Carta Universal de Juan de la Cosa

A single manuscript sheet, roughly a meter high and just under two meters wide, made of two pieces of calfskin parchment sewn together and backed with canvas. It is not symmetrical: the seam falls near Italy and North Africa, and the western half is narrower because it follows the natural shape of the animal's neck. Down that seam, the chart carries the whole of the known world as a Castilian pilot understood it at the turn of the sixteenth century, from the Caribbean in the west to the Arabian Sea in the east.

The New World is painted in green. Everything else is left uncolored parchment, decorated in ink and watercolor in the ornamental idiom of the Majorcan portolan school: wind roses, flags, ships under sail, African kings enthroned under canopies, and legendary figures placed where medieval geography expected them, Prester John in Africa and the Three Magi in Asia. A wind rose set in the middle of the Atlantic carries a small image of the Virgin and Child. It is, in other words, both a working navigational chart and a piece of court decoration, and the two purposes sit on the same sheet without much tension between them.

Structurally the object is two charts joined into one. The Old World portion follows the conventions of ordinary Mediterranean portolan charts of the period: the African coastline reflects the most recent Portuguese reconnaissance, with a note near the Cape recording that the coast was reached under King John II of Portugal, a reference to Bartolomeu Dias's voyage of 1487-1488. Asia is drawn only as far as the Arabian Sea, cutting the continent off before China or Japan, with a note crediting King Manuel of Portugal for the coast beyond, a reference to Vasco da Gama's voyage of 1497-1499.

The American portion is drawn at a different, larger scale than the Old World, a discrepancy some historians estimate at roughly 1.4 to 1, which stretches the apparent size of the new discoveries relative to everything else on the sheet. The map supplies two unmarked dot-scales in the margins of its western half, thought to represent intervals of about fifty miles, but nowhere states the coordinate system or projection it uses; a 2010 cartometric study by Luis A. Robles Macias proposed a specific mathematical projection consistent with the Atlantic coastlines and found the chart's American latitudes considerably more accurate under that reading than earlier commentators had judged them by eye.

Where a cartographer with better information would have drawn a coastline, the chart instead places a large framed image of Saint Christopher carrying the infant Christ across a river, positioned squarely over the region that would become Central America. Below it sits the signature legend. The pairing is usually read as a visual pun: the saint who carried Christ across the water stands in for Christopher Columbus, who carried Christianity across the Atlantic.

The image also does cartographic work. By covering the isthmus rather than drawing it, de la Cosa avoided having to commit to a continuous coastline joining North and South America, a decision that would have foreclosed the possibility, which Columbus and others still argued for, of an open sea passage through to the Indian Ocean and the Spice Islands. The same evasion appears at the map's eastern edge, where the deliberate cutoff before China and Japan spares the chart from having to take a position on whether the lands Columbus reached were the edge of Asia or something new. Both choices are treated by cartographic historians as considered omissions rather than gaps in knowledge, a chartmaker declining to draw a conclusion the political and geographic argument of the day had not yet settled.

The Old World portion is easy to treat as filler beside the New World data that gives the chart its fame, but it carries its own information worth reading closely. Below the Indian peninsula a note credits King Manuel of Portugal with the coast beyond, tying that section of the chart directly to Vasco da Gama's 1497-1499 voyage, which means the Old World half of the sheet was current to within a year or two of the chart's own completion date, not copied wholesale from an older exemplar. Ceylon appears as a large triangular island under the name Trapobana, and two further large islands in the middle of the Indian Ocean, named Zanabar and Madagascoa, are drawn in a form close to their equivalents on Martin Behaim's globe of 1492, evidence of a shared stock of Ptolemaic and early-Renaissance geographic convention that both works were drawing on independently.

The chart's Asia stops at the shores of the Arabian Sea, cutting off before Indochina, China, or Japan. Historians read this as a second version of the same evasion at work over Central America: by declining to draw the Asian mainland any further east, de la Cosa avoided having to take a position on whether Columbus's discoveries were the edge of that same continent or something genuinely new. Both cutoffs, the American isthmus hidden under Saint Christopher and the Asian coast trimmed short, serve the same function, letting the chart carry the live geographic argument of 1500 without forcing an answer to it.

Among the Bahamas, which de la Cosa charts with unusual care and individually names, sits the island of Guanahani, the first land in the Indies that Columbus claimed to have sighted on his first voyage. It receives no special graphic treatment on the chart at all, no distinguishing mark, no flag, no cartouche, nothing to set it apart from any of its neighbors. For an island that would become the anchor point of the entire Columbian landfall debate in later centuries, its complete ordinariness here is itself a small piece of evidence: to de la Cosa, sailing the Bahamas in the 1490s and drawing them in 1500, Guanahani was simply one island among the group, not yet the symbolically loaded point of first contact it was later made into.

Read as a strict grid of longitude and latitude, the chart's longitudes for the Atlantic crossing are close to correct: the apparent longitudinal gap between the Iberian peninsula and Hispaniola works out to about 62 degrees against a true value of roughly 59, and Africa sits about 16 degrees from the northeast coast of South America against a true separation of about 17.75, both reasonably tight for the period. Its American latitudes, by the same simple grid reading, are considerably worse. Puerto Rico is drawn sitting on the Tropic of Cancer when it in fact lies about five degrees south of it. Hispaniola is placed at the same latitude as the Canary Islands, though the two differ by roughly eight degrees in reality. Cuba's north coast appears as far north as 36 degrees, some twelve degrees north of its true position.

Two things complicate any simple verdict of error. First, Columbus made comparable mistakes independently: his March 1493 letter describing his own return voyage claims Hispaniola lies 26 degrees from the equator, and he repeated a similar figure in a 1494 letter to the Crown, when the island in fact sits between about 17.5 and 20 degrees north, suggesting these latitude errors were common currency among Spanish navigators of the moment rather than a fault unique to this chart. Second, a 2010 cartometric study by Luis A. Robles Macias argued that reading the chart against a specific, non-obvious projection rather than a simple rectangular grid produces considerably more accurate latitudes across both the Old and New World, with errors generally under two degrees; if that projection is the right one to apply, most of the apparent latitude error is an artifact of reading the chart the wrong way rather than evidence de la Cosa measured badly. The two findings sit uneasily together and this record does not adjudicate between them; both are reported because a reader doing diligence on the chart's technical accuracy needs both halves of the argument.

The legend is unambiguous about who made the chart and when. Whether the surviving sheet is what it claims to be is a genuine, long-running argument among cartographic historians, not a settled footnote.

The challenge was first formulated by George E. Nunn in a 1934 monograph, which argued that Cuba could not have been circumnavigated before 1509, so de la Cosa could not have known in 1500 that it was an island, which the chart nonetheless shows correctly. Nunn's argument has been contested since: Ricardo Cerezo Martinez and others point out that Pietro Martire d'Anghiera, writing in 1501, already reported that many claimed to have sailed around Cuba, which undercuts Nunn's chronology. The counter-argument from the other direction is that the chart omits events that had happened by the time it was supposedly finished, most pointedly de la Cosa's own third voyage of 1501-1502, which reached the Gulf of Uraba and the Darien coast and left no trace on the map at all; if the sheet were retouched as late as 1508-1509, that omission is harder to explain.

A separate and more widely held position does not dispute the 1500 date so much as the hand: that the surviving sheet is a copy of a lost original rather than de la Cosa's own autograph work. The chief evidence is the state of the New World place names, many of which are illegible in a way consistent with a copyist unable to make out the original author's handwriting rather than with damage or fading. The Spanish military historian Hugo O'Donnell pushed this further, arguing that de la Cosa's own contribution may have amounted to no more than a working sketch, later completed and elaborated by a different, unnamed cartographer.

Material science has weighed in without fully resolving the question. In 1987 the Documentation Laboratory of the Museo del Prado analyzed the chart's pigments and found them compatible with an execution date of 1500, with no signs of later repainting. That result supports an early date for the physical object itself, but it cannot distinguish a single 1500 autograph from a faithful, contemporaneous copy made from the same period materials, so it narrows the dispute without closing it. A 2017 study by Joaquim Alves Gaspar approaches the same problem from the institutional side, asking whether the chart represents the first known Spanish padron real, the official royal master chart from which authorized copies were struck, or instead the last of an older, informal charting tradition before the padron real system was established; his framing treats the sheet's early depiction of the equator and the Tropic of Cancer, on a chart of this date, as evidence that observed astronomical latitude had already begun entering Spanish nautical cartography.

The dispute over Cuba is worth setting out on its own because it is the single most-argued detail on the chart. De la Cosa draws Cuba as an elongated island, pinched at two points, with numerous bays and a curved western tip enclosing a gulf scattered with small islands, an outline cartographic historians consider reasonably accurate for the period. The difficulty is that Columbus himself had made every member of his second voyage, de la Cosa included, swear in June 1494 that Cuba was not an island but a peninsula of the Asian mainland. A chart correctly showing it as an island in 1500, made by a man who had sworn otherwise six years earlier, is either evidence the chart postdates that oath by years, or evidence that de la Cosa simply knew better than the oath he had been made to take.

Historians divide on which reading is correct. Those who doubt de la Cosa could have known Cuba was insular in 1500 point to the fact that the definitive circumnavigation of the island is credited to Sebastian de Ocampo's voyage of 1508-1509. Those on the other side argue that partial Castilian voyages had already sailed around portions of Cuba's coast well before Ocampo's confirming circumnavigation, and that de la Cosa, present in the Caribbean through the 1490s, had ample opportunity to learn the truth independent of any single official voyage. Neither side has produced a document that ends the argument, and this record does not resolve it either; the honest description is that the chart's Cuba is the strongest single piece of evidence on both sides of the dating dispute, not proof for either one.

Set beside the corpus's other early world maps, this chart's place is specific and narrow. It predates the Cantino planisphere by two years and depicts markedly less of the American coastline, without that chart's more complete Portuguese data on Brazil. It predates the Waldseemuller world map of 1507 by seven years and, unlike Waldseemuller, does not draw the New World as a separate continent at all, let alone name it America; de la Cosa's chart is a full generation older than the name it eventually made possible. It has no genealogical relationship to the Piri Reis map of 1513, an Ottoman compilation which claims Columbus's own charts among its twenty-odd sources, though the claim illustrates how fast Spanish Atlantic data of the kind on this chart could travel once it existed.

What distinguishes this chart from all of them is that its maker was there. Alonso de Ojeda and Amerigo Vespucci sailed the northern coast of South America with de la Cosa in 1499-1500, and de la Cosa had crossed the Atlantic with Columbus twice before that. No other surviving map of the Americas from this early was made, even in part, by someone who had navigated the coasts it depicts. That is the basis for the claim, made across the literature on this object, that it is the earliest surviving European map to show any part of the Americas from direct exploration rather than from theory or secondhand report; the caveat attached by every serious source that makes the claim is the dating dispute itself, since a map retouched after the fact draws on more than one voyage's worth of direct knowledge, and to that extent stops being a single eyewitness snapshot.

De la Cosa was born in Santona, on the Cantabrian coast, at some point in the mid-fifteenth century; no source consulted for this record gives a firmly documented birth year, and estimates in the literature range from about 1449 to 1460. He later lived much of his working life out of the port of Santa Maria, near Cadiz, the port his chart names as the place of its making.

His documented role in Columbus's first voyage of 1492 was not that of an ordinary sailor. He owned and was master of the Santa Maria, Columbus's flagship, while Columbus himself held the overall command. The ship wrecked off the coast of present-day Haiti on the night of 24-25 December 1492. Columbus's own account of the wreck accused de la Cosa of abandoning the sinking ship rather than helping to save it, a charge de la Cosa disputed; he was compensated for the loss of the vessel in 1494. He sailed again on Columbus's second voyage, 1493-1496.

In 1499 de la Cosa joined Alonso de Ojeda as chief pilot on an independent expedition to South America, in which Amerigo Vespucci also took part. They sailed in May 1499 and charted a substantial stretch of the Atlantic coast of South America, from roughly the Essequibo river to Cape de la Vela. The return date is itself disputed: the traditional account gives June 1500, but the historian Demetrio Ramos has argued for a considerably earlier return, around November 1499. The chart is generally thought to have been drawn in the summer or autumn of 1500, after de la Cosa's return and before he sailed again, this time with Rodrigo de Bastidas, on an expedition some sources date to late 1500 and others to mid-1501.

Its evident quality and rich decoration lead historians to conclude it was made on commission for someone of real standing at court, quite possibly Bishop Juan Rodriguez de Fonseca, who administered Spain's American affairs on behalf of the Crown throughout this period. One direct piece of testimony survives from a pilot named Juan de Jerez, who stated in a 1513 legal deposition connected to the Columbus lawsuits that he had personally seen the chart de la Cosa made on that 1499-1500 voyage and the coastline it recorded, evidence that de la Cosa was carrying at least a working sketch of the coast home with him from the voyage itself.

The next documented moment comes in 1503, when de la Cosa presented two sea charts of the Indies to Ferdinand and Isabella at Segovia. Whether either of those two charts was this exact sheet, or a later, updated version of the same underlying chart, is not established in the sources consulted for this record; the presentation is documented, the identity of the object presented is not certain.

The two charts are then thought to have passed into the custody of Bishop Fonseca, where the historian and papal chronicler Pedro Martir de Angleria examined them in Fonseca's office. Writing on 4 December 1514, four years after de la Cosa's death, Martir recorded having seen a solid globe of the world and several sea charts there, among them work by de la Cosa, whom he names, alongside charts attributed to the Portuguese with Vespucci's involvement and one begun by Columbus and continued by his brother Bartolome. It is the earliest outside assessment of de la Cosa's cartographic work to survive, and it treats his charts as being kept among the most valued in Fonseca's collection.

After 1514, the trail goes dark. No document places the chart anywhere for roughly three centuries. It resurfaces already in commercial hands, in a shop in Paris, bought by the French naturalist and diplomat Baron Charles Athanase Walckenaer early in 1832. How it left Spain, whether through an ordinary sale, an inheritance passed down without record, or an outright theft from a Spanish archive, is not documented, and this record does not guess at it. A 2024 article by Luisa Martin-Meras Verdejo, published in the Revista de Historia Naval, argues for a specific theft hypothesis covering this gap, framed by the article's own title as the first of two thefts in the chart's history, the second being its wartime departure from Spain in 1936. That is a recent scholarly proposal, not yet a settled consensus, and this record flags it as such.

In the spring of 1832, shortly after acquiring it, Walckenaer showed the chart to friends visiting Paris, among them the German naturalist Alexander von Humboldt, who recognized its importance and was the first to bring it to scholarly attention. The Spanish scholar Ramon de la Sagra published the chart's western, American portion in 1837, and the Portuguese scholar and diplomat the Viscount of Santarem reproduced its African portion in 1842. For the rest of Walckenaer's life the chart stayed in his private collection in Paris.

Walckenaer died in 1852 and his library went to auction. Ramon de la Sagra alerted the Spanish Ministry of Marine to the sale and to the chart's importance, and the Ministry authorized La Sagra to bid on Spain's behalf. Spain won the chart for 4,321 francs. Whether that figure was a hammer price or an all-in price, and which auction house conducted the Walckenaer estate sale, are not specified in the sources consulted for this record. The chart entered the Museo Naval in Madrid in 1853 and went on display in the museum's Sala de Descubrimientos, where it has remained the museum's central exhibit ever since, with one significant interruption.

That interruption came during the Spanish Civil War. In November 1936 the chart was evacuated from Madrid to Valencia along with the rest of what came to be called Spain's Tesoro Artistico, following the Republican government as it relocated. From Valencia the collection moved to Barcelona, then to Girona province, and in February 1939, as the Republic collapsed, across the border to the headquarters of the League of Nations in Geneva. When the war ended, the League handed the Tesoro Artistico, this chart among it, over to Franco's government, and the chart returned to the Museo Naval in Madrid on 3 June 1939.

The episode is the subject of Luisa Martin-Meras Verdejo's 2024 study, which treats it as the second of the two departures from Spain named in the article's title, the first being the chart's undocumented exit from Spain before its reappearance in Paris in 1832. Whatever happened to the chart during its years abroad in the 1930s, it survived intact and came home; the surrounding record of exactly how it was held and transferred between 1936 and 1939, beyond the outline above, was not available in the sources consulted for this record.

The chart has never been restored. Its condition is generally described as good, with some flaking losses near the top of the sheet and damage attributed to an unspecified fire at some point in its history. Several facsimile editions have been produced, the first in 1892, timed to the fourth centenary of Columbus's landfall, and the most recent in 1992, for the fifth centenary. Between those anniversaries, a handmade copy made by the museum's own cartographer, Juan Pedro Suarez Davila, stood in for the original on public display from 1988 to 1992, while the original itself was kept off view.

In 2021 the Factum Foundation digitized the chart using a Lucida 3D scanner combined with composite photography, capturing both its surface relief and its color, and produced a facsimile using elevated-print technology for display at the Spanish Gallery in Bishop Auckland. Reporting connected to that project has suggested the Museo Naval may eventually substitute a facsimile for the fragile original on permanent display, echoing the 1988-1992 precedent; whether and when that happens is a live, unresolved question rather than a settled fact, and belongs among the open questions below rather than in the record of what has already occurred.

Scholarly attention to the chart has arrived in three concentrated bursts rather than as a steady stream, and each is tied to an anniversary or a rediscovery rather than to a change in the object itself. The first wave followed its 1832 resurfacing, when Humboldt, Ramon de la Sagra, and the Viscount of Santarem gave it its first scholarly description and reproduction through the 1830s. The second followed the 1892 fourth centenary of Columbus's landfall, which produced the chart's first facsimile edition and an accompanying study by the researcher A. Vascano. The third gathered around the 1992 fifth centenary, which produced the most recent facsimile and much of the technical and cartometric literature, including Alves Gaspar's later Padron Real argument, that this record draws on. Reading the historiography this way explains an otherwise odd feature of the literature: long stretches with almost nothing published, interrupted by short periods of intense, overlapping scholarship, all clustered around round-number Columbus anniversaries rather than around new archival finds.

De la Cosa's career after the chart continued to be spent at sea. New World Encyclopedia's account, the only source found for this specific voyage, places him in command of an expedition to the Pearl Islands and the Gulf of Uraba around 1504-1506; this record flags that detail as thinly sourced rather than independently confirmed, since no second source corroborating it was found.

His last voyage is much better documented. In 1509 he sailed once more with Alonso de Ojeda, departing Santo Domingo on 10 November 1509 with roughly two hundred colonists aboard three ships, bound for the coast of present-day Colombia. On arrival, de la Cosa advised against landing at Cartagena and argued for the more peaceable coast of the Gulf of Uraba instead, on the grounds that the local population used poisoned arrows extensively and effectively. Ojeda overruled him. The party pushed inland to the village of Turbaco, roughly twenty kilometers from Cartagena, where they were attacked by the Indigenous inhabitants. De la Cosa was struck by poisoned arrows and died of the wounds; most surviving accounts agree the attack was a rout for the Spanish party and that de la Cosa did not survive it. Ojeda escaped and reached a passing expedition, which returned with him to Turbaco to retaliate. De la Cosa died on 28 February 1510. This record found no documented earlier ambush at Turbaco that de la Cosa survived before this fatal one; every source consulted describes a single engagement there, in late 1509 or very early 1510, that killed him.

The largest unresolved question is the oldest one on this record: whether the sheet in the Museo Naval is de la Cosa's own 1500 work, a faithful copy of it made later, or, as Hugo O'Donnell has proposed, a sketch by de la Cosa finished by another hand. A century of argument over Cuba's insular coastline and the state of the chart's toponyms has not settled it, and the 1987 pigment analysis, while informative, addresses only the materials, not the authorship of the drawing itself.

Underneath that larger dispute sit several smaller, equally open gaps: the roughly three centuries between Pedro Martir's 1514 sighting and Walckenaer's 1832 purchase, across which no document places the chart anywhere; the exact circumstances of its departure from Spain that Martin-Meras's 2024 study frames as a theft; and the precise handling of the chart during its 1936-1939 absence from Spain, beyond the broad outline of Valencia, Barcelona, Girona, and Geneva. None of these gaps should be filled with a guess, and none are.