A Date With No Ship Attached

Archaeologists date ancient shipwrecks by dating the cargo, the fittings and the seabed deposit that survived the sinking, because the ship itself almost never survives to be dated. Wood floats, rots and is eaten by worms in warm water, so the vessel that gives a wreck its name is usually the one object that cannot be recovered, measured or sampled. The date attached to a wreck therefore comes from jars, coins, ingots, anchors and timbers interpreted as a group, and its reliability depends on one prior question that excavation must answer first: did these objects actually travel together, and did they arrive on the seabed in a single event? A wreck is dated by its youngest trustworthy object, never by the ship itself, and the strength of the date is the strength of the argument that the objects were together when the vessel went down.

That argument sounds like a qualification, yet it is the entire subject. Every famous wreck date in the archaeological literature is a compressed version of an argument about association, built from the positions, conditions and kinds of the finds. When the argument is strong, when a hull lies under a sealed mound of cargo and several independent clocks agree, the date can be stated as a narrow range and defended for decades. When the argument is weak, when a scatter of pots on a slope is all that remains, the honest date is a wide range attached to a careful statement of what the evidence cannot yet establish. Readers who learn to see the argument inside the date can judge any wreck claim, including claims that arrive dressed as records.

The Mediterranean, where most ancient wreck archaeology has been practiced, is both generous and unkind to evidence. Its warm, oxygenated shallows destroy unprotected wood within centuries through the action of the shipworm Teredo navalis and gribbles, small crustaceans that graze timber surfaces. Its deep anoxic basins, by contrast, can preserve hulls almost intact, which is why the wrecks of the Black Sea, where oxygen vanishes below a certain depth, have transformed expectations of survival. Between those extremes sits the standard case: a cargo of fired clay, stone and metal that outlasts its carrier by thousands of years, arranged on a slope or a reef in a pattern that preserves information even after the ship has gone. The discipline of nautical archaeology exists to extract that information without inventing it, and dating is where its standards are tested hardest.

When the Hull Is Gone, the Wreck Must Be Proven

A seabed scatter becomes a shipwreck through evidence, not through announcement. Excavators look for a cluster of objects that could not have arrived together by any process other than a vessel sinking, and each kind of object contributes a different kind of proof. A concentration of storage jars of one fabric and one period, spread in an elongated pattern that follows a slope, suggests a hull outline even where no timber remains. Stone anchors lying at one end of the pattern strengthen the reading, because anchors are carried at the bow and stern of a working ship and are lost or abandoned in configurations that a random dumping of cargo rarely reproduces. Ballast stones, galley wares, personal possessions such as lamps and balance weights, and the sheer weight of metal cargo all add to the case, since a merchant crew’s kitchen equipment and a captain’s scales are not cargo and would only reach the seabed with a ship.

The alternative explanations must be excluded one by one, and the exclusion is part of the date. Cargo jettisoned in a storm leaves jars without anchors and without a galley. A harbor dump leaves mixed material of many periods in a confined area. A later wreck can deposit its cargo on top of an earlier one, producing a layered site whose objects span centuries and whose youngest items date only the upper event. Fishermen’s nets drag objects across sites and gather them into false clusters. Each of these processes is documented, and an excavation report that does not address them has not established a wreck, whatever its press coverage says. The standard is the one applied on land, where an archaeologist who called a plowed scatter of pottery a building would be expected to show foundations, and the sea does not lower it.

Scale matters to the proof as well. A ten metre trading vessel of the Early Bronze Age carried cargo measured in dozens of jars, not hundreds, so a modest scatter can represent a whole shipload. A Roman grain freighter carried thousands of amphorae, and its wreck forms a mound visible in survey sonar. Judging whether a scatter is complete, partial or illusory requires knowing what a complete example of its class looks like, which is why the history of excavated wrecks functions as the reference collection for every new claim. The excavated benchmarks discussed below, at Dokos, Cape Gelidonya, Uluburun and Kyrenia, are not merely famous sites. They are the calibration set against which the completeness of every later scatter is judged.

The Seabed Is an Editor: How Wreck Sites Form

Keith Muckelroy, whose 1978 book Maritime Archaeology gave the young discipline its first general theory, argued that a wreck site is not a frozen moment but the surviving fraction of a system, filtered twice. The first filter is the wrecking process itself: a ship strikes, breaks up in a sequence governed by its construction and the energy of the sea, and distributes its contents according to weight, buoyancy and the shape of the seabed. Buoyant timbers drift and strand elsewhere. Heavy ingots sink where the hull opens. Jars roll downhill until friction and sediment stop them. The second filter is everything that happens afterward across the centuries: storms that rework the deposit, currents that sort small objects away, marine organisms that bore, encrust and cement, and human salvage, ancient and modern, that removes the most valuable material first. Muckelroy called these extracting and scrambling filters, and the terms have stayed in use because they name what every excavator observes.

The consequence for dating is direct. A wreck assemblage is a biased sample of the ship’s contents, biased in knowable directions. Metal survives and ceramic survives, while textiles, grain, timber and most food vanish unless sealed inside a jar or buried under sediment quickly enough to exclude oxygen. The youngest object principle, which governs wreck dating, must therefore be applied to a sample whose completeness is uncertain. If the late material was preferentially removed by salvage or scattered by a storm, the excavated date can err on the early side; if later material intruded from a second wrecking or from dumping, it can err late. A careful report states which filters are visible in the site and adjusts its confidence accordingly, rather than presenting the date as if the seabed had kept perfect minutes.

Concretion is the seabed’s own form of editing, and it cuts both ways. Iron objects corrode into bulky concretions that swallow neighboring finds, cementing a coin to a jar or a tool to an ingot in a single mineral mass. The cementing destroys surfaces and complicates recovery, since the mass must be X rayed before it is opened, but it also proves association with a rigor no argument can match: two objects inside one concretion lay together while the minerals formed around them. Excavators have learned to read concretion as a witness rather than as damage, mapping which finds it joins before any of them are lifted, and a dating claim supported by concreted groups carries a weight that loose distribution alone cannot supply.

Muckelroy also supplied the discipline’s most useful distinction, between the ship as a machine for transport, the crew as a social unit, and the cargo as an economic cargo system, each leaving different material traces. A date built only from cargo says when the cargo was made and shipped, and the other two systems check it: galley pottery dates the crew’s equipment, hull construction dates the shipbuilding tradition, and personal seals and weights date the merchants. Where all three systems point to the same generation, association is strong. Where they diverge, the excavator has found either a long lived ship carrying old equipment, which is common and interesting, or two events merged into one site, which is a dating failure that only honesty can salvage. Divergence is information, provided it is reported rather than smoothed away.

The Youngest Object Sets the Date

The logic that converts objects into a date is borrowed from stratigraphy and is stated in Latin because generations of archaeologists have found the Latin harder to misuse: terminus post quem, the time after which. A deposit cannot have formed before its youngest component was made. A wreck containing a coin struck in a named ruler’s tenth year cannot have sunk before that year, however ancient its other cargo may be. A wreck containing amphorae of a shape first produced in a given generation cannot predate that generation. The date of the sinking is therefore a floor built from the newest datable object, combined with an argument about how much later than that floor the sinking plausibly occurred. Amphorae were reused for years, coins circulated for decades, and ships themselves served for a generation or more, so the floor is a beginning, not an answer.

Against the floor stands a second concept, the closed context. A group of objects sealed together by a rapid event, undisturbed since, can be treated as a single moment for dating purposes, and a capsized hull collapsing onto its cargo on a soft seabed is one of the best closed contexts archaeology possesses. The celebrated land parallel is Pompeii, where a single eruption sealed households in 79 CE, a comparison the archive explores in its account of the destruction of Pompeii by Vesuvius. Underwater, the seal is sediment rather than ash, and its integrity must be demonstrated by excavation rather than assumed: concretions that bridge two objects prove they lay together while the concretion formed, and a cargo mound with an unbroken internal order, heavy goods low and fragile goods protected between them, is difficult to explain except as a hull’s contents settling in place.

Dating therefore proceeds by accumulation rather than by a single instrument. Each method contributes a constraint, and the published date is the intersection of the constraints that survive scrutiny. A typological range for the jars, a radiocarbon range for a timber, a coin’s issue year and a stamped handle’s named magistrate do not average into a date; they overlap into one, and the overlap is narrower and more honest than any of its parts. Where the constraints fail to overlap, the site is telling its excavators that the association is wrong, and the correct response is to split the site, not to widen the error bars until the contradiction disappears. This is why experienced wreck archaeologists read a suspiciously precise date as a warning and a frankly stated range as a credential.

Ceramics as Clocks: Typology and Seriation

Fired clay is the wreck archaeologist’s most faithful servant. It is nearly indestructible in seawater, it was produced in enormous quantities, and its shapes changed constantly under pressure from fashion, function, taxation and the mechanics of stacking and pouring. A potter’s choices about rim, neck, handles, body and base were never arbitrary: they solved problems of filling, sealing, carrying, standing in sand or in a rack, and being recognized at a market. Because the solutions changed at rates measured in decades, a trained eye can place an amphora within a generation or two from its silhouette alone, and can often place its production region from its clay. Typology, the classification of these changing forms, is the oldest dating instrument in the wreck toolkit and still the first one applied on any new site.

Why do broken pots make better clocks than gold?

Broken pottery makes a better clock than gold because pottery was cheap, abundant and constantly redesigned, while gold objects were rare and kept for generations. One rim fragment can place a deposit within decades, and sherds survive in quantities that allow comparison, whereas a precious object may be an heirloom far older than its ship.

The method behind typology is seriation, placed on a scientific footing by Flinders Petrie in his 1899 study of the cemeteries at Diospolis Parva in Egypt. Petrie ordered graves by the gradual appearance, flourishing and disappearance of pottery types, producing a relative sequence that later absolute dates could anchor. The same logic orders wreck cargoes. Shapes have careers: a form appears, spreads, dominates, coarsens and vanishes, and a cargo caught mid career can be located along the curve. Heinrich Dressel gave Roman amphora studies their enduring classification in 1899, numbering forms whose production spans are still refined by excavated kiln sites and dated deposits on land. Underwater assemblages borrow their chronology from these land sequences, which is why a wreck date is only as strong as the land archaeology behind its pottery, a dependency that careful reports state openly.

Ceramic dating also carries a built in check against the heirloom problem that weakens every other material. Cargo pottery was packaging, made to be sold with its contents or discarded, and a merchant did not ship wine in jars a century old when new jars were cheaper than the risk of leakage. Galley pottery, the crew’s own cooking and serving wares, skews older and is interpreted with that caution. Where cargo and galley wares agree, the date firms; where the cargo is uniform and the galley mixed, the pattern matches a working ship’s life and still dates the sinking by its cargo. The Kekova claim examined below rests entirely on this instrument, on amphora shapes matched to dated parallels in Egypt and the Levant, and its strength and its limits are the strength and limits of typology itself.

Kiln sites give typology its floor of certainty. A kiln abandoned with its final load inside, or a waster heap of misfired vessels discarded at the production site, ties a form to a place and, where the kiln’s own context is dated, to a span of years. Production spans matter as much as shapes: a form made for three centuries at a conservative workshop cannot date a wreck within a generation however crisp its profile, and the honest typologist reports the span alongside the silhouette. The great reference collections, built from excavated kilns, harbor dumps and dated graves across the Mediterranean, exist so that a form’s full career, birth, spread, dominance and disappearance, can be quoted rather than its most convenient decade. Wreck dates inherit those careers entire, uncertainties included, and a report that quotes a tight range from a long career form has spent confidence it never held.

Amphorae, the Barcodes of the Ancient Sea

The amphora deserves its own section because it is three dating instruments in one object. Its shape gives a typological range. Its clay gives a production region, identified by the mineral inclusions visible in a broken edge and confirmed by laboratory analysis of the fabric. Its stamps, where the producing state required them, can give a year. The island of Rhodes operated the most informative stamping system of the Hellenistic world: handles were impressed before firing with the names of an annual magistrate, the eponym, and of the fabricant who made the jar, so a stamped Rhodian handle found on a wreck names a specific year of manufacture. Virginia Grace, the American scholar who spent a career decoding these stamps at the Athenian Agora, built the reference chronology that wreck excavators still use, correlating stamp sequences with dated deposits until the eponyms could be placed in order and, at many points, in absolute years.

Stamps also expose the difference between dating a jar and dating a sinking. A jar stamped in a magistrate’s year was filled, shipped, emptied and refilled across a working life that could span decades, and amphorae turn up in secondary use as building material, drainage pipes and coffins. The wreck context disciplines the stamp: a hold full of jars stamped within a narrow band of years, uniform in fabric and filled for a single voyage, dates that voyage tightly, while a single stamped handle in a mixed scatter dates almost nothing. Excavators therefore report stamp evidence as a population, with counts, rather than as a trophy find. A report that quotes one glamorous object and is silent about the population has inverted the evidential order, and readers are entitled to discount its date accordingly.

The economic geography written into amphorae supports the association argument as well. A coherent cargo, jars of one region carrying one commodity toward a market that consumed it, reads as a shipload; a miscellany reads as a harbor. The licensed Greek trading post at Naucratis in the Nile Delta, examined in the archive’s account of Naucratis and Egyptian trade with Greece, shows the consuming end of such routes, where Aegean jars arrived in quantity and their fabrics and stamps recorded the traffic. Wreck cargoes and port deposits are the two halves of one archive, and each dates the other: ports supply the dated parallels that order the shapes, and wrecks supply sealed cargoes that test whether the port sequences hold at sea.

Radiocarbon: A Clock Inside the Wood

Radiocarbon dating measures the decay of carbon 14, an unstable isotope created in the upper atmosphere and absorbed by every living thing. While an organism lives, it exchanges carbon with its environment and holds the atmospheric proportion of carbon 14; when it dies, exchange stops and the isotope decays with a half life of about 5,730 years, so the surviving proportion measures the time since death. Willard Libby published the method in 1949 and received the Nobel Prize in Chemistry in 1960, and its arrival transformed wreck archaeology wherever organic material survives: hull timbers, dunnage brushwood, rope, basketry, food remains and the wooden plugs and stoppers of jars all carry the clock. Accelerator mass spectrometry, which counts atoms directly rather than waiting for decays, reduced the required sample to milligrams and made it possible to measure a single seed or a fragment of cordage.

Why does charcoal date a fire but not necessarily a voyage?

Charcoal registers the year its wood stopped exchanging carbon with the atmosphere, the year the tree tissue grew, not the year it burned or sailed. Heartwood from an old tree, or reused timber, carries a date generations older than the ship, so excavators sample outer rings, twigs and short lived materials tied to the voyage itself.

The discipline imposed by that answer is called the old wood problem, and wreck reports manage it by sample selection. A hull plank cut from a mature oak contains rings formed across two centuries, and a bulk sample averages them into a date older than the felling, let alone the sinking. Short lived samples, twigs, leaves, seeds, resins and food remains, grew in a single season close to the voyage and date it within the method’s precision. The Uluburun excavation set the standard by dating multiple sample types from one sealed cargo and showing that the short lived materials clustered where the typology said they should. Calibration then converts the radiocarbon age into calendar years, because atmospheric carbon 14 production has varied with solar activity and the Earth’s magnetic field. Tree ring sequences of known age provide the calibration curve, maintained internationally and revised as IntCal20 in 2020, and a single radiocarbon age can intersect that wavy curve at several calendar points, producing ranges that honesty requires excavators to publish in full.

A refinement called wiggle matching strengthens the method where a timber preserves a run of rings. Each ring’s radiocarbon age must follow the calibration curve’s shape in sequence, so a series of measurements from consecutive rings can be fitted to the curve like a key in a lock, narrowing the felling date to a span far tighter than any single measurement allows. Applied to wreck timbers and to the brushwood dunnage packed between ingots, wiggle matching has produced some of the most precise prehistoric dates in archaeology. Its condition is association once more: the rings must belong to wood that was alive near the sinking, which is why excavators prize the outermost rings and the bark edge above all other samples.

The Marine Reservoir Problem

Ocean water complicates the clock in a way that land archaeology never faces. Surface ocean water exchanges carbon with the atmosphere slowly, and deep water that wells up in some regions has been isolated from the air for centuries, so marine organisms build their shells and bones from carbon that is already old. A shell or a fish bone from a wreck therefore returns a radiocarbon age older than its death, by an offset called the marine reservoir effect. The global average offset for surface waters is roughly four centuries, but the figure is an average of a variable quantity: upwelling zones, enclosed seas and estuaries each carry their own correction, expressed by specialists as a local deviation added to the global curve. The Mediterranean, an enclosed sea with limited exchange through the Strait of Gibraltar, has its own regional values, measured from shells of known collection date gathered before nuclear testing altered the atmosphere.

For wreck dating, the reservoir effect is less a trap than a rule of sample discipline. Hull timber, cordage and terrestrial food remains grew on land and follow the atmospheric curve, so they are preferred. Marine samples enter a wreck date only with their correction attached, and a report that dates a sinking from uncorrected shell or fish bone has built an error of centuries into its foundation. The effect also explains a recurring pattern in the literature, where dates from a ship’s stores of grain agree with the pottery while dates from its ballast shells drift early, and the drift itself becomes evidence that the samples have been correctly understood. A clock that fails predictably, once its bias is measured, remains a clock; the failures that damage wreck chronology are the unmeasured ones.

There is a converse benefit. Because the reservoir offset is a property of the water mass, paired samples from one context, a terrestrial seed and a marine shell sealed in the same deposit, measure the local offset at the moment of sinking and contribute to the regional correction tables that future excavations will use. Wreck sites, precisely because their moments are fixed by other clocks, calibrate the sea itself. The traffic between methods is two way throughout this subject: pottery dates the wreck, the wreck dates the pottery sequence, radiocarbon checks both, and the paired samples refine the marine curve that the next wreck will need.

Tree Rings and the Uluburun Dendro Date

Dendrochronology, the dating of wood by its annual growth rings, offers the precision radiocarbon cannot reach, because a ring sequence matched against a master chronology names the calendar year in which each ring formed. A. E. Douglass, an astronomer studying sunspots and climate, built the first master sequences in the American Southwest in the early twentieth century by overlapping patterns of wide and narrow rings from living trees, beams and older timbers. The method requires a regional master chronology, a long reference sequence for the species and climate zone, and the Aegean and Anatolian chronology built by Peter Kuniholm and the Cornell Aegean Dendrochronology Project supplied that reference for the eastern Mediterranean Bronze Age. Where a wreck timber preserves enough rings and, ideally, its bark edge, dendrochronology can state the felling year, and a felling year for repair timbers or dunnage cut shortly before a voyage stands very close to the sinking.

The Uluburun wreck, excavated off the Turkish coast near Kaş between 1984 and 1994, received the method’s most famous wreck application. Brushwood and firewood carried aboard, short lived material cut for the voyage, preserved ring sequences that Kuniholm’s laboratory matched to the Anatolian master chronology, placing the felling of the youngest wood in or near 1305 BCE. Because dunnage is cut green and used once, its felling date crowds the sinking date closely, and the result agreed with the radiocarbon wiggle match and with the pottery and scarab evidence from the same sealed cargo. Three independent clocks converging on one generation is the strongest dating position prehistoric archaeology can occupy, and Uluburun is cited across the discipline precisely because the convergence was demonstrated rather than asserted.

Dendrochronology’s limits are as instructive as its triumphs. Many wreck timbers preserve too few rings, or rings complacent from a sheltered growth site that match nowhere. Floating chronologies, sequences not yet anchored to calendar years, can date timbers relative to each other without naming a year. Salt water and marine borers attack exactly the outer rings that matter most. The method therefore appears in wreck reports as a crown witness when conditions favor it and as an absence to be explained when they do not, and an excavator who claims a dendro date owes the reader the ring count, the master chronology used and the statistical strength of the match.

The harbor works at Istanbul supplied the method’s largest single test. Excavation of the Theodosian harbor at Yenikapı, undertaken between 2004 and 2013 ahead of railway construction, recovered 37 Byzantine vessels, from fifth century coasters to tenth century galleys, buried in the silted basin where they had sunk at their moorings or been abandoned. Timbers from the fleet carried ring sequences that the Cornell laboratory and Turkish colleagues matched into the master chronology, dating construction and repair episodes across five centuries of one port’s life and cross checking the coin and pottery sequences from the same layers. A harbor that buries its own ships in order is a stratigraphy lesson in itself, and Yenikapı’s fleet demonstrated that dendrochronology scales from a single wreck’s dunnage to the biography of a port.

Coins and Inscriptions: Years You Can Read

A coin is a dated document struck by a state. Its issuing authority, its type and often its magistrate or regnal year place its minting within a narrow span, and because coins were struck in quantity and studied in depth, numismatists can usually say which span. For the wreck excavator, a coin assemblage supplies the hardest terminus post quem available: the sinking cannot precede the youngest coin, and a purse or strongbox of coins lost together reflects the currency of its moment with a fidelity no other artifact matches. The Antikythera wreck, examined below, was anchored by coins from Pergamon and Ephesus struck in the decades before the sinking, and the Kyrenia ship’s few coins similarly disciplined a date built mainly from pottery and hull carpentry.

Inscriptions extend the same gift beyond money. A bowl inscribed with a date in a reign or an era converts a cargo into a document, as the Belitung wreck demonstrates, where a Changsha ware bowl bearing an inscription corresponding to 826 CE fixed a ninth century dhow’s cargo within a reign period. Stamped amphora handles are inscriptions of a bureaucratic kind, and lead seals, merchants’ marks and inscribed ingots contribute wherever they survive. The evidentiary condition is constant: the inscription dates the object’s manufacture or issue, and the step from object to sinking remains an argument about circulation and use. Coins were hoarded, inscriptions were carved on heirlooms, and a single old coin in a young wreck is a commonplace of the literature. Assemblages date; singletons decorate.

There is also a negative use, quieter and just as binding. The absence of coin types that flooded a region after a known reform can argue that a wreck predates the reform, provided the cargo’s origin makes their absence surprising. Arguments from absence are weaker than arguments from presence and are presented as such, with the size of the assemblage stated, since a small sample’s silence proves little. In combination, presence sets the floor, absence suggests the ceiling, and the space between them is the date range the evidence supports. That discipline, floor and suggested ceiling rather than a single theatrical year, is the form in which honest wreck dates are published, and learning to read it is the reader’s main defense against the headline version of archaeology.

Reading the Last Cargo: Residues in the Clay

The contents of an amphora outlast its wine. Organic compounds soak into the porous ceramic wall during a jar’s working life and persist there, protected from oxygen and light, long after the liquid has drained, evaporated or been consumed by the sea. Residue analysis recovers them with solvents and reads them by gas chromatography and mass spectrometry, instruments that separate a mixture into its compounds and identify each by its molecular weight and fragmentation pattern. Certain compounds act as biomarkers: tartaric acid in quantity points to grape wine, specific fatty acid ratios point to olive oil or animal fats, and diterpenoid compounds point to pine resins used to seal jars or flavor their contents. Patrick McGovern of the University of Pennsylvania built a research career on such signatures, reconstructing ancient fermented beverages from sherds and jar interiors across the Near East and Mediterranean.

Why do archaeologists keep the mud inside a recovered jar?

The sediment inside a recovered jar is a sample, not dirt: it holds pollen, seeds, absorbed oils and the chemical signature of the last contents, plus the concretion that can prove the jar lay undisturbed beside its neighbors. Excavators seal jars intact because washing one on deck destroys evidence no later technique can recover.

Residues serve dating indirectly and identification directly. They rarely carry a clock of their own, though radiocarbon can be attempted on short lived organic fractions, but they establish what a cargo was, and knowing the commodity disciplines the typology: a jar form associated with a specific product in a specific region narrows both origin and period. At Kekova, the viscous residues still clinging to amphora interiors when Karin Sowada of Macquarie University examined the raised jars were part of the announcement’s force, since identifiable contents would tie the scatter to the long distance carriage of oil, wine or resins rather than to casual loss. A residue result also carries a caution the excavators themselves stated: amphorae were reused repeatedly, and a jar’s absorbed signature may record several cargoes across its life, with the strongest or latest signal dominating. The chemistry, like every other witness in this subject, testifies about the object first and about the voyage only through argument. For readers who want to follow how such laboratory identifications are reported, a companion reference summary gathers the methods in plain form.

Stone Anchors and the Shape of What Vanished

Stone anchors are the most durable part of an ancient ship. A pierced stone, its hole worn by rope, sinks with the vessel or is lost overboard in an emergency, and stone does not rot, burn or interest salvors. Honor Frost, who drew the finds at Cape Gelidonya in 1960 and devoted decades to anchor studies, showed that anchor forms evolve with ship size and rig: simple pierced weights give way to composite anchors with wooden flukes and stocks, and eventually to the lead stocked anchors of the Roman world. An anchor assemblage therefore carries chronological information, cruder than pottery but independent of it, and the positions of anchors relative to a cargo scatter reconstruct the orientation of a vanished hull, bow from stern, in a way no other surviving element can.

At Kekova the three stone anchors did quiet but essential work in the claim. Jars alone might be jettison, as the skeptics noted, but jars in association with multiple anchors fit the equipment of a vessel that anchored, dragged or foundered at that slope. The counter argument, that anchors too are lost singly and accumulate at known anchorages, is also documented, which is why anchor evidence corroborates rather than decides. Frost’s studies of anchorages around the Levantine and Turkish coasts mapped exactly such accumulations, giving both sides of the Kekova debate their reference material. The pattern recurs throughout wreck dating: each class of evidence is ambiguous alone, and the date stands where independent ambiguities resolve in the same direction.

Anchors contribute one more service, negative evidence of a respectable kind. A scatter claimed as a very early wreck that includes anchor forms developed millennia later is not an early wreck, whatever its pottery suggests, and mixed anchor types across a site can expose the palimpsest of several losses that a cargo only sample would merge into one. Excavation reports that publish their anchors in full, with weights, stone types and positions, give later scholars the means to test the claim independently, and the Kekova team’s publication of its anchor count and positions allowed the debate to proceed on evidence rather than on authority.

Kekova 2026: The Claim and the Evidence

In October 2026, Turkey’s Minister of Culture and Tourism, Mehmet Nuri Ersoy, announced that divers working off Kekova Island in Antalya Province, the island known in antiquity as Dolichiste, had identified the remains of a trading vessel dated to between 4,400 and 4,700 years ago. The excavation, led by Hakan Öniz of Akdeniz University, recovered more than 30 ceramic amphorae and three stone anchors from a seabed slope about 30 metres deep. The vessel was estimated at about ten metres in length, a coasting trader of the Early Bronze Age, lost in the same broad era in which the Great Pyramid of Giza was under construction, an era whose building workforce the archive examines in the account of who really built the pyramids. The claim, if sustained, makes the Kekova assemblage the oldest shipwreck identified anywhere, older than the Dokos wreck that had held that standing for decades.

The evidence offered was typology and association. No hull timber survived, so radiocarbon and dendrochronology had nothing to measure. The amphorae, with twin handles, a central ridge and flat bases, were matched to dated parallels from the Jordan Valley, Egypt, Troy and Crete, and their clay suggested production in the Jordan Valley region. Residues inside several jars awaited laboratory identification, with olive oil, wine and tree resins named as candidates. Öniz described the site in a widely quoted phrase, calling it a time capsule on the seabed, and the planned Mediterranean Underwater Archaeology Museum in Antalya was announced as the collection’s destination. The announcement traveled through the Turkish press, the BBC and specialist outlets including the Biblical Archaeology Society, whose report of October 19, 2026, by Lauren K. McCormick set out both the finds and the external caution in unusual detail for a news account.

Measured against the standards of the preceding sections, the Kekova position is legible. The association argument rests on the patterned co location of jars and anchors on one slope. The dating rests on a single instrument, ceramic typology, applied to a small assemblage, with no independent clock yet available and the residue results still to come. The excavators’ own framing, a date range spanning three centuries, honestly reflects typology’s precision at that depth of time. What the claim cannot yet supply is convergence, the agreement of independent methods that distinguishes Uluburun and Kyrenia, and the outside response, examined in the complication section below, focused precisely on that gap. Kekova is therefore best understood as a well formed hypothesis at the start of its testing, published by a ministry before the laboratory work that could strengthen or weaken it.

Dokos: The Benchmark Kekova Is Measured Against

The wreck off Dokos, a small island near Hydra in the Argolic Gulf, held the standing of oldest known shipwreck from its excavation until the Kekova announcement, and its history shows how such standings are built. Peter Throckmorton, the American photographer and wreck hunter whose surveys transformed Aegean archaeology, located the site in 1975, and the Hellenic Institute of Marine Archaeology under George Papathanasopoulos excavated it in the seasons that followed. The finds were a cargo scatter rather than a hull: thousands of pottery fragments of Early Helladic II types, obsidian blades, millstones and lead ingots, spread where a vessel had struck and broken. The pottery dated the loss to the third millennium BCE, commonly summarized at around 2200 BCE, making the Dokos assemblage roughly a millennium older than Uluburun.

The Dokos cargo also carried a sourcing lesson that later work generalized. Its obsidian, the volcanic glass prized for blades before metal became cheap, could be traced by its chemistry to the island of Melos, the Aegean’s great obsidian source, tying the cargo to exchange routes that land sites had already mapped. Its pottery fabrics pointed to production centers around the Argolic and the Cyclades. A cargo, properly sourced, becomes a route; a dated cargo becomes a dated route, and the accumulation of dated routes is how the Bronze Age exchange system acquired its chronology. The Kekova jars, with their suggested Jordan Valley clay, promise the same conversion if the association holds, a route from the Levantine interior to the sea and westward, dated by the very jars that traveled it.

Dokos matters to the Kekova debate for two reasons beyond chronology. First, its acceptance as a wreck rested on the same class of evidence Kekova offers, a cargo scatter without a surviving hull, interpreted through pottery typology and the pattern of its distribution. The discipline has therefore already crossed the bridge that Kekova’s critics question, and the argument is about the quality of the scatter and its parallels, not about whether hull less wrecks can exist. Second, Dokos taught the field the cost of coarse parallels: its pottery types belonged to a long lived Early Bronze Age repertoire, and its date has always been carried as a broad range rather than a point. Kekova’s claimed range, three centuries wide, sits inside the same convention. The benchmark thus disciplines both sides, reminding supporters that the standing was always provisional and reminding skeptics that provisional standings, honestly ranged, are how this science normally proceeds.

Cape Gelidonya 1960: The Excavation That Set the Standard

The modern standard for proving and dating a wreck was set in the summer of 1960 at Cape Gelidonya, on Turkey’s Lycia coast, where George Bass directed the excavation of a Late Bronze Age vessel lost at about 1200 BCE. Throckmorton had found the site in 1959 during his survey of the coast, noting bronze objects concreted to the rocks at about 27 metres, and he brought it to Bass, a young scholar at the University of Pennsylvania, with a proposition that sounds obvious only in retrospect: that a wreck could be excavated underwater to the same standard as a site on land, with a grid, measured drawings, photographic recording and finds lifted in context rather than as treasure. The expedition’s team, including Honor Frost as archaeological illustrator, spent the season demonstrating that divers could do the work of excavators rather than the work of salvors.

The cargo, copper and bronze ingots, scrap bronze for recycling, tin, tools and pottery of Cypriot and Levantine types, dated the vessel to the end of the Bronze Age by the same typological means that Kekova relies on, supported by the ingot forms and the pottery parallels. What Gelidonya changed was not the clock but the context: because every object was plotted before lifting, the association argument could be made from a plan rather than from an impression, and the resulting publication allowed scholars who never dived to test the dating for themselves. Bass drew the institutional conclusion and in 1972 founded the organization that became the Institute of Nautical Archaeology, later based at Texas A&M University, which professionalized wreck excavation worldwide. Every later standard invoked in this subject, including the standards by which Kekova will be judged, descends from the Gelidonya summer and its insistence that a wreck is an excavation that happens to be wet.

Gelidonya also left a substantive legacy in the study of the Bronze Age metal trade. Its cargo of ingots and scrap, a smith’s working stock of tools and partly melted bronze, suggested a traveling metalworker’s vessel or a recycling consignment rather than a palace shipment, a reading that reframed how scholars imagined the circulation of metal at the end of the Bronze Age. The reading depended entirely on context: the same bronzes lifted without positions would have been a diver’s trophy pile, while plotted among ingots, tools and balance weights they became an argument about an economy. That difference, between objects and evidence, is the whole content of the Gelidonya standard, and it is why the expedition’s grid and its drawn plans are still reproduced in textbooks long after its pottery dates ceased to be controversial.

Uluburun: When Every Clock Agreed

The Uluburun wreck is the discipline’s reference case for what a fully dated wreck looks like. A sponge diver, Mehmet Çakir, reported metal biscuits on the seabed off Uluburun, near Kaş, in 1982, recognizing copper ingots from his work with earlier survey teams. Excavation by the Institute of Nautical Archaeology, directed by George Bass and then by Cemal Pulak across eleven seasons from 1984 to 1994, recovered a cargo of about twenty tons: some ten tons of copper ingots, about a ton of tin, glass ingots, ebony logs, ivory, gold and faience jewelry, resin in jars by the hundredweight, and pottery from Cyprus, the Levant, Egypt and the Aegean. The ship had struck a rocky promontory and slid down a slope, spreading its cargo across a gradient that the excavators mapped object by object, more than 22,000 dives in total.

The cargo’s lesser items argue as eloquently as its ingots. Canaanite jars carried terebinth resin in quantities that spoke of the incense and unguent trade; hippopotamus teeth and elephant ivory waited as raw material for carvers; ostrich eggshells, ebony logs and a stone scepter head mapped a network reaching from Egypt and the Levant to the Aegean and beyond. A writing tablet of boxwood with ivory hinges, its wax surfaces lost, testified to the paperwork of Bronze Age commerce. Each class of object carried its own production geography and its own chronological span, and the excavators’ catalogue treated a potsherd and a gold pendant under the same discipline of position and count, which is why the cargo could be reassembled in print as a system rather than admired as a hoard.

The dating of Uluburun is a lesson in convergence. The pottery, Canaanite jars and Mycenaean and Cypriot wares, placed the sinking in the late fourteenth century BCE. A gold scarab bearing the name of Queen Nefertiti supplied a terminus post quem from the Amarna age. Radiocarbon measurements on short lived samples, wiggle matched against the calibration curve, agreed. Dendrochronology on the brushwood dunnage returned the felling date near 1305 BCE described earlier. The keel plank, of cedar, and the hull’s mortise and tenon construction matched Bronze Age shipbuilding known from Egyptian depictions. No single strand carried the date; the strands braided. When later scholars disputed details of the Uluburun chronology, they disputed them inside a structure of mutually checking evidence, adjusting the range by years rather than by centuries, which is what mature dating looks like.

The contrast with single instrument dates is the point. Uluburun’s excavators could afford precision because the site was rich, sealed and excavated over a decade, conditions that no announcement driven find enjoys in its first season. Judging Kekova by Uluburun’s finished standard would be unfair to a site in its first year; judging it by Uluburun’s method, asking how many independent constraints exist and whether they intersect, is exactly fair, and it is the standard the field applies to itself. The Bronze Age world that Uluburun sailed, the palace economies whose collapse the archive traces in the Sea Peoples and the Bronze Age collapse, is reconstructed substantially from cargoes like this one, each dated wreck adding a fixed point to a network otherwise known from texts and tomb paintings.

Kyrenia: A Hull That Survived to Be Questioned

The Kyrenia ship, a small Greek merchantman lost off the north coast of Cyprus, supplies the complementary case: a wreck whose hull survived and could be interrogated directly. Sponge diver Andreas Cariolou located the wreck in 1967, scattered amphorae and millstones marking it on a sandy bottom, and Michael Katzev of the University of Pennsylvania directed its excavation and raising in 1968 and 1969. About three quarters of the hull survived, preserved where it had settled into mud that excluded the worms, and the timbers were lifted, recorded and reassembled ashore, an operation that turned a wreck into a ship that scholars could walk around. The hull’s pine planking, joined edge to edge in the shell first manner of Greek construction, carried its own evidence of a long working life, repairs and a lead sheathing added against borers.

The small finds aboard filled out the human scale against which the clocks would be read. Tableware survived in four settings, with wooden spoons, a lamp and cooking gear, the equipment of a crew of four eating together in a space a modern reader would find cramped. More than 400 amphorae had been stowed as cargo, and 29 millstones rode as saleable ballast, heavy goods that paid their way. Iron spears concreted near the stern raised questions the excavators debated in print, defensive weapons or cargo, without forcing an answer the evidence could not carry. A ship this completely inventoried disciplines its own interpretation: the crew’s gear, the cargo and the hull form three populations whose dates must be reconciled together, and a stray object that fits none of them announces itself as an intrusion to be explained rather than a date to be averaged in.

Dating Kyrenia was an exercise in agreement among modest witnesses. The amphorae, mostly Rhodian and Samian, placed the cargo in the late fourth century BCE. Coins found in the hull’s region belonged to the years after Alexander the Great. Radiocarbon measurements on the timbers and cargo organics were consistent. The almond cargo, some 9,000 almonds recovered from the hold, and the millstones and iron spears among the finds filled out a working coaster’s last voyage at around 300 BCE. Nothing about Kyrenia required a record claim; its importance is methodological and, for shipwrights, experimental, since a full size sailing reconstruction, Kyrenia II, later tested the hull’s performance and confirmed the lines the timbers preserved. A dated hull that can be rebuilt and sailed is association made tangible: the clocks dated the objects, and the objects, reassembled, behaved like a ship.

The Ship as Its Own Document: Construction as a Clock

Where timbers survive, the ship itself testifies, because shipbuilding traditions changed in datable ways. Ancient Mediterranean hulls were built shell first, the planks edge joined with mortise and tenon fastenings to form a self supporting skin into which frames were fitted afterward, and the density, spacing and workmanship of those tenons shifted across centuries and regions in patterns specialists can read. The gradual transition toward skeleton first construction, in which frames define the shape and planking follows, unfolded across late antiquity and the early medieval centuries, unevenly and with regional overlap, so a hull’s building logic can place it within that long change. The Byzantine wrecks at Yenikapı displayed the transition in sequence within one harbor, hulls of the fifth century holding to older practice while later vessels mixed and then abandoned it.

Construction evidence disciplines cargo dating in both directions. A hull built in an old fashioned technique may be a conservative product, a repair heavy survivor or a genuinely early vessel, and the cargo decides among those readings; conversely, a cargo of mixed periods above a coherent hull suggests a long working life, the normal condition of merchant shipping in every age. Fastenings carry their own chronology: treenails, iron nails clenched or driven, copper alloy spikes and lead sheathing each belong to technological horizons, and the metallurgy of a nail can be tested against regional production. Even tool marks testify, adze patterns and saw types leaving signatures that ship historians compare across wrecks. None of these witnesses names a year, yet together they give the excavator an argument independent of every portable object, the argument of the machine itself, and on hull rich sites it is often the construction sequence that exposes a mixed deposit, by showing that the timbers form one vessel while the objects above them form two.

The Madrague de Giens wreck, a large Roman freighter excavated off the French coast in the 1970s, showed how much a hull can say. Its double planking, its mortise and tenon joinery at close spacing and its internal framing allowed scholars to reconstruct Roman shipbuilding at industrial scale, and its amphora cargo, some 6,000 jars of Italian wine, dated the loss to the first century BCE by forms and stamps that the hull’s carpentry corroborated. A vessel of that size was a fortune in timber and labor, and its construction record explains the economics of the trade that filled it. Where a Kekova offers objects without a ship and a Kyrenia offers a ship with modest objects, Madrague de Giens offers both at scale, and the three cases together map the range of testimony a wreck can give.

Antikythera: A Date Written in Cargo

The Antikythera wreck demonstrates dating where the cargo is famous and the ship is nearly anonymous. Sponge divers from Symi, blown off course in 1900, found bronze and marble statues on the seabed off the island of Antikythera, and the recovery that followed, directed by Greek authorities with Elias Stadiatis among the divers who first reported the find, brought up the corroded lump that proved to be the Antikythera mechanism, the geared astronomical calculator whose complexity reframed ancient technology. The ship itself left few timbers, and its date had to be assembled from its portable wealth: coins, pottery, glassware and the mechanism’s own inscriptions.

The coins did the heavy work. Issues from Pergamon and Ephesus among the finds clustered in the decades around 70 to 60 BCE, supplying the terminus post quem, while pottery and amphora forms of eastern Mediterranean types fitted the same generation. Derek de Solla Price, whose 1959 study Gears from the Greeks began the mechanism’s modern investigation, worked within that cargo derived frame, and the Antikythera Mechanism Research Project, applying computed tomography from 2005 onward, refined readings of the inscriptions without displacing the frame. The mechanism itself was later dated from within, by the inscriptions its makers engraved on its plates. The texts name calendar cycles, eclipse predictions and the dials’ own instructions, and their letter forms and astronomical references were argued by the research project to fit the decades the cargo indicated, a rare case of an artifact cross examining its own context. Where the inscriptional argument and the coin argument met, the frame hardened; where early scholars had proposed dates a century apart on mechanical grounds alone, the cargo’s discipline prevailed. The episode generalizes cleanly: an object’s internal evidence, however ingenious, is read inside the deposit’s chronology, never instead of it, and the deposit is dated by its least ambiguous members.

The wreck is a standing reminder that a date belongs to the loss event, not to the most interesting object: the mechanism may have been made years before the voyage, the statues may have been older still, as looted or traded art often was, and the youngest coins, unglamorous and precisely studied, carry the chronology.

Belitung: An Inscribed Bowl Dates a Dhow

The Belitung wreck extends the method beyond the Mediterranean and beyond antiquity. In 1998, fishermen and then a commercial recovery team located a wreck in the Gelasa Strait off Belitung Island, Indonesia, holding some 60,000 pieces of Chinese ceramics, gold and silver of the Tang dynasty. The vessel itself was the sensation for ship historians: an Arab style dhow, its planks sewn together with coir cordage rather than nailed, the earliest sewn vessel of its kind found in those waters and direct evidence of the maritime route between the Abbasid world and Tang China. Hull and cargo told different halves of one story, and the date came from the cargo’s humblest quarter.

The cargo’s composition mapped the trade it served. Changsha bowls formed the bulk, painted ware produced for export markets, accompanied by finer Yue celadons and northern white wares, bronze mirrors, gold and silver vessels that included an elaborately worked gold cup, and organic remains such as star anise that spoke of the return cargo economy of the southern seas. The mixture matched the manifest of a vessel loaded at a Chinese port, Guangzhou being the usual candidate in the scholarship, for the run westward toward the Abbasid world. Each ware type carried its own kiln chronology from land excavation, so the cargo arrived with a choir of typological voices already in rough agreement before the inscribed bowl spoke.

Among the Changsha ware bowls, produced in Hunan for export and painted with quick commercial brushwork, one bore an ink inscription giving a date corresponding to 826 CE, the second year of the Baoli era of the emperor Jingzong. A single inscribed bowl among tens of thousands fixed the cargo’s production horizon, and radiocarbon results from the hull timbers and cargo organics were consistent with a ninth century loss. The wreck’s later history also instructs: commercially recovered under license, with its cargo sold as a collection, Belitung became the discipline’s standard case in arguments about what is lost when context is recorded quickly for sale rather than slowly for science, a debate that frames the ownership questions addressed below. For dating alone, the bowl stands with the Rhodian stamp and the Pergamene coin as proof that bureaucratic habits, potters marking batches and states marking money, are the archaeologist’s most generous benefactors.

Mazarrón: Phoenician Ships in Lead and Pine

The two Phoenician wrecks off Mazarrón, on the Murcian coast of Spain, show the method working at the western edge of the ancient world. The first, located in 1988 and excavated in the following decade, survived as a partial hull with a cargo of lead ingots; the second, found in 1994, proved to be the most complete Phoenician vessel known, a small freighter of about eight metres whose pine timbers, preserved under their own cargo and a blanket of sand and seagrass, retained even the caulking and lashings of its construction. Both were dated to the seventh century BCE, the era in which Phoenician traders from the Levant were planting factories and colonies along the Iberian coast to reach the silver and lead of the interior.

The dating combined the usual witnesses in their western forms. The lead ingots, stamped and shaped to Iberian trade patterns, belonged to the mining economy that drew Phoenicians west. The pottery, Phoenician amphorae and table wares of types ordered by excavated sequences at Tyre, Carthage and the Iberian colonies, placed the cargoes in the seventh century, and radiocarbon measurements on the second wreck’s timbers and cargo organics agreed. The hull itself contributed construction evidence: sewn and pegged planking in the Phoenician manner, a technology whose distribution across the Mediterranean is itself a chronological and cultural marker. Iván Negueruela, who directed the Mazarrón work for Spain’s National Museum of Underwater Archaeology, presented the second wreck as a time witness of Phoenician shipbuilding precisely because hull, cargo and context survived together.

Mazarrón also demonstrates a conservation truth that feeds back into dating. The second wreck was left protected on the seabed for years while its excavation and a purpose built conservation facility were prepared, because raising waterlogged pine without a treatment plan destroys the evidence the raising was meant to save. Polyethylene glycol impregnation, the standard treatment that replaces water in degraded wood cell by cell, takes years for a hull, and the Vasa in Stockholm and the Mary Rose in Portsmouth stand as the great precedents for both the cost and the yield. A wreck raised in haste may keep its cargo and lose its timbers, and with them its dendrochronology, its construction sequence and its repair history, which is to say it keeps the objects and loses the ship. The Mazarrón decision to wait is the discipline’s patience made visible, and it preserved exactly the classes of evidence on which any future redating will depend.

The Complication: When the Jars Might Not Be a Wreck

The strongest objection to the Kekova claim deserves its best form, because it comes from specialists whose own work built the standards this subject depends on. Justin Leidwanger of Stanford observed that the published evidence is thin: a limited number of jars and three anchors that resist dating, with no hull to supply an independent clock, and that amphora forms of the relevant types were long lived, so shape matching alone leaves a wide margin. Sean Kingsley raised the sharper possibility that the jars may simply have been thrown overboard, jettisoned cargo or anchorage refuse rather than the remains of a lost ship, noting parallels between the Kekova jars and material from much later wrecks. Between them, the objections attack both halves of the claim: the association, whether a wreck exists there at all, and the date, whether Early Bronze Age typology can carry a record.

The objection has force because the history of the field contains its own false dawns, announced wrecks that dissolved under scrutiny into dumps, ballast piles and mixed anchorages, and because the oldest wreck label carries institutional rewards in funding, tourism and national prestige that give skepticism a disciplinary duty rather than a merely cautious one. Turkey’s announcement came from a ministry, with a museum already attached to the finds, a sequence that reverses the Gelidonya order in which excavation and publication preceded acclaim. None of that makes the claim wrong. Öniz’s team knows the Dokos precedent, in which a hull less scatter was accepted on pottery and pattern, and the Kekova distribution may well strengthen as excavation continues and the residue results arrive. It does mean the thesis stated at the outset must be narrowed for this case: a date is only as strong as its association argument, and at Kekova the association argument is the part still under construction.

The narrowing generalizes beyond one Turkish slope. Every wreck date in this subject is a confidence statement wearing a number, strong where independent clocks converge, provisional where one clock speaks alone, and honest in proportion to how plainly the report says which condition holds. The Kekova announcement, read with its range and its pending laboratory work stated, largely honors that convention; the headline version of it, oldest shipwreck found, discards exactly the qualifications that make the underlying claim scientific. The complication thus ends as a defense of the discipline’s own practice against its publicity, and the reader who holds the distinction can follow the Kekova testing as it proceeds, holding the right question in hand: not whether the team is believed, but what the next independent constraint will be as excavation and laboratory work proceed.

From Salvage to Science: The Discipline of Nautical Archaeology

For most of their history, wrecks were mines rather than archives. Salvors from antiquity onward recovered cargo, cannon and coin from sunken ships, and sponge divers across the Aegean brought up antiquities as a byproduct of their trade, the route by which both Antikythera and Kyrenia first came to notice. The transformation of that inheritance into a science was the work of the generation of Bass and Throckmorton: Throckmorton, surveying the Turkish coast in the late 1950s, catalogued wreck sites from fishermen’s knowledge and argued they constituted an archive of ancient trade awaiting scholars; Bass supplied the scholarly method at Gelidonya in 1960 and then the institution, founding in 1972 the body that became the Institute of Nautical Archaeology, whose expeditions and journal set professional expectations for excavation, conservation and publication.

The legal frame followed the professional one, slowly. The UNESCO Convention on the Protection of the Underwater Cultural Heritage, adopted in 2001, declared in situ preservation the first option for wrecks a century old or more, ruled commercial exploitation incompatible with protection, and required that recovered material be documented to scientific standards. States have adopted it unevenly, and treasure hunting under salvage law persists in many jurisdictions, but the convention states the discipline’s consensus: a wreck’s value lies in its context, and context destroyed for cargo is information destroyed for everyone. The galleons of the Spanish treasure fleets make the point in both directions, scientifically excavated Armada wrecks off Ireland yielding evidence about sixteenth century naval warfare that the archive surveys in the Spanish Armada of 1588 explained, while bullion first recoveries elsewhere left hulls and associations unrecorded beyond recovery.

The professionalization also changed what counts as a publishable claim. A modern excavation report carries its dating evidence in auditable form: plans showing every plotted object, catalogues with counts of diagnostic and undiagnostic material, sample lists for every laboratory date with the material sampled and the laboratory that measured it, and the calibration details that let another scholar recompute the range. Claims announced without that apparatus are understood within the field as preliminary, whatever their wording, and are weighted accordingly. The standard is demanding because the excavation is destructive: a wreck dug is a wreck unrepeatable, and the report is the only site that survives.

Survey technology changed the supply of sites faster than excavation could follow. Side scan sonar and magnetometers towed over likely coasts turned fishermen’s lore into mapped anomalies by the hundred, and remotely operated vehicles opened depths no diver team could work, including the Black Sea’s anoxic layer, where a Byzantine wreck located by a team led by Robert Ballard in 2000 stood with its hull and amphora cargo eerily intact in water that excludes the organisms of decay. Discovery, however, is the cheap part of the discipline. Each located wreck joins a queue for excavation funding that stretches across careers, and the growing register of recorded but unexcavated sites is itself a dating resource, since survey photography can establish a cargo’s character and broad period without disturbing the association that a future excavation will need.

Sourcing the Cargo: Clay Fabrics and Lead Isotopes

Dating tells when; sourcing tells where from, and the two disciplines reinforce each other on a wreck. Ceramic petrography, the microscopic study of a sherd’s clay fabric, identifies the mineral inclusions a potter’s clay carried from its geology, matching vessels to production regions whose kilns and clays have been sampled. Neutron activation and X ray fluorescence extend the matching chemically. A cargo whose jars share one fabric and one form is a single consignment, evidence for association; a cargo of mixed fabrics tells a story of collection at several ports, and its dating must respect the mixture. At Kekova, the attribution of the amphora clay to the Jordan Valley region was offered in exactly this spirit, tying the scatter to the eastern Mediterranean exchange system whose other nodes, in Egypt, Troy and Crete, supplied the shape parallels.

Metal sourcing has had a more turbulent career, and its turbulence is instructive. Lead isotope analysis, developed for Bronze Age metals by Noel Gale and Zofia Stos Gale at Oxford, promised to fingerprint copper and lead to their ore bodies by the isotope ratios that geology fixes in each deposit. Applied to the oxhide ingots of Uluburun and Cape Gelidonya, the method attributed copper to Cypriot ores and reorganized debate about Bronze Age metal trade. Critics replied that ore fields overlap, that mixing and recycling blur signatures, and that early claims outran the reference databases, and the field absorbed the correction by treating isotope results as probability statements against published data rather than as verdicts. The episode is the sourcing version of the dating lesson: an instrument’s output is a constraint to be combined, not an oracle to be quoted, and its authority grows exactly as its error structure is published with it.

Thermoluminescence and the Other Clocks

Thermoluminescence dating answers a question the other clocks cannot: when was this ceramic last fired? Clay minerals accumulate trapped electrons from environmental radiation at a steady rate after firing resets the accumulation to zero, and reheating a sample in the laboratory releases the stored energy as light proportional to the time since firing. For wreck material the method dates the pot rather than the voyage, with precisions typically wider than typology can achieve, so its wreck applications are specialized: authenticating raised objects whose context was lost, testing whether a sherd belongs to an early or a late firing tradition where forms overlap, and checking suspect pieces in museum collections. Its laboratory home for much of its development was Oxford’s Research Laboratory for Archaeology and the History of Art, and its role in exposing forged antiquities is a large part of its reputation.

Around these principal instruments orbits a constellation of lesser clocks, each with a wreck niche. Archaeomagnetic dating reads the direction of the Earth’s magnetic field frozen into fired clay at its last heating and matches it against regional field curves. Obsidian hydration measures the water absorbed into a freshly flaked obsidian surface, useful where blades form part of a cargo, as at Dokos. Corrosion and concretion growth have been proposed as relative indicators and are treated with caution, since their rates vary with local chemistry. Radiocarbon applied to the organic fraction of residues and to short lived cargo remains, seeds, grain and basketry, remains the workhorse among the auxiliaries, chosen wherever a sample’s brief life ties it tightly to the voyage. The pattern across all of them repeats the central discipline: match the clock to the event it actually measures, publish its error, and let independent clocks vote.

The six clocks can now be set side by side. The table settles, for each method, which event it actually dates, the precision it characteristically delivers, and the weakness a reader should probe first.

Method	What It Dates	Typical Precision	Main Weakness
Ceramic typology	Manufacture of the pottery forms in the cargo	A generation to three centuries, widening with age	Long lived forms and regional lag blur the range
Radiocarbon	Death of the sampled organism, wood, seed or residue	Decades to two centuries after calibration	Old wood, reservoir offsets and calibration plateaus widen ranges
Dendrochronology	Felling year of the sampled timber or brushwood	A single year where the match is secure	Needs long ring sequences and a regional master chronology
Coins and inscriptions	Minting or inscription year of the youngest piece	A year to a reign for the floor it sets	Circulation and hoarding separate issue from loss
Residue analysis	Identity of the last contents held by a vessel	Identifies commodity rather than year	Reuse layers several cargoes into one signature
Thermoluminescence	Last firing of the ceramic sample	Broad, often a tenth of the age or wider	Dates the pot, not the voyage, with wide error

A Team at Work: How a Date Gets Built

Consider a composite case assembled from standard practice, a survey team that has located a scatter of amphorae and anchors on a slope in 30 metres of water and must decide what it has found. The first season is survey rather than excavation: the team grids the site with lines and tags, photographs every visible object in place, and builds a photogrammetric model from thousands of overlapping images, so that the distribution exists as data before anything moves. The pattern is read against the questions of association. Do the jars form an elongated cluster consistent with a hull? Do the anchors sit at its ends? Are there galley wares, ballast, or fittings that a jettison would not explain? A scatter that fails these questions is recorded and left as a scatter, a result the team must be willing to publish.

If the pattern holds, targeted excavation follows, trenches placed to test the cluster’s edges and its burial. Diagnostic objects are lifted with their positions recorded; samples are taken under protocol, outer ring wood and short lived organics for radiocarbon, sherds for fabric analysis, sealed jars kept whole for residue work. The laboratory returns arrive as ranges and probabilities, and the team assembles the constraint intersection: typology’s span for the jars, the radiocarbon ranges calibrated, any stamps or coins read for their floor. Divergence is investigated as a site problem, intrusion, mixture or reuse, before it is published as a date problem. The report then states the range the constraints support and the confidence the association warrants, and deposits the catalogue where other scholars can recompute both. Every stage of that sequence is a place where a date can be inflated by silence, and the sequence is therefore also a reader’s checklist for any claim, including the claim that opened this subject at Kekova.

The same sequence, run on a rich Roman wreck, shows the machinery at full power. A first century BCE freighter’s mound yields amphora stamps naming consular years, coins of the late Republic, dendrochronology from its framing timbers and radiocarbon from its dunnage, and the constraints intersect within a decade. Run on a sparse Bronze Age scatter, the machinery returns three centuries and a research program. The machinery does not change; the evidence does, and the discipline’s maturity consists in reporting the machinery’s output without embarrassment in either case. Teams that publish their sparse results at full width train the typologies and the regional chronologies that will narrow the next team’s range, which is how the calibration set grows: one honestly ranged wreck at a time, each depositing its constraints into the common stock.

What a Date Range Actually Means

A published wreck range, 2700 to 2400 BCE or 1305 BCE with a small margin, compresses three different kinds of statement, and readers profit from separating them. The first is measurement: an instrument returned a value with an error, the radiocarbon laboratory’s plus or minus, the dendrochronologist’s match probability. The second is interpretation: the measured event, a tree’s felling or a jar’s firing, has been argued to stand close to the event of interest, the sinking. The third is association: the dated objects have been argued to belong to one event at all. Headline precision usually borrows the first statement’s confidence and spends it on the third, where it was never earned. The methods table given earlier in this subject, setting each clock beside what it dates and where it weakens, exists to keep the three statements apart, and returning to it with a specific claim in hand is the fastest way to locate the soft point in that claim.

The dating of origins in other fields follows the same grammar, which is why the archive’s treatment of when animals first evolved reads so similarly despite its different evidence: fossils supply floors, molecular clocks supply ranges with model dependent errors, and the published answer is an intersection argued from association of evidence rather than a reading from one instrument. In both subjects, the expert habit is identical, to ask of any date which event was measured, how the measurement reaches the event of interest, and what would have to be true for the chain to break. Wreck archaeology simply makes the chain visible, because its links lie on the seabed in rows, waiting to be plotted.

There is a final statistical courtesy that good reports observe and readers should expect: ranges are given at stated confidence, conventionally the 95 percent range for calibrated radiocarbon dates, and the range’s width is information rather than failure. A three century range for an Early Bronze Age scatter is not a confession of ignorance; it is the honest output of typology at that depth, and a narrower claim from the same evidence would be worse science wearing better tailoring. Precision is purchasable only with independent constraints, and the history of the discipline is largely the history of learning to buy it honestly, one clock at a time.

Who Owns a Wreck: Law, Treasure Hunters and the Record

Ownership law shapes what dating evidence survives. Admiralty salvage doctrines, built for rescuing ships in peril, reward recoverers in proportion to cargo value and historically granted rights over wrecks treated as abandoned property, a framework that fits a foundering freighter and misfits a Bronze Age scatter whose value is knowledge. Sovereign and state claims complicate the picture further: warships remain the property of their flag states wherever they lie, coastal states assert jurisdiction over territorial waters, and the Abandoned Shipwreck Act of 1987 in the United States transferred title to most abandoned wrecks in state waters to the states themselves. Each legal answer decides an archaeological outcome, because a salvor paid by recovered value will lift coins and ingots and leave the plotted distribution, the amphora counts and the anchor positions that make dating possible.

The Belitung recovery is the case the discipline cites when the stakes need stating plainly. The cargo was recorded to the standards a commercial license required and conserved and sold as a collection, and scholars gained a dated dhow and 60,000 Tang objects whose broad context survived, while losing the fine grain association evidence that a slower excavation would have preserved. Reasonable specialists weigh that bargain differently, given that licensed recovery also financed conservation that might otherwise never have occurred, and the honest literature presents both the gain and the loss. For dating alone the moral is narrower: every legal regime that touches a wreck should be judged partly by whether the record it produces would let a later scholar recompute the date. A claim that cannot be recomputed is a claim on trust, and trust is the one currency this subject tries not to spend.

The frigate Nuestra Señora de las Mercedes, sunk by a British squadron in 1804 and recovered commercially by Odyssey Marine Exploration in 2007, tested the alternative principle in court. Spain sued for the cargo of some 594,000 coins, arguing that a sovereign warship and its contents had never been abandoned, and in 2012 the United States courts ordered the entire recovery returned to Spain. The legal result turned on sovereignty rather than on archaeology, yet its archaeological consequence was plain: a recovery conducted for cargo value, with its site documentation contested by scholars throughout the litigation, ended with the collection repatriated and the wreck’s context debated rather than established. Set beside Yenikapı’s silted fleet or the Mazarrón wrecks waiting under protection for their conservators, the Mercedes record marks the distance between a wreck treated as a pit and a wreck treated as an archive, and the distance is measured in exactly the evidence that dating consumes.

Reading a Shipwreck Headline: Questions for the Skeptical Reader

A reader meeting an oldest, earliest or first wreck claim can test it in minutes with the apparatus assembled here, and the test requires no diving and no laboratory. Begin with the association: does the report describe a pattern, anchors, galley material, fittings and a hull shaped distribution, or does it show a handful of attractive objects? Continue with the clocks: how many independent methods contribute, and do their ranges intersect without being stretched? Ask what the dated samples were, since outer rings and seeds argue for a voyage and heartwood and shells argue for caution. Ask who disputes the claim and whether the report quotes the dispute fairly. A claim that survives those questions with its range intact has earned its headline; one that sheds its qualifications at each step has located its weakness for you.

What does a wreck date actually promise, and what does it not?

A wreck date promises that the loss cannot be earlier than its youngest securely associated object, and that the combined evidence places the loss inside a stated range at a stated confidence. It does not promise the year the ship was built, the age of every object aboard, or immunity from revision when a new constraint arrives.

Applied to Kekova, the questions return a mixed and therefore credible verdict. The association case is moderate: jars and three anchors in one slope distribution fit a small trader, while the jettison alternative remains open until excavation tests the scatter’s edges and burial. The clock count is one, ceramic typology, applied by specialists with deep regional knowledge, and its three century range is honestly stated. The samples that could add a second clock, residues for identification and any organics for radiocarbon, were named as forthcoming laboratory work rather than silently assumed. The skeptics were quoted in the specialist coverage with their objections intact. A reader applying the test lends the claim provisional belief at its stated range, withholds belief from the record label, and waits for the second constraint. That posture, neither credulous nor dismissive, is the one the evidence has earned.

The practical habit underneath the questions is to read dates as arguments and ranges as honesty. Where a report names its excavators, its samples and its skeptics, a reader can follow the reasoning and lend belief in proportion; where a claim arrives as a ceremony, belief should wait for the catalogue. Readers who keep research notes across many such claims will find that a simple field notebook habit, recording the range and the clocks behind each date rather than the headline year alone, converts a stream of announcements into a personal chronology that can be checked when later evidence lands. The discipline rewards that patience, because its own chronology was built by exactly that accumulation of checked claims, each one surviving the questions that removed its weaker contemporaries.

The Ship Is Gone, the Argument Remains

A shipwreck is dated by its youngest trustworthy object, never by the ship itself: that sentence, stated at the outset, has now been walked through ceramics and coins, tree rings and residues, anchors and isotope ratios, and it holds because each method was found measuring an object and reaching the sinking only through an argued association. The Kekova scatter may in time join Dokos, Gelidonya, Uluburun and Kyrenia in the calibration set, its date narrowing as independent constraints accumulate, or it may resolve into an anchorage deposit whose jars still teach the typology that dated them. Either outcome leaves the method intact, which is the deeper point. Ships are lost once and briefly; the argument about their dates is rebuilt by every generation of excavators, and it is the argument, tested and published and recomputed, that constitutes the knowledge.

There is a discipline of reading hidden in that conclusion, and it transfers to every dated thing. A date in archaeology is a relationship between an object, an event and an argument, and it deserves exactly as much confidence as the weakest of the three. The seabed makes the relationship visible because its evidence arrives scattered, encrusted and incomplete, refusing the false neatness of a label in a case. What the excavators of Gelidonya understood in 1960, and what every honest wreck report since has practiced, is that the scatter is not an obstacle to knowledge but its form: knowledge of the past is what survives after association has been proven, clocks have been cross checked and doubt has been published alongside the range. The ship is gone. The standard of proof remains, and it is enough.

Frequently Asked Questions

Q: What counts as a shipwreck if the hull is gone?

A shipwreck is established by pattern and association rather than by surviving timber. Excavators look for a cluster whose parts make sense only as a vessel’s remains: cargo jars of one period arranged in an elongated spread, anchors positioned where a ship carried them, ballast, cooking wares and personal equipment that no cargo consignment explains. Each element is weak alone, since jars are jettisoned and anchors are lost singly, but their combination in one distribution is difficult to produce by accident. A scatter that cannot exclude dumping, jettison or repeated losses at an anchorage is recorded as a scatter, and its finds are dated as objects rather than as a sinking.

Q: How does ceramic typology date pottery from a wreck?

Typology dates pottery by matching a vessel’s profile, rim, handles, base and fabric against a reference sequence built from hundreds of dated deposits on land. Amphora shapes changed steadily because potters answered shifting demands of transport, taxation, stacking and fashion, so a form’s career of appearance, dominance and disappearance can be charted across generations. A wreck jar placed inside that chart inherits the chart’s dates, with precision that depends on how fast the form evolved and how well the land sequence is anchored. The method gives ranges rather than years, narrow for fast changing Roman and Hellenistic forms, wide for conservative Bronze Age shapes.

Q: What is seriation in archaeology?

Seriation is the ordering of assemblages by the changing popularity of the types they contain. Styles are born, flourish and fade, so deposits can be arranged in a sequence in which each type’s frequency rises and falls in a smooth curve, an arrangement Flinders Petrie demonstrated in 1899 with Egyptian cemetery pottery. The sequence is relative at first, saying only what came before what, and becomes chronological when independently dated deposits pin points along it. On wrecks, seriation orders cargoes against each other and against port deposits, and disagreements between a wreck’s seriation position and its other dates are treated as evidence of mixture, reuse or intrusion.

Q: How accurate is radiocarbon dating on shipwreck material?

Radiocarbon dating of wreck material typically delivers calibrated ranges spanning several decades to two centuries, depending on the period, the sample and the shape of the calibration curve at that point. Accuracy is limited less by laboratory measurement, which is precise, than by the relationship between the sample and the sinking. Heartwood from a long lived tree can predate the voyage by generations, while seeds, twigs and food remains grew within a season or two of it. Marine samples carry an additional reservoir offset. Published ranges at 95 percent confidence that name their sample material deserve trust in proportion to how short lived that material was.

Q: What is the marine reservoir effect?

The marine reservoir effect is the offset that makes marine organisms appear older to radiocarbon dating than they are. Ocean surface water mixes slowly with the atmosphere, and deep water rising in upwelling zones has been cut off from the air for centuries, so shells, fish and sea mammals build tissue from carbon that is already partly decayed. The global average offset is roughly four centuries, with regional corrections that vary by basin, current and estuary. Wreck excavators handle the effect by preferring terrestrial samples such as timber, seeds and cordage, and by applying a published local correction whenever a marine sample must be used.

Q: How does dendrochronology date a shipwreck?

Dendrochronology dates wood by matching its pattern of annual growth rings against a master chronology built for the same species and climate region. Wet and dry years widen and narrow rings in a shared regional rhythm, so a timber’s sequence locks into the master at one position, naming the calendar year of each ring. Where the bark edge survives, the felling year is known exactly, and felling dates for brushwood and repair timbers cut near a voyage stand close to the sinking. The method gave the Uluburun wreck its celebrated precision, and it fails gracefully where timbers are too short, too eroded or from regions lacking a master sequence.

Q: Why are coins so useful for dating a wreck?

Coins carry their issuing authority, type and often their year or magistrate on their faces, and numismatic scholarship has ordered those issues into tight chronologies. A coin therefore supplies a hard floor for a wreck’s date, since the sinking cannot precede the minting of the youngest coin lost with the ship. Groups of coins are stronger than single finds because a purse or payment reflects the currency circulating at one moment, while a single coin might be an heirloom kept for generations. Coin evidence anchored the Antikythera wreck’s chronology, and its discipline is constant: assemblages date events, singletons only decorate reports.

Q: What does terminus post quem mean?

Terminus post quem is Latin for the time after which, and it names the earliest possible date of a deposit, set by the youngest object the deposit contains. A wreck holding a jar form first made in a given generation cannot have sunk before that generation, however old its other contents may be. The concept disciplines wreck dating by separating the floor, which evidence fixes firmly, from the sinking itself, which may follow the floor by the years a coin circulated, a jar was reused or a ship remained in service. Responsible reports state the floor and then argue the interval, rather than presenting the floor as the date.

Q: Can scientists tell what an ancient amphora once held?

In many cases, yes, within limits set by reuse and preservation. Organic compounds from wine, oils and resins soak into an amphora’s porous wall and survive there for millennia, and laboratory instruments separate and identify them by their molecular signatures. Tartaric acid indicates grape wine, characteristic fatty acid patterns indicate olive oil or animal fat, and resin compounds indicate the pine products used to seal or flavor. Because amphorae were refilled across their working lives, a result may blend several cargoes, with the last or strongest dominating. A residue identification therefore describes the jar’s history first and the final voyage only where context supports the step.

Q: What are stone anchors and why do they matter to dating?

Stone anchors are pierced weights, among the oldest anchor forms, that held early ships by mass and friction before iron anchors spread. Their shapes, sizes and arrangements changed as vessels grew, and specialists such as Honor Frost charted that development, so an anchor assemblage contributes an independent, if coarse, chronological signal. Anchors matter equally for identification: their positions relative to a cargo scatter can reveal a vanished hull’s orientation and support the argument that a wreck exists at all. They corroborate rather than decide, since anchors were also lost singly at anchorages, and careful reports publish their counts, weights, stone types and positions for others to test.

Q: How old is the Kekova wreck said to be?

The Kekova assemblage off Antalya Province in Turkey was announced in October 2026 as dating to between 4,400 and 4,700 years ago, in the Early Bronze Age span of roughly 2700 to 2400 BCE. The excavation led by Hakan Öniz of Akdeniz University recovered more than 30 amphorae and three stone anchors from about 30 metres of water, with the vessel estimated at ten metres. The date rests on amphora shapes matched to parallels in Egypt, the Jordan Valley, Troy and Crete, since no hull timber survived for radiocarbon or tree ring measurement. Outside specialists urged caution, and residue analysis was still pending when the claim was announced.

Q: What was the Dokos shipwreck?

The Dokos wreck, near the island of Hydra in Greece, was for decades the oldest shipwreck known to archaeology. Located by Peter Throckmorton in 1975 and excavated by the Hellenic Institute of Marine Archaeology under George Papathanasopoulos, it consisted of a cargo scatter without a surviving hull: thousands of Early Helladic pottery fragments, obsidian blades, millstones and lead ingots. Its pottery placed the loss in the third millennium BCE, commonly summarized around 2200 BCE. Dokos established that a hull less wreck could be recognized and dated from its cargo pattern, the precedent against which later claims, including Kekova, are measured.

Q: What did the Uluburun excavation prove about wreck dating?

Uluburun proved that independent dating methods, applied to one sealed cargo, converge into a range far stronger than any method alone. The wreck off Kaş in Turkey, excavated by the Institute of Nautical Archaeology between 1984 and 1994, yielded pottery placing it in the late fourteenth century BCE, a scarab of Queen Nefertiti setting a floor, radiocarbon measurements on short lived samples in agreement, and a tree ring felling date near 1305 BCE from its brushwood dunnage. Because the strands checked each other, later disputes moved the date by years rather than centuries. The excavation became the discipline’s standard example of a date demonstrated rather than declared.

Q: How was the Antikythera ship dated?

The Antikythera ship was dated from its cargo, since little of the vessel survived its discovery by sponge divers in 1900. Coins from Pergamon and Ephesus among the finds clustered in the decades around 70 to 60 BCE and set the floor for the sinking, while amphorae, pottery and glassware of eastern Mediterranean types fitted the same generation. The celebrated geared mechanism, studied by Derek de Solla Price from 1959 and imaged by tomography in the twenty first century, was interpreted inside that cargo derived frame. The case teaches that a wreck’s date belongs to its youngest reliable evidence, not to its most famous object.

Q: How did an inscribed bowl date the Belitung wreck?

The Belitung wreck, a sewn plank dhow lost off Indonesia with some 60,000 Tang dynasty ceramics aboard, was dated by a single inscribed bowl. Among its Changsha ware, one bowl carried an ink inscription corresponding to 826 CE, fixing the cargo’s production horizon within a known reign period. Radiocarbon results from hull timbers and organic remains agreed with a ninth century loss. The inscription worked like a coin or a stamped amphora handle, a bureaucratic habit preserved in clay, and the case shows why excavators prize mundane marked objects over precious unmarked ones when a chronology must be built.

Q: Who decides whether a scatter of jars is really a wreck?

The decision belongs to the excavating team in the first instance and to the scholarly community in the long run, through publication and critique. An excavation report must show that a scatter’s pattern, contents and setting exclude jettison, dumping and anchorage loss, and specialists test that showing against excavated benchmarks when the report appears. Announcements by ministries or museums carry no evidential weight of their own. The Kekova debate followed this path exactly: the excavators presented jars and anchors as a wreck, and outside specialists such as Justin Leidwanger and Sean Kingsley asked whether the pattern excluded a jettison explanation.

Q: What is site formation in maritime archaeology?

Site formation describes everything that happens between a sinking and an excavation. A vessel breaks up according to its construction, the sea’s energy and the seabed’s shape, scattering buoyant and heavy material differently; then centuries of storms, currents, marine boring, encrustation and salvage filter what remains. Keith Muckelroy formalized the idea in 1978, naming the extracting and scrambling processes that bias every assemblage. For dating, formation studies matter because the surviving sample is not the ship’s full contents: late objects may have been salvaged away and intrusive objects dragged in, and an honest date states which distortions the site itself displays.

Q: How does thermoluminescence dating work on pottery?

Thermoluminescence dating measures the time since a ceramic was last fired. Firing empties the energy traps in the clay’s mineral crystals, and environmental radiation refills them at a steady rate afterward; reheating a sample in the laboratory releases the stored energy as light, and the brightness totals the years since firing. The method dates the pot rather than any voyage it made, and its precision is broad, so wreck specialists use it mainly to authenticate context less finds, to test suspect museum pieces and to check broad attributions where typology is ambiguous. Its development at Oxford made it a standard weapon against forged antiquities.

Q: What does the UNESCO convention say about old shipwrecks?

The UNESCO Convention on the Protection of the Underwater Cultural Heritage, adopted in 2001, covers traces of human existence under water for at least a century. It makes preservation in place the first option, declares protection incompatible with commercial exploitation, and requires that any recovery follow scientific standards with full documentation. States apply it unevenly, and salvage law still governs wrecks in many waters, but the convention records the professional consensus that a wreck’s context is its principal value. For dating, its practical effect is to protect the distributions and associations on which chronology depends from being dispersed for sale.

Q: How should a reader judge a new oldest shipwreck announcement?

Judge the announcement by its association argument and its clocks rather than by its record. Ask whether the report describes a wreck shaped pattern with anchors, fittings and galley material, or only a few striking objects. Count the independent dating methods and check whether their ranges genuinely intersect, and note what material was sampled, since seeds and outer tree rings speak for a voyage while heartwood and shells need correction. Look for named skeptics quoted fairly, a range stated with its confidence, and a catalogue others can recompute. Announcements that keep those elements visible have earned provisional belief; those that shed them have answered the question themselves.