws only the conclusion both sides can share: in the early decades it was entirely possible for different parts of one animal to be assigned ages differing by fifteen thousand years, and the disagreements were resolved not by the clock but by judgement about which results to trust.

The inter-laboratory exercises. Radiocarbon's own community has tested itself, blind, in the international intercomparison programmes (FIRI, and VIRI reported by Scott and colleagues in Radiocarbon, 2010). Identical materials sent to dozens of laboratories came back with spreads larger than many laboratories' quoted errors would allow, and the organisers' own analyses found statistically significant differences between laboratories on the same material. The differences are usually hundreds of years rather than thousands — this is a mature method for young samples, and no one honest claims otherwise — but they demonstrate the standing gap between the quoted precision of a radiocarbon date and its actual reproducibility. A date is printed as, say, 12,340 ± 40 years. The intercomparisons say the ±40 is the laboratory's hope, not its history.

11. Potassium-Argon and Argon-Argon: The Excess Argon Problem

For rocks too old for carbon, the workhorse has been potassium-argon. Potassium-40 decays with a half-life of about 1.25 billion years; roughly a tenth of its decays produce argon-40, a gas. The textbook picture is clean: molten rock cannot hold gas, so a lava flow starts its clock at zero argon; every argon-40 atom found trapped in the solid rock afterwards was born there by decay, and the ratio of argon to potassium gives the time since solidification. The argon-argon refinement irradiates the sample to convert potassium-39 to argon-39, lets the analyst measure both gases in a mass spectrometer, and can release the gas in heating steps to build an internal "isochron" — but it depends on a standard mineral of supposedly known age irradiated alongside, and those standards' ages come from other radiometric methods, so the argon family calibrates itself against its own cousins.

The zero-argon assumption is the load-bearing one, and it fails in the open literature — beginning with the literature's own side. G. Brent Dalrymple, one of the method's foremost defenders, studied 26 historic lava flows of known age (Earth and Planetary Science Letters, 1969) and found five of them carrying "excess" radiogenic argon — argon that was in the rock when it erupted, making the flows read old. His numbers, as tabulated from his own data in Austin's 1996 review:

• Hualalai basalt, Hawaii (erupted AD 1800–1801): 1.6 ± 0.16 million years, and 1.41 ± 0.08 million years
• Mount Etna basalt, Sicily (122 BC): 0.25 ± 0.08 million years; (AD 1792): 0.35 ± 0.14 million years
• Mount Lassen plagioclase, California (AD 1915): 0.11 ± 0.03 million years
• Sunset Crater basalt, Arizona (AD 1064–1065): 0.27 ± 0.09 and 0.25 ± 0.15 million years

Dalrymple concluded that excess argon is "rare" — one flow in five, by his own table. Other documented young-flow anomalies in the literature run far higher (submarine basalts, quenched under pressure, are the extreme case, returning apparent ages of hundreds of millions of years, since water pressure keeps the gas from escaping at all).

Mount St Helens (Austin, 1996). Creationist geologist Steven Austin took dacite from the new lava dome — rock that solidified in 1986 and was sampled in 1992, ten years old — and sent whole-rock and mineral-concentrate portions to Geochron Laboratories, a commercial dating firm, without disclosing its origin. The whole rock read 0.35 ± 0.05 million years; the mineral concentrates read from 0.34 ± 0.06 million (feldspar–glass) up through 1.7 ± 0.3 million (pyroxene) to 2.8 ± 0.6 million years (pyroxene concentrate) — five portions of one ten-year-old rock disagreeing with each other by nearly an order of magnitude. The mainstream replies are real and should be heard: conventional potassium-argon is not designed for rocks under a few million years (the laboratory's own stated working limit was in that range), and Austin's concentrates contained phenocrysts — crystals that grew in the magma chamber before eruption — which can carry inherited argon. But notice what the reply concedes: phenocryst-bearing lavas are the normal case in nature, not the exception; inherited argon is a physical fact of the method, not a creationist invention; and on a rock of unknown age there is no independent way to see how much of its argon was inherited. The method can only detect its own failure when someone already knows the answer. Similarly, Snelling (1998) reported that lava flows erupted from Mount Ngauruhoe, New Zealand, in 1949, 1954, and 1975 returned potassium-argon ages from 0.27 to 3.5 million years — explained by the same inherited-argon mechanism, which does not cease to operate merely because it has been named.

12. Uranium-Lead: Concordia and Its Escape Hatches

Uranium-lead dating is called the gold standard because it carries two clocks in one mineral: uranium-238 decays to lead-206 (half-life 4.47 billion years) and uranium-235 to lead-207 (half-life 704 million years). Plotted against each other, the two clocks trace a curve called concordia; a zircon crystal that has stayed perfectly closed since it formed must lie on the curve, and the ratio of the two lead isotopes can yield an age even without measuring uranium at all.

The escape hatches are built into the practice. Real zircons very often do not lie on concordia. They fall along a line — a discordia — and the standard interpretation is that the crystal leaked lead at some later disturbance: the line's upper intersection with concordia is then declared the true crystallisation age, and the lower intersection the date of the disturbance. Sometimes the lower intersection is geologically meaningful; often it falls at or near zero (interpreted as recent or ongoing lead loss) or at a time with no known event to attach it to. Either way the construction guarantees a usable answer: points on the curve give an age, points off the curve give an age plus a story. Add the further correction every lead age requires — subtracting "common lead" present at crystallisation using an assumed initial lead composition, commonly the two-stage model of Stacey and Kramers (1975), whose parameters are themselves fitted to the presumed history of the earth — and the gold standard turns out to rest on a model of the very history it is used to establish. Inherited zircon cores, older crystals swept up by younger magma, are likewise documented in the literature and likewise handled by choosing which part of the crystal to believe. None of this means uranium-lead dates are random. It means their celebrated precision is precision within a set of assumptions about closed systems and initial compositions that the crystals themselves frequently testify against.

13. Isochrons: Rubidium-Strontium, Samarium-Neodymium, and the Mixing Line

The isochron methods were invented to escape the initial-daughter problem. Instead of assuming how much daughter isotope the rock started with, the analyst measures several samples from the same rock unit with different parent/daughter ratios, plots daughter ratio against parent ratio, and reads the age from the slope of the line; the intercept gives the initial ratio as a bonus. Rubidium-strontium (half-life about 48.8 billion years) and samarium-neodymium (half-life about 106 billion years) are the workhorses.

The method is elegant, and it has one structural weakness that the textbooks themselves acknowledge: a straight line on an isochron diagram is not uniquely a clock. If two materials with different isotopic compositions mix — two magmas blending, magma assimilating wall rock, a sediment averaging two source terrains — the mixture's samples also fall on a straight line in the same diagram, because mixing between two end-members is mathematically linear in these ratios. Such a "pseudo-isochron" has a slope that reflects the compositions of the end-members, not any lapse of time, yet it yields a tidy "age" with an impressive correlation coefficient. The isochron's other requirements are equally unearned from the diagram itself: all samples must have started with the same initial daughter ratio (homogenised at formation) and must have remained closed ever since. Metamorphism, fluid flow, and weathering all violate closure, and the standard response — as with lead loss — is to read the line that survives as the age and the disturbance as the explanation for any point that spoils it. Samarium-neodymium model ages add a further assumption: they are calculated against an assumed composition for the earth's primitive mantle (the "chondritic uniform reservoir"), so the age of a crustal rock is in part the age of a model of the mantle it is presumed to have melted from.

14. The Grand Canyon Test Case: One Canyon, Four Clocks, Four Ages

Nowhere is the discordance of the methods on full display like the Grand Canyon, where creationist and mainstream teams have dated the same walls for decades.

The Cardenas Basalt. This lava unit sits low in the canyon's strata. Its published rubidium-strontium whole-rock isochron age is 1.07 ± 0.07 billion years (Elston and McKee, 1982) — long quoted as the unit's definitive age. Austin and Snelling (1998) assembled the potassium-argon data — published analyses plus thirteen new ones — and the potassium-argon isochron came out at 516 ± 30 million years: the same rock, half the age, by a different clock. The mainstream resolution is that fluids associated with the overlying Chuar Group disturbed the potassium-argon system; current reference works settle the unit near 1,104 million years on newer work. That may be so. But observe the order of operations: the rubidium-strontium number was believed first, and the potassium-argon number was explained because it disagreed.

The Brahma amphibolites (RATE, 2005). The RATE project submitted samples of one metamorphosed lava unit in the canyon's basement to two commercial laboratories for four independent methods. The isochron results for a single rock unit: rubidium-strontium 1,240 ± 84 million years; samarium-neodymium 1,655 ± 40 million years; lead-lead 1,883 ± 53 million years — a spread of over 600 million years between methods, each with small quoted errors — while the potassium-argon model ages for samples of the unit ranged from 405 million to 2,574 million years, including two samples taken less than a metre apart that differed by more than a billion years. Mainstream geologists attribute the spread to the unit's violent metamorphic history — resetting is exactly what metamorphism would do. The creationist team answers that a clock which can be reset by heat and fluids, in rocks whose thermal history is inferred largely from the clocks themselves, cannot then be cited as independent proof of the timescale. The student can weigh both readings; what cannot be denied is the raw fact that four celebrated methods, run on one rock unit by professional laboratories, returned four irreconcilable answers.

The Uinkaret lavas. On the canyon's western rim sit young basalt flows that poured over the rim and dammed the Colorado River — geologically recent, erupted into a canyon already carved. Their potassium-argon model ages scatter widely, as young lavas' do; but their rubidium-strontium "isochron" age, as compiled by Austin (1994), is 1,143 ± 220 million years — a billion-year age, statistically indistinguishable from the Cardenas Basalt's, for lava that flowed into a canyon that did not then exist if the age were real. The mainstream explanation is that the line is a mixing or source-inheritance artefact, not an age. Perhaps it is. But that answer, granted here, must be granted everywhere: a straight line on the diagram does not certify an age, and when the line gives a billion years for a young lava the line — not the lava — is discarded. The diagram cannot be proof when it agrees and artefact when it does not.

15. Fission Track, Thermoluminescence, and the Counting Methods

Three shorter treatments, each with the same anatomy: a sound physical effect, an assumption about the unobserved past, and a calibration borrowed from the radiometric system it is supposed to check.

Fission track. Uranium-238 occasionally splits by spontaneous fission, blasting a damage trail through the surrounding crystal. Count the fossil tracks under a microscope (after acid etching), measure the uranium (by inducing fresh fissions in a reactor), and the ratio gives an age. But tracks fade — anneal — when the crystal is warmed, and partial annealing over a burial history shortens and erases them, biasing ages young by amounts that depend on the very thermal history the dating is meant to discover. More telling still is the calibration: the practical constant of the method, zeta, is not measured from physics but determined empirically against standard samples whose ages are assigned by potassium-argon and argon-argon dating (the Hurford and Green zeta approach, 1983). Fission track therefore does not independently confirm the argon timescale; a share of its agreement was built in at the factory.

Thermoluminescence (and its cousin, optically stimulated luminescence). Minerals accumulate trapped electrons from environmental radiation; heating or sunlight empties the traps with a flash of light proportional to the total dose received since the traps were last emptied. The age is the dose divided by the dose rate — and the rate must be assumed across the whole burial: soil moisture, which shields radiation, comes and goes with climate; the radioactivity of the surroundings changes as groundwater moves uranium in and out; and feldspars suffer "anomalous fading," losing signal with time for reasons the method corrects by estimate. The result is an age whose numerator is measured and whose denominator is substantially a guess about ten thousand unobserved years of weather.

The counting methods. Tree-ring dating is the soundest chronology in this study within its true range — a genuinely annual record, cross-matched between trees — and honesty requires saying so. Its limits are equally real: the continuous master chronologies reach back roughly twelve to fourteen thousand years even on the conventional account; beyond them lie floating chronologies, sequences of rings whose internal order is sure but whose placement in time is fixed by radiocarbon wiggle-matching — that is, by the very method they are then used to calibrate. Cross-matching itself involves judgement: missing rings, double rings, and drought-deformed sequences occur, and candidate matches are accepted or rejected by statistical thresholds a researcher chooses. Varves — annual sediment layers in lakes — carry the same structure further: the Lake Suigetsu varve count underpins the calibration curve beyond the tree rings, yet the count's own anchoring has been revised between publications, and its link to absolute time runs partly through the radiocarbon it calibrates. Ice-core layer counting (the Greenland GICC05 timescale) is candidly published with a "maximum counting error" that grows with depth into the thousands of years — an official admission that the deepest "annual" layers are identified by judgement among ambiguous strata, with flow-thinning models filling the gaps. Counting is honest work while the layers are unambiguous. It is not an independent witness when its far end is tied, by construction, to the clocks it is called in to correct.

16. The Meta-Pattern: A Circle of Clocks, and the Fate of Discordant Dates

Step back from the individual methods and watch the system as a whole. No dating method in this study stands alone. Radiocarbon is corrected by tree rings; tree rings beyond their continuous range are placed by radiocarbon. The far calibration curve is built from uranium-thorium-dated corals and speleothems; uranium series and uranium-lead share the lead-correction models. Argon-argon is calibrated on standards dated by potassium-argon and astronomical tuning; fission track's zeta is calibrated on argon dates; ice-core timescales are synchronised to the radiocarbon-calibrated records by matching volcanic ash and gas signals between them. Each clock's agreement with the others is therefore partly an achievement and partly a construction. The system is a circle of clocks, each set by its neighbours, and the circle as a whole is anchored to history only at its near edge — the last few thousand years, where written records and true tree-ring counts can check it, and where (as Chapter 3 showed) the checking found errors of centuries that had to be corrected by hand.

Within such a system, what happens to a result that disagrees? The literature's own vocabulary answers: the sample was contaminated; the system was disturbed, reset, partially reset; the rock suffered argon loss, or carried excess argon; the zircons inherited old cores; the line is a mixing line, not an isochron; the horizon was reworked. Each of these is a real physical phenomenon, documented in particular cases. Together they form a complete set of escape hatches: there is no conceivable measurement that the vocabulary cannot explain away, and a system that can explain every outcome can be confirmed by none of them. The history of the KBS Tuff, the ash layer above the famous Koobi Fora fossils in Kenya, shows the selection process running in the open. Fitch and Miller's first potassium-argon result (1970) put the tuff at roughly 212–230 million years. The fossils beneath it said that was impossible, and the date was rejected as excess argon contamination. Fresh samples gave 2.61 million years, which was accepted and independently "confirmed." Then the skull KNM-ER 1470 was found beneath the tuff looking too advanced for 2.61 million years, fossil pig correlations pointed younger, the tuff was redated to about 1.8–1.9 million years, and the confirming methods obligingly confirmed the new figure too. Mainstream textbook authors Prothero and Schwab (1996) describe the episode without embarrassment: the dates "had to be redone many times in several laboratories by three methods over 15 years before all the results were consistent and undisputed." Read that sentence twice. The criterion of success was not that a method produced the same answer repeatedly; it was that the answers eventually became consistent — consistent, in context, with what the fossils were believed to require. Fitch and colleagues (1985) state the working rule from inside the profession: geological error "is evident when repeat samples from the same stratigraphical horizon don't produce consistent ages, and when the ages obtained from a given horizon are incompatible with those from strata immediately above and below." Which is to say: a date is judged wrong by its disagreement with its neighbours, in a system whose neighbours were all set within the same circle.

This is not an accusation of fraud. It is a description of how the enterprise works when it works as designed — and it caps what the enterprise can ever claim. Radiometric dating is a method for assigning numbers inside a framework. It is not, and by its own operating rules cannot be, an independent proof of the framework.

17. Conclusion: What the Evidence Will Bear

Gather the threads. Carbon-14, the most honest and most checkable of the methods, is built on a starting ratio that has demonstrably changed within living memory (diluted by industry, doubled by bombs, swung by sun and magnetic field across the older record); it is computed with a half-life known to be wrong by three percent, by convention; it must be corrected by a calibration curve assembled from other dating methods; caught in the act, it dates living snails at 27,000 years, living mollusks at 2,300, and a fresh-killed seal at over a thousand, whenever an organism's carbon came from old water; its arithmetic hands a 37,000-year age to a pinch of modern contamination in an ageless sample; and at the deep end of its range its signal cannot be separated from the background of the machine that measures it. The long-range methods fare no better under examination: argon that was supposed to start at zero is found, by the method's own champions, in one historic lava flow in five; one Grand Canyon rock unit returns four irreconcilable ages from four methods; an "isochron" certifies a billion years for lava that flowed into a canyon younger than the date; and the whole family of methods is calibrated in a circle and defended by a vocabulary that converts every disagreement into a disturbance. None of this shows that the earth is young, and this study has not claimed that it does. Scripture does not date the earth, and the man who honours Scripture will not pretend otherwise — "In the beginning God created the heaven and the earth" is a statement about the Author, not about the year.

What the evidence does show is that the certainty is unwarranted. The dating methods are instruments of real but limited power, reliable in proportion as their results can be checked against known history, and speculative in proportion as they leave that check behind. A dogma built on them — the dogma that sneers at Genesis from the security of "settled science" — is built past its evidence. Peter's prophecy stands over the whole subject: the scoffer's case rests on the claim that all things continue as they were from the beginning of the creation, and the laboratory record of the dating methods is, in the end, a two-hundred-year footnote discovering that they do not. Let the clocks be used for what they can do. Let them not be worshipped for what they cannot.

But, beloved, be not ignorant of this one thing, that one day is with the Lord as a thousand years, and a thousand years as one day. — 2 Peter 3:8