top of page

Portolan Chart of the world by Juan de la Cosa, 1500; detail: Western Hemisphere only

Sultan Husayn Bayqara 1470–1506
Stacked Wooden Logs


The original parchment of this map or chart, a piece of ox-hide measuring 37.5 x 72 inches (96 X 183 cm), superbly illustrated in ink and water colors, was found in 1832 in a shop in Paris by Baron Walckenaer, a bibliophile and the Dutch Ambassador, and was brought to the attention of the world the following year by Alexander Humboldt, the famous German scholar. Upon the death of Baron Walckenaer in 1853 the map was purchased by the Queen of Spain, and though greatly deteriorated, is now the chief treasure of the Museo Naval in Madrid. Notwithstanding several large holes, the map may be said to be in a good state of preservation. There is, however, a regrettable gap on the northern coast of Brazil, where a piece two inches wide, containing names, has been torn off. Originally a manuscript map, it was never engraved or printed until recent years, and therefore may have exercised little influence on the cartography of the sixteenth century, except for those privileged few who were allowed to study it.

The name of the maker of the map is in the legend under the picture of St. Christopher, at the left, which reads: Juan de la cosa la fizo en el puerto de S: ma en año de 1500 [Juan de la Cosa made it in the port of Santa Maria in the year 1500]. Some scholars accept this date, others do not. One of the latter is George E. Nunn. In The Mappemonde of Juan de la Cosa he held that the map is a copy and not an original work of La Cosa, and that it probably dates from about 1508 instead of 1500. He argued, (a) that the insular nature of Cuba shown on the map does not fit the exploration record as of 1500, (b) that the map makes of South America the peninsula of southeastern Asia believed in by Martin Behaim (Slide #258), and that no navigator or explorer is known to have held this concept before Columbus' fourth voyage, (c) that it shows evidence of exploration of the coast west of Cabo de la Vela to the vignette of St. Christopher and beyond it west of Cuba, by which exploration Nunn meant the mainland discoveries of Rodrigo de Bastidas and La Cosa in 1501-02 and of Columbus in 1502-03, (d) that it shows evidence of both of the Cabot voyages, (e) that it shows evidence of exploration of southern Brazil after 1503 at least, and (f) that it shows the island group in the South Atlantic Ocean which corresponds to the Tristan da Cunha islands. Nunn concluded that some of this information cannot have been available to La Cosa at the earliest before 1504, and some of it probably not before 1507 or 1508.
Another theory claims that this is a copy by a draftsman who was unable to decipher La Cosa's original lettering. This theory is born out by the many names, even on the South American coasts that had been explored by Cosa himself, that are unintelligible and meaningless.

As far as its publication date, not only does the map carry on it the date of 1500, but it has features which would have been utterly different had it been compiled in 1508. Thus in 1501-02 Cosa explored the Gulf of Uraba and the coast of Darien, yet these are not shown on his map nor are there any names on the hypothetical coasts drawn thereon. Again Columbus in 1502-03 explored from Yucatan to Darien and his coasts and names are not shown on the Cosa map. The weakest feature of the Cosa map is its grotesque representation of Brazil south of the equator. This was excusable in 1500, but by 1508 Amerigo Vespucci was pilot-major of Spain and his great voyage from 5° south to 42° south would have been available for La Cosa to correct his outlines of South America. There is no cause to doubt that the Cosa map was compiled in 1500. La Cosa returned to Spain with Alonso de Ojeda [Hojeda] in July 1500 and left again with Bastidas for Colombia in October of that year, which means that the map was drawn between these months.

La Cosa has been called "the most expert mariner and unrivaled pilot of his age." He was the owner and mate of the Santa Maria, the flagship of Columbus on his first voyage, and the official cartographer and captain of the Niña on the second voyage. With others he signed the famous affidavit, demanded by Columbus, that he believed Cuba to be a part of the mainland of Asia. In 1499, he was chief pilot of Ojeda's expedition along the northern coast of South America. On this voyage, he was associated with Amerigo Vespucci. Upon his return from this expedition in 1500, he made his famous marine chart for Ferdinand and Isabella. Later he went on four other voyages to the new world and in 1509 was killed by the Indians in Venezuela, "pierced by more than twenty poisoned arrows."

This map is the oldest, now known, made since 1492, which shows the discoveries in the new world. It is an excellent example of the portolan charts which came into use among Italian sailors in the thirteenth and fourteenth centuries. The La Cosa chart displays compass roses and direction lines. The scale is given by a line of dots, unnumbered and unexplained; the distance between the points however is apparently intended to represent fifty miles. The northern tropic and the equator are drawn, but degrees of latitude or longitude are not indicated. The Italian portolan originated in the Greek periplus, a harbor-book or set of sailing directions describing harbors, shoals, distances, currents and winds, but containing no map. The Italian portolan was made on the model of the periplus but contained a map or chart showing the coast line and a few places along the coast, but, as it was intended only for seamen, it gave little information of the interior of even the most populous countries. After the invention of the compass, portolans became numerous. At least one hundred made before 1500 are still in existence. Modern maps developed from these portolans. In addition to Juan de la Cosa, some of the most distinguished makers of Renaissance portolan charts were Alberto Cantino (Slide #309), Nicolo de Canerio (Slide #310), Diego Ribero (Slide #332), Vesconte de Maiollo (Slides #315/329) and Battista Agnese.

La Cosa represents upon his map only the line of the Equator and the incorrectly located Tropic of Cancer but no other lines of latitude, and no lines of longitude except the liña meridional crossing the tip of Brazil, which is thought to be the line of demarcation established by the Treaty of Tordesillas of 1494. This line is also plainly indicated with a suitable inscription on the Cantino map made in 1502 (Slide #309). La Cosa, who was Basque by birth, placed most of the New World under Spanish juristiction. The lack of parallels and meridians, and the apparent compilation of the map from separate portolan charts of the Old and New Worlds with differing scales but no attempt to reconcile them, reveal the extreme difficulties involved in constructing a world map, or planisphere, during the early period of the Age of Discovery.

The chief interest in the map is its delineation of the coast line of the New World so recently discovered and so little known at that time. North America, which in most early maps (up to 1506) normally consisted of separate islands, only later fusing into a cohesive whole, is shown on the Cosa map as a solid landmass extending far into the North Atlantic. In the west are the discoveries of Cabot in the north and of Columbus and the Spaniards in the West Indies and along the northeastern coasts of South America.

What is shown of South America also suggests that is propably a continental landmass. Off the coast of Brazil, the island with the legend, Ysla descubierta por portugal [Island discovered for Portugal], indicates La Cosa's belief concerning the location of the land discovered by Cabral in 1500. The news of this discovery was probably brought to Spain while La Cosa was still working on his map. On the coast of the mainland farther north is the legend: Este cavo se descubrio en año de mily IIII XCIX por Castilla syendo descubridor vicenttians. [This cape was discovered in the year 1499 for Castile, Vicente Yañez being the discoverer thereof.] This refers to the discovery made by the Spaniard, Vicente Yañez Pinzon, January 20, 1500 (1499 old style), three months before the Portuguese Cabral sighted the same coast on his way to India via the Cape of Good Hope. Mar Dulce commemorates the current of fresh water at the mouth of the Orinoco River seen by La Cosa far out to sea when he was with Columbus on his second voyage. Costa de perlas, or the Pearl Coast, on the northern coast of South America, was first discovered by Columbus on his third voyage in 1498 and had been visited again by La Cosa, with Ojeda and Vespucci, in 1499.
The Caribbean islands already show, in contrast to Columbus' ideas, clear hallmarks of the way they appear in much later maps. The Bahamas group is shown with some accuracy but necessarily on a small scale. It includes the island Guanahani, Columbus' first landfall, alternatively known as San Salvador and now identified with Watling Island. No special emphasis is given to this memorable locality. Haiti and Cuba are located north of the Tropic of Cancer. The latter island is now first known by that name. Columbus had called it Iuna. Its representation as an island, instead of a part of the mainland of Asia, indicates that La Cosa had changed his opinion since he signed the famous affidavit. The historian Henry Harrisse sees in this insular character of Cuba strong confirmation of the much disputed story of Vespucci's voyage along the coast of the mainland in 1497. He points out that there is no record of any other voyage prior to 1500, the date of the map, which had revealed the insularity of Cuba, and that La Cosa may have obtained his information directly from Vespucci, while they were together under Ojeda, in 1499, on the Pearl Coast. The later world maps of Cantino, and of Canerio, 1502, agree with La Cosa in putting water and a mainland west of Cuba; indeed these two maps are even more explicit than that of La Cosa in that they place more than a score of names on the supposed mainland, which may be identified as Florida.

The earliest and, for more than half-a-century, the most complete description of Cuba is the one which that celebrated mariner has inserted in his famous planisphere, designed during the autumn of 1500. La Cosa was considered in Spain as the greatest cartographer of his day, and the pilot best conversant with the West India seas. He had been, moreover, Christopher Columbus' chief pilot on several transatlantic expeditions, and even owned and was mate of the flag-ship during the memorable voyage in which the island of Cuba was discovered by the great Genoese. No seaman, therefore, could then make a more reliable chart of the Antillies than Juan de la Cosa; and his map of Cuba must be considered as embodying all that which was known concerning its ports, rivers, capes, and other sea-board localities, from the time of the discovery to the close of the year 1500. On his cartographical representation of Cuba, the great Basque pilot has inscribed as many as twenty-seven names of landings, estuaries, streams, harbors, headlands, towns, or hamlets. In Cantino's map, drawn in 1502, the coast line, alleged as aforesaid to be a duplicate of the island of Cuba, contains also numerous nomenclature. Here again, if that region in the Cantino chart is really Cuba, we should find among its legends and designations the identical names which are inscribed on the Cuba of La Cosa, especially as both maps were delineated within a year of each other. An analysis of such nomenclature by Harrisse has determined that there is not even a single matching place-name among the twenty-two. The fact, of itself, were it not supported by the other proofs which have been documented, is sufficient evidence that the northwest coast in the Cantino map and Cuba in La Cosa's map were intended to represent two entirely different countries.

Nomenclatures play such an important part in the identification of cartographical documents; they enable us to ascend so surely to the origin not only of names, but also of the configurations on which we find them inscribed, that no better means can be employed by critics to solve the numerous problems which are involved in every ancient map, chart, and globe, without a single exception. And even when the names are scarcely legible, or evidently corrupted by the inattention of cartographers, and their ignorance, oftentimes, of the language employed in the prototype, they still serve to indicate the source from which were borrowed important geographical averments. But La Cosa's chart was made before October 1500, yet Cuba is depicted therein as it is in reality, elongated, deeply indented (at Nipe, Nuevitas, Turiguana, Cardenas, Matanzas, etc.), depressed or strangulated in two places (Manzanilla-Jbara and Sabanilla-Jalibonico), the western extremity curved and forming at its southwest end a very large bay , which is studded with islands.

La Cosa avoided the problem of the possible existence of straits in Central America that might provide a sea link to the Far East. The vignette at the left of the map represents St. Christopher, with a pine tree as a staff, carrying the infant Jesus over the deep water, as Christopher Columbus carried the knowledge of Christ over the sea to the natives of the newly discovered lands. Below the drawing of St. Christopher in the neck of the skin, is the inscription: Juan de la Cosa la fizo en el puerto de s. maria en ano de 1500. It has been suggested that the face of the saint is a portrait of Columbus. It is a matter of argument whether La Cosa intended to represent the mainland behind Cuba as the eastern coast of Asia. If so, it seems strange that he placed no Asiatic names upon it.

While Cosa could draw upon the voyages of Columbus and Spanish explorers for the Antilles and much of South America, no Spanish ships had by 1500 visited Central and North America. The Cosa map represents them as a continuous landmass extending to the Arctic, which was cosmographically correct, but the actual coastlines are clearly hypothetical in outline and without nomenclature, proving they were not based on charts of actual voyages which had explored such coasts. The one exception is the east-west coast in the northwestern portion of La Cosa's map with English flags shown at about 52° north. These English coasts derive from the voyages of John Cabot from Bristol in 1497 and in 1498. The map historian R.A. Skelton wrote: "The only map which unambiguously illustrates John Cabot's voyage of 1497 . . . is the world map signed by Juan de la Cosa and dated 1500". A study of the Cabot voyage of 1498 by Williamson reveals that this voyage coasted from Greenland to New England and that Cabot's ship went down off Newfoundland, leaving the consort vessel to bring news of his end to England in the early summer of 1499. The English coasts on the Cosa map consequently represent the discoveries made in the 1497 voyage. Cabot left Bristol on 2nd May 1497 in a little ship called the Matthew, with a crew of only eighteen men. Information as to his discoveries in English records is meager but certain statements provide a skeleton by which they may be reconstructed.

1. It is known that he tacked against the westerlies for 54 days and made landfall on 24th June (Bristol records). Sebastian Cabot, in his map of 1544, indicated this landfall as Cape Breton island: This land was discovered by John Cabot the Venetian and Sebastian Cabot, his son, in the year 1497 on the 24th June in the morning, to which they gave the name "Land First Seen". Since Sebastian had no motive for lying about the location of the landfall there is no reason to doubt this statement.

2. Lorenzo Pasqualigo to his brothers in Venice on 23rd August 1497: That Venetian of ours who went with a small ship from Bristol to find new islands has come back and says he has discovered mainland 700 leagues away, which is the country of the Grand Khan, and that he coasted it for 300 leagues and landed and did not see any person, ... and on the way back he saw two islands...

Unfortunately, those cartographical data are not sufficiently precise to enable one to locate the landfalls with adequate exactness. Nor is the kind of projection adopted, without explicit degrees of latitude, of such a character as to aid someone much in determining positions. One is compelled, therefore, to resort to inferences.

Much attention has been attracted to La Cosa's representation of the north-eastern coastline of America. The principal features are (I) a prominent cape, Cavo da Yngleterra, about 1,300 mlles from south-west Ireland, and approximately in the same latitude; (2) to the west of this Cape, an extent of coastline, running about due west for approximately 1,200 miles: a number of features along this coast are named, and this is the only portion of the North American coastline on which names occur; (3) beyond this coast, a stretch without names continues for another 700 miles, forms a bay, Mar descubierta por Yngleses, and then turns southwards.

The Cavo da Yngleterra is shown in about 56° N. latitude. Since, however, the latitudes of many places in Europe are out by several degrees (Land's End, for example, is shown 4.5° too far north) the Cavo may be assumed to be not further north than 51° 30' N., which would put it in the neighborhood of Belle Isle Strait. On the other hand, the 1,200 miles of explored coastline is in all probability southern Newfoundland or Nova Scotia, so that the Cavo de Yngleterra must have lain further south, and Cape Race at once suggests itself, but as nothing more than a possibility. J.A. Williamson, however who credits this charting to the Cabots in 1497-8 believes that the Cavo was Cape Breton, while G. E. Nunn identifies it with Cape Farewell in Greenland.
The northwestern portion of La Cosa's map sets forth twenty inscriptions, seven of which are the names of capes, whilst one refers to a river (r° longo), another to an island (isla de la trinidad), and a third to a lake (lago fore ?). Although many of those designations convey no meaning to the scholar (apparently on account of imperfect transcriptions), and are not to be found on any other map, they must be considered as proving that the coast had been actually visited before 1500. On the other hand, the northernmost names represent certainly the points marked by John Cabot during his first voyage, whether we place them on the north coast of Labrador or on the east shores of Newfoundland. But as the line of English flagstaffs covers a space by far too extensive for the Cabot voyage of 1497, which lasted only three months, the legends placed further south necessarily apply to the expedition of 1498.

When preparing himself to return to the newly discovered regions, John Cabot told Raimondo di Soncino that his intention was to pursue the undertaking as follows:- "From the place already possessed [discovered] he would proceed by constantly following the shore, until he reached the east, and was opposite an island called Cipango, situate in the equinoctial region."
All that is clear in this vague description, and which must be retained just now, is that John Cabot's ultimate objective, when he set out from England in 1498, was an equatorial or southern region situated south of the point reached by him in 1497. To this interpretation must be added the fact that the line of British flags in La Cosa's map, corroborates such an intention, as it indicates plainly a southward coasting.

How far south then did John Cabot go in 1498 ? Taking the distance from the equator to the extreme north in La Cosa's map as a criterion for measuring distances, and comparing relatively the points named therein with points corresponding for the same latitude on modern planispheres, the last English flagstaff in the southern direction seems to indicate a vicinity south of the Carolinas.

The five English standards in the north and the inscriptions, mar descubierta por yngleses [sea discovered by the English] and cavo de ynglaterra [Cape of England, corresponding with a point almost as high north as the entrance to Davis Strait], indicate the explorations of the Cabots in 1497 and 1498, probably either Newfoundland or Nova Scotia, i.e., 50°-53°N. It is believed that La Cosa drew this portion of the coast from Cabot's map, now lost, which Pedro de Ayala, the Spanish Ambassador at London, sent to King Ferdinand.

In latitude the Old World portion of the map, it extends from the Scandinavian peninsula to the southwards of the African continent. The African coast as far as the Cape of Good Hope is represented with fair accuracy, from Portuguese sources. The eastern coast however seems to be entirely imaginary. Well out in the Indian Sea, almost in the center, are two large islands, Zanabar and Madagascoa, as on Behaim's globe. La Cosa shows the coast line of Africa and the Cape of Good Hope with considerable accuracy. Only the previous year Vasco da Gama had returned from the first voyage to India. The inscription on the southern coast of Asia, tierra descubierta por el Rey don Manuel Rey de portugal, refers to Vasco da Gama's voyage to Calicut on the western coast of India from which he had returned to Portugal the previous year. The numerous names on the eastern coast of Africa cannot be attributed to da Gama.

Ceylon/Sri Lanka appears in the large triangular island of Trapobana. The non-peninsular coast of India is from the Ptolemaic cartographic tradition; the outline of the Asian coast, however, is no improvement on that of the Catalan Atlas of 1375 (Slide #235). In the various kingdoms into which Asia and Africa are divided by La Cosa are pictures of the reigning sovereigns, some seated on thrones. In eastern Asia are two strange figures, one that of a man without a head, with his eyes in his chest and his mouth in his stomach, and the other with the snout of a dog. R. Got and R. magot suggest the Biblical Gog and Magog. At Babylon is seen the tower; near thee eastern shore of the Red Sea, which on the original parchment is properly colored red, is the Queen of Sheba with drawn sword; and crossing Asia towards Syria are the Three Wise Men, bearing gifts.

The planisphere does not extend eastward beyond the northern border of the Arabian sea, omitting, therefore, Hindostan, the Malay Peninsula and China. Yet the La Cosa map depicts the Rio Ganges, but where one would expect the Indus River. All of the Asian portion of the Cosa map is extremely inferior, particularly compared with other contemporary world maps such as the Cantino chart, although there is only a difference of eighteen months between the two. However, it should not be inferred from such an important omission that La Cosa considered the continent, depicted by him west of Cuba, as identical with the east coast of Asia, and, consequently, could not have added the latter to his eastern configurations without repeating what he had already marked on the same map. Had such been his geographical conception of the world, he would not have omitted the India intra and extra Gangem, and especially the Molucca regions, nor to name Cathay, Mangi and all the cities and provinces rendered famous by Marco Polo and which had figured in all mappamundi of the fifteenth century.

The map in fact has every appearance of having been put together from at least two sections: the western portion comprising the American discoveries and perhaps the West African coast having been joined to a portion of a world map resembling those of fifty years earlier which display the influence of Ptolemy. If we use the distance between the tropic and the equator to obtain a scale of degrees, and apply this to the map, we find that in the western section, though there are discrepancies, the general picture is not wildly inaccurate. The newly discovered lands are placed in fair relationship to those of western Europe. The longitudinal difference between the Iberian coast and Hispaniola is apparently about 62°, instead of 59°, and between the African coast and the northeast coast of South America approximately 16°, instead of 17.75°. For a reason which has never been satisfactorily explained, Hispaniola and Cuba are placed well to the north of the tropic; the north coast of Cuba being shown in approximately 36° N., some 12° too far north. Whatever the reason for this, it would appear that the Central and South American portion is on a larger scale than the rest of the map.

The representation of Africa is distorted by the excessive lengthof the Mediterranean. The general shape of the western coastline is good, though, in relation to the west-east extent of the Gulf of Guinea coast, the coastline southwards to the Cape is too short. This was a characteristic of early Portuguese charts of this region: owing to adverse sailing conditions, it was usual to underestimate distances run.

LOCATION: Museo Naval, Madrid, Spain

Crone, G. R., Maps and their Makers, pp. 84-86
Cummings, W.P., The Southeast in Early Maps, p.64
*Cummings, Exploration of North America, pp. 53ff, Figures 52, 53
Davies, A., "The 'English Coasts on the Map of Juan de la Cosa", p. 26
*Fite, E., and Freeman, A., A Book of Old Maps . . ., p. 11, no. 4
*Ganong, W.F., Crucial Maps in the Early Cartography . . ., pp. 8-43; 142-175
*Goss, J., The Mapmakers Art, pp. 72-73, Plate 3.16 (color)
*Harrisse, H., The Discovery of North America, pp. 42-3, 89, 98, 256, 412-415, 711-12
Hoffman, B., Cabot to Cartier, p. 87
*Humble, R., The Explorers, pp. 12-13 (color)
*Nebenzahl, K., Atlas of Columbus, pp. 30-33, Plate 10 (color)
*Nordenskiöld, A.E., Periplus, p. 149, Plates XLIII, XLIV
*Nunn, G., The La Cosa Map and the Cabot Voyages
Roukema, E., "Some Remarks on the La Cosa Map", p. 38ff
*Shirley, R.W., The Mapping of the World, pp. xxii, xxiii, Plate 5 (color)
Skelton, R.A., Looking at an Early Map, p. 16
Williamson, J.A., The Cabot Voyages. . ., pp. 59ff
*Winsor, J., Narrative and Critical; History of America, vol. II, pp. 106-109.
Wolf, H., Circa 1492, "Manuscript World Maps", p. 45
*Wroth, L.C., The Voyages of Giovanni da Verrazano, p. 37, Plate 2

Mughal Library

Rate This BookDon’t love itNot greatGoodGreatLove itRate This Book

Your content has been submitted

Post Comment
Ratings & Review
Click To Close Comment Box
Click To Post Your Comment
Show Reviews

average rating is 5 out of 5

Ismail Mazari

average rating is null out of 5

Very good information.

average rating is null out of 5

Shah Sharaf Barlas


The Mughal Maps takes you back to a historical time of different eras, where you can see historical events based on the map and location of the event with all the details of the area of that time. You can add your maps to The Mughal Library here.

Mughal Library brings readers of our history and related subjects on one platform. our goal is to share knowledge between researchers and students in a friendly environment.

bottom of page