Did Romans Reach America Before Columbus?

Jars recovered from the waters near Rio de Janeiro have fuelled claims that Roman sailors reached Brazil long before Christopher Columbus arrived in the Caribbean in 1492. During the 1970s and 1980s, reports of amphora-shaped objects in Guanabara Bay drew attention from people familiar with ancient Mediterranean shipping. Some specialists thought the jars resembled Roman transport containers. The resemblance prompted the suggestion that a merchant vessel, driven off course across the Atlantic, had sunk off the South American coast.

No Roman shipwreck or Roman archaeological site has been securely established anywhere in the Americas. A vessel that looks Roman, even if its manufacture in antiquity could be confirmed, would not by itself tell us who brought it into Guanabara Bay or when. Trade and collecting have moved ancient objects across oceans many centuries after their manufacture. An excavated ship with associated cargo and independent dating would provide a much firmer basis for claiming an ancient arrival.

Laboratory analysis can establish important things about pottery. The minerals and elements in its clay can help locate a production region, and chemical residues may reveal something about its contents. These methods have clarified the manufacture and distribution of Roman goods. Yet neither test determines when a jar first entered waters off Brazil. The Guanabara Bay controversy depends as much on excavation records and a find’s later history as on the composition of the ceramic itself.

How far did Roman ships actually sail?

Roman merchants carried goods far beyond the Italian peninsula. Their amphorae were large two-handled ceramic vessels used to transport wine, olive oil and fish products through the Mediterranean and to distant provinces such as Britain. Workshops frequently stood near the estates or industries supplying those goods. Jar shapes, manufacturing fabrics and stamps can thus help reconstruct trade routes. Roman merchandise also circulated beyond imperial borders, and a Roman-made object did not necessarily travel with a Roman crew throughout its journey.

The Roman trading world also included the Atlantic shore of Europe. Strabo, a Greek geographer writing around the beginning of the first century CE under Roman rule, described the maritime activity around Gades, now Cádiz, and the River Baetis, today’s Guadalquivir. In his account of southern Iberia, Strabo linked the region’s prosperity to its navigable river and access to ships. The existence of this Atlantic commerce is well established. It provides no record of a crossing to the coast of Brazil.

Long journeys over open water were possible in other parts of the Roman world. The anonymous first-century Periplus of the Erythraean Sea, especially sections 56–57 describes merchants using seasonal winds to sail between Red Sea ports and the Indian coast. Those voyages demonstrate knowledge of Indian Ocean navigation and a willingness to cross stretches of open sea. They do not show that crews possessed the same Atlantic routes, or that anyone undertook one. Their navigational experience alone is not evidence of a Roman Atlantic voyage.

What pottery chemistry can reveal

Pottery analysis begins with the clay from which a vessel was made. Local geology leaves patterns of minerals and chemical elements in that raw material. Instrumental neutron activation analysis, or INAA, determines concentrations of selected elements following neutron irradiation; X-ray fluorescence, or XRF, provides another way of measuring elemental composition. Researchers compare such measurements with pottery and clay from known regions and workshops. A close match may indicate a production area, especially when extensive reference collections exist and other possible sources have been tested.

A matching chemical profile is rarely decisive on its own. Workshops in different places can use comparable clay, makers may mix materials, and firing, burial or contamination can affect the measurements. A 2002 inter-laboratory study by Andreas Hein and colleagues compared techniques including neutron activation and XRF and investigated whether different laboratories could produce compatible pottery-provenance results. Its emphasis on calibration, statistical analysis and reliable reference samples explains why the result of an instrument is a starting point for an archaeological identification, rather than a unique geographical address.

Petrography examines the minerals and textures within the ceramic itself. Archaeologists polish a small ceramic sample into a thin section and inspect it with a polarising microscope. Mineral grains, fragments of rock and pores can reveal clay preparation and additions of temper, the non-plastic material mixed into it. Those characteristics may distinguish manufacturing traditions even when whole jars look alike. M. S. Tite’s 2008 review of ceramic production and provenance explains why materials analysis is most effective when read alongside the excavated context.

The contents of a vessel require a different investigation. Gas chromatography-mass spectrometry, usually shortened to GC-MS, can separate and identify organic compounds left in porous ceramic walls or surface coatings. Lipids may suggest plant oils or animal fats; resins might reveal pitch used to seal a jar; other molecules can be consistent with wine and plant products. In an examination of amphorae from wrecks near Marsala in Sicily, researchers reported plant-oil and pine-resin biomarkers in a 2021 study. Those biomarkers appeared in only a minority of the 33 sampled vessels, illustrating how unevenly such traces survive.

A positive residue result still needs care. Fatty acids change over time and can come from several foods. A vessel might have been reused, while a protective lining could influence the analysis. A study by Zuzana Chovanec on distinguishing olive oil from other plant oils examined how ageing alters the fatty-acid profiles used to identify a vessel’s contents. Comparison with suitable samples is therefore necessary before assigning a specific commodity.

The manufacture and use of an amphora must be distinguished from its later arrival in a harbour. A vessel made in Roman times could have been transported across the Atlantic in the nineteenth century. Conversely, an imitation could have been manufactured much later using clay with a similar composition. Deposition, movement and associated archaeological material must be examined separately from the manufacturing evidence.

The amphorae reported from Guanabara Bay

Reports of the Brazilian finds predate the involvement of Robert Marx. In 1976, the diver José Roberto Teixeira was said to have brought up jar-shaped objects from Guanabara Bay. Some recalled the elongated forms of Mediterranean transport pottery. Marx, an American underwater explorer and treasure hunter, publicised additional material in 1982 and argued that a Roman ship had sunk there. The United Press International report dated 15 December 1982 describes his request to excavate, the sonar anomalies he interpreted as a possible wreck, and his belief that the pottery supported an ancient Atlantic crossing.

Elizabeth Lyding Will, a University of Massachusetts Amherst scholar specialising in Roman amphorae, reportedly compared some fragments with pottery made around Kouass on Morocco’s Atlantic coast. Her preliminary comparison raised the possibility of a Mediterranean origin. It could establish neither an ancient date for every piece nor a single common wreck, much less continuous submersion since Roman times. Reports at the time even disagreed about the proposed dating, so Will’s reported assessment should remain a tentative attribution.

Marx also claimed to have observed substantial quantities of pottery fragments and sonar signs of a buried hull. Sonar can reveal shapes below water and sediment but cannot identify them as Roman. A shipwreck investigation would have to document wood, fastenings, cargo patterns and the deposits in which those remains lay. No independently verified excavation publication has established a Roman hull with an associated Roman cargo in Guanabara Bay. The supposed buried ship remains a report attributed to Marx.

Press accounts supplied a possible modern origin for some of the objects. Brazilian businessman Américo Santarelli reportedly said he ordered replica amphorae from Portugal, placed 16 in the bay in 1961 to age their appearance, and subsequently recovered only four. His account could explain some jar-shaped material. It does not scientifically establish that every fragment described by Marx was one of those replicas. Santarelli’s account must be linked to specific artefacts before it can explain their presence.

Official disputes further complicated the story. Marx accused Brazil’s navy of hiding the find; officials denied doing so. Press reports also described disagreements over recovery permits and over removing or selling material from shipwreck sites. These allegations help explain the controversy’s persistence but cannot substitute for an excavation. There is still no published, independently assessable record of each object’s location, associations and dating. The scholarly discussion also includes Mills in the Journal of Field Archaeology.

What would count as evidence of a Roman arrival?

Underwater archaeology depends on recorded findspots and associated deposits. The location of an object on a map is not enough: archaeologists need to know what lay with it and whether the site has been disturbed. Water currents, dredging, anchors and wrecks of different ages can mingle material on the seabed. A Roman-looking jar found beside an eighteenth-century vessel would have to be explained in relation to that later wreck before it could support a claim about an ancient crossing.

The precise archaeological position and setting of an object at discovery are its provenience. The ownership, handling and movements of the object after recovery belong to its provenance. The first could establish whether an amphora came from an undisturbed deposit, while the second could confirm that a laboratory examined the same artefact that divers actually recovered. Both records matter when evaluating an underwater discovery. The annex to UNESCO’s 2001 Underwater Cultural Heritage Convention describes standards for planning investigations, recording sites, conserving finds and treating underwater remains responsibly. Although adopted after the Guanabara dispute, those principles help explain what evidence is now expected.

A suspected wreck must also be dated through evidence securely associated with it. Jar typology may suggest a broad manufacturing period without identifying when the container was deposited in Atlantic waters. Radiocarbon measurements on short-lived organic objects, tree-ring dating of suitable timber, and coins or other diagnostic objects in a closed assemblage can help establish a chronology. Investigators must establish that the dated evidence belongs to the alleged wreck rather than a separate event. An old amphora reused by a later ship may predate the ship by centuries.

Investigators would need consistent evidence from the hull and cargo, secure dates, typology, chemistry and independent examination. A short accidental Roman visit need not leave a settlement or numerous buildings. Nevertheless, any proposed crossing requires careful testing of later object movements, possible intrusions and mistaken identifications. The absence of a Roman town in Brazil cannot exclude every imagined landing; the excavated evidence currently available from Guanabara Bay does not establish one.

How a genuine Roman object could reach America much later

Roman artefacts could have arrived in the Americas centuries after their manufacture. Collectors, merchants, museums, private owners and ships carried Mediterranean antiquities across the ocean in modern centuries. If a genuinely ancient amphora arrived that way, its clay would retain the characteristics of the Roman production site. A laboratory could identify its manufacture accurately while leaving the date of its Atlantic journey completely unresolved.

Ship ballast is another possible route by which old pottery could have been deposited. Ships required weight for stability, and ballast could be dumped or exchanged in harbours. An antique vessel could theoretically have moved as part of such a load. At Guanabara Bay, however, ballast must remain a hypothesis unless records or the site’s context show that this actually happened. The possibility alone is not an explanation of the finds.

Modern reproductions create another source of uncertainty. A recently made jar may become stained or acquire marine organisms once submerged. Barnacles and mud demonstrate exposure to water and sediment, not a Roman date. Researchers would need to consider modern workshops and the artefact’s history even if its chemical composition resembled an ancient reference group. Without those checks, a photograph of an encrusted amphora cannot establish when or how it reached the seabed.

What about the supposed Roman head found in Mexico?

Another debated object comes from Tecaxic-Calixtlahuaca in central Mexico. A small terracotta head discovered in 1933 has been compared by some specialists to Roman work of the second or third century CE. It was reportedly found as part of a burial offering, rather than among scattered and poorly recorded underwater material. The burial and its associated objects are generally assigned to the late fifteenth or early sixteenth century, long after the date suggested for the head’s manufacture. That buried context raises different questions from the underwater finds in Brazil.

In a 1999 paper in Ancient Mesoamerica, Romeo Hristov and Santiago Genovés argued that the head might indicate contact before Columbus. Their discussion included thermoluminescence testing, a technique that estimates the time since pottery was last fired. The resulting age range was broad, and subsequent scholarship disputed the way it had been reported. The test cannot identify a Roman maker, and the sculpture’s appearance cannot establish when it reached Mexico.

Questions about the original excavation and possible later intrusion or deliberate placement remain contested. Hristov and Genovés defend the burial context in their extended review of the evidence and answer critics, among them archaeologist Michael E. Smith. Their disagreement concerns both the context and the interpretation of the object. The head is nevertheless a single disputed object, with no independently documented series of events linking it to an Atlantic voyage in Roman times. It offers no independent confirmation for the Guanabara amphorae, nor do those jars resolve the Mexican dispute.

Why archaeologists accept the Norse presence in North America

At L’Anse aux Meadows in northern Newfoundland, excavated remains demonstrate a pre-Columbian European crossing of the Atlantic. Identified in 1960, the site was investigated in excavations led by Anne Stine Ingstad. Archaeologists uncovered eight structures with turf walls, including dwellings, workshops and a forge. Iron boat rivets and remains of ironworking fit Norse communities known from Greenland and Iceland. Parks Canada’s account of the excavations explains how the buildings and material remains were interpreted as a coherent settlement, rather than from the resemblance of a single artefact.

The dating evidence is equally specific. Research released online in October 2021 and published in the January 2022 issue of Nature studied three pieces of tool-marked wood. The samples contained the radiocarbon signal of an event in 993 CE, identifiable in their growth rings. Researchers counted from that ring to the surviving outer edge and found the trees were felled in 1021 CE. The study provides a precise year of Norse activity, although it does not establish the year of first arrival or how long Norse people stayed.

Buildings, manufacturing debris, recognisable artefacts and independent dates make the Norse identification persuasive. The UNESCO World Heritage record recognises L’Anse aux Meadows as the earliest confirmed European settlement in North America. An accidental Roman landing might have left far less material than the Norse settlement. It would, however, need comparably reliable evidence appropriate to the kind of visit claimed. That level of documentation is what separates the Newfoundland site from the contested Brazilian material.

What Roman amphorae genuinely tell us about trade

Roman amphorae show the extent of Mediterranean trade without requiring hypothetical Atlantic voyages. Dressel 20 jars, large and rounded, were mainly used to ship olive oil from Baetica in southern Spain. Pottery workshops on the Guadalquivir supplied vessels found across the empire. The London Museum’s record of a stamped Dressel 20 handle associates a stamped handle of this form with the production region. Britain has yielded vessels of this type, while huge quantities discarded in Rome eventually helped create Monte Testaccio, a mound made from broken containers.

Jar shapes varied with cargo and period. Dressel 1 jars are especially associated with wine transport in the late Roman Republic; families of vessels from Baetica and Lusitania carried fish products. Archaeologists call the comparison of rims, handles, bases and body shapes typology. Petrography and chemical testing can then investigate whether similar shapes share a manufacturing region. Written marks might identify a workshop, merchant, owner or contents, depending on their wording. Such evidence is strongest when studied with an excavated context and associated material.

A jar designed for one product might later be filled with another. Shape therefore identifies a likely original purpose rather than guaranteeing the contents at every stage of use. Surviving organic residues sometimes clarify the question, though many have degraded. Research on Lusitanian amphora contents has prompted reassessment of forms traditionally linked mainly with fish products by considering production sites, inscriptions, associated objects and chemical evidence together. The strongest interpretations draw on all those forms of evidence together.

Roman pottery and other products also reached peoples outside imperial territory. Goods could pass through multiple traders, routes and regional networks, a process explored in Spoken Past’s article on ancient trade routes. A Roman object at a distant destination therefore does not identify the ancestry or nationality of the person who transported it there. The empire’s documented maritime routes were extensive, but extending a line from known Atlantic ports to Brazil on a map is not evidence that ships sailed it.

Can chemistry ever help prove a Roman Atlantic crossing?

If a securely recorded Roman-period wreck were found in American waters, chemistry could help establish where its cargo was made. A ship found off the American coast with associated cargo, hull timbers and an undisturbed Roman-period deposit would allow investigators to compare its amphorae with Mediterranean production groups. Thin-section petrography might identify manufacturing traditions, residue testing could identify commodities, and construction details could be compared with excavated Roman vessels. Independent laboratories would be particularly useful given the importance of the conclusion.

Laboratory results would still need to be linked to the wreck’s archaeological date. Elemental composition does not date when a pot was made or when a vessel entered an American harbour. Residues establish possible uses without tracing a ship’s route. Thermoluminescence, when applicable, concerns the firing of ceramic material rather than its final deposition. The central question would remain whether the complete archaeological context demonstrates an arrival from the Roman world before 1492.

For Guanabara Bay, no published evidence currently satisfies those requirements. The reported amphorae have prompted preliminary expert assessments, incompatible accounts and unresolved questions about origin and recovery. No evidence yet demonstrates that a Roman ship crossed the Atlantic and sank near Rio de Janeiro. The disputed pottery remains an object of investigation, not an established record of Roman landfall.

Archaeological evidence can change the accepted history of long-distance navigation. The Norse remains in Newfoundland show how an unexpected voyage was confirmed through excavated structures, objects and independent dates. The Brazilian amphora claims would need the same attention to a find’s context, whatever chemical tests reveal about the vessels themselves. Even a genuinely ancient ceramic vessel cannot, by itself, date its journey across the Atlantic.