Some of the best-known misleading maps in history are the Mercator world map, which makes Greenland look as big as Africa when Africa is about 14 times larger; the Mountains of Kong, a range that appeared across West Africa on about forty maps from 1798 to 1892 but does not exist; the Island of California, drawn as an island on hundreds of 17th- and 18th-century maps; Sandy Island in the Coral Sea, a phantom island that French charts dropped in 1974 but Google Maps kept until 2012; Agloe, New York, a fake hamlet planted as a copyright trap in the 1920s; and the election maps of the German Empire, where drawing constituencies by land area hid where the votes actually were. Each one fails differently. One distorts through projection, two through copying errors, one through a database, one by design and one through the choice of what to scale. Taken together, they show the main ways a map can mislead, and in five of the six cases nobody set out to deceive.
Greenland and Africa on Mercator’s 1569 chart
Mercator’s projection made its debut on his 18-sheet world map of 1569, whose Latin title promised a new and augmented description of the Earth corrected for the use of sailors. The title states its purpose. On a Mercator chart, a line of constant compass bearing is a straight line, so a navigator can rule a course and hold it. By the 18th century it was the standard map for navigation at sea.
To keep those bearings true, Mercator had to stretch the map more and more towards the poles. The result is the best-known distortion in cartography. Greenland, 2,166,086 km², is drawn about the same size as Africa, which covers about 30.3 million km². Greenland is actually closer in size to the Democratic Republic of the Congo, at 2,345,408 km². Antarctica turns into an endless white band along the bottom of the map. Nothing on the map is wrong for its original purpose. It misleads when it is taken off the ship’s chart table and hung on a classroom wall as a picture of how big the countries are.
Arno Peters’ 1973 map and Boston classrooms in 2017
The Greenland gap drove one of the 20th century’s fiercest map arguments. In 1973 the German historian Arno Peters presented a cylindrical equal-area map that was essentially the projection the Scottish clergyman James Gall had described in 1855. Peters argued that the Mercator made wealthy Europe and North America look too large next to Africa and South America. Groups such as the National Council of Churches took up his map, and Boston Public Schools began phasing in Gall–Peters maps in March 2017. In 2024 Nebraska passed a law requiring its public schools to display Gall–Peters, a similar equal-area projection or AuthaGraph. Critics pointed out that, like every equal-area projection, Gall–Peters distorts most shapes in turn. A different projection does not remove distortion. It only changes which property is given up.
The Mountains of Kong across West Africa, 1798–1892
In 1796 the Scottish explorer Mungo Park, travelling along the Niger, noted some very distant mountains to the south and was told they lay in a kingdom called Kong. In 1798 the English cartographer James Rennell turned that remark into a solid west-to-east mountain range on a map of Park’s journey. The range suited Rennell’s own theory that the Niger dried up inland at Wangara instead of reaching the sea.
The error spread by copying. Aaron Arrowsmith made the Mountains of Kong a main physical feature of West Africa on his 1802 map, and other publishers followed. Johann Reinecke drew the Gebirge Kong in 1804. In 1805 the London engraver John Cary joined them to the equally imaginary Mountains of the Moon in Central Africa, producing a chain that crossed most of the continent. Explorers including René Caillié and Hugh Clapperton kept drawing the range on their own maps after their journeys. The historians Thomas Bassett and Philip Porter have identified forty maps, published between 1798 and 1892, that show it.
Cartographers began to drop the range after the French explorer Louis Gustave Binger travelled through what is now Côte d’Ivoire in 1887–89 and established that it was fictitious. Even so, Bartholomew’s Oxford Advanced Atlas still listed the mountains in its 1928 index, at 8°40′N, 5°0′W. The lesson is that a map gains authority every time it is copied. Once a feature is printed by a respected publisher, leaving it off looks like a mistake, so later mapmakers keep it in.
The Island of California on maps from 1622
The name California first appears in a novel. The 1510 romance Las sergas de Esplandián by Garci Rodríguez de Montalvo describes an island called California, close to the Earthly Paradise and ruled by Amazon-like women. When Spanish expeditions reached the southern tip of Baja California in the 1530s, the name and the idea of an island attached themselves to the real coast.
The evidence pointed the other way early on. In 1539 Francisco de Ulloa sailed to the head of the Gulf of California and reached the mouth of the Colorado River, and a 1548 Spanish map shows a peninsula. Yet from 1622, maps showing California as an island became common. The historian Dora Beale Polk blamed uncritical repetition. A 1626 Portuguese map got it right while a 1630 British map got it wrong. Stanford’s Glen McLaughlin Collection holds slightly over 1,000 maps of California as an island, the largest such collection in the world.
The Jesuit Eusebio Francisco Kino travelled overland from Sonora towards the Colorado delta between 1698 and 1706 and drew California as a peninsula in 1701. Even then, European mapmakers stayed divided until Juan Bautista de Anza’s expeditions of 1774–1776 settled the question. The hero image on this page, Joan Vinckeboons’ map of about 1650, shows the island in full. It is an example of a map that looks precise, with a detailed coastline and named bays, while getting the basic shape of the land wrong.
Sandy Island in the Coral Sea, reported 1876 and removed 2012
The whaling ship Velocity reported Sandy Island in 1876, in the eastern Coral Sea between the Chesterfield Islands and Nereus Reef, west of New Caledonia. At the time it was standard practice to add every reported hazard to the charts as a precaution. An 1881 German map and an 1895 British Admiralty chart both included it.
No one could find it again. Charts added the abbreviation “ED”, for existence doubtful. The French Hydrographic Service removed it in 1974, and the Australian Hydrographic Service did the same in 1985. Australian seafloor mapping put the minimum depth around the charted position at between 1,488 and 2,353 m. But the island had already been copied from paper charts into the World Vector Shoreline database, a global coastline dataset built by the US National Imagery and Mapping Agency, and digital maps kept it.
On 22 November 2012, scientists aboard the Australian research vessel Southern Surveyor noticed that their maps disagreed and sailed to the charted position. They found open ocean, with depths never less than about 1,300 m. Google Maps removed the island on 26 November, and National Geographic struck it from its maps on 29 November. This is the digital version of the Mountains of Kong problem. A database looks objective, but it can only be as accurate as the paper charts that were scanned into it.
Agloe, New York: a Catskills copyright trap
Some misleading maps are wrong on purpose. In the 1920s Otto G. Lindberg, founder of General Drafting, and his assistant Ernest Alpers made an anagram of their initials and placed a hamlet called Agloe at a dirt-road junction in the Catskills, where NY 206 meets Morton Hill Road, north of Roscoe in Delaware County. It was a fictitious entry: if Agloe turned up on a rival’s map, the rival had copied General Drafting’s work. Agloe was printed on the road maps General Drafting made for Esso.
Then the trap became real. In 1930 a business called Agloe Lodge Farms was incorporated and took over a fishing lodge at the site, renaming it Agloe Lodge. According to the cartographer Frank Brown, when Agloe later appeared on a Rand McNally map and General Drafting complained, Rand McNally said it had taken the name from Delaware County records, which by then listed a business called Agloe. The hamlet stayed on maps into the 1990s, briefly appeared on Google Maps, and was added to the US Geological Survey’s Geographic Names Information System as “Agloe (Not Official)” on 25 February 2014. John Green’s 2008 novel Paper Towns made it widely known. Trap streets and paper towns like Agloe show that commercial maps may contain small deliberate errors.
Berlin, Hamburg and Saxony in the 1898 Reichstag maps
The subtlest way a map misleads is by sizing everything by land area when the subject is people. After the 1898 German Reichstag election, Hermann Haack and Hugo Weichel published two maps of the results ahead of the 1903 vote. The first drew the constituencies at their true size, so the large rural districts dominated the picture. The second, the earliest known contiguous cartogram, resized each constituency so that the densely populated areas around Berlin, Hamburg and Saxony swelled up.
The two maps showed the same results and gave very different impressions. The cartogram was meant to make visible a controversial pattern: the mainly urban Social Democrats tended to win the popular vote, while the mainly rural Zentrum won more seats. The same effect appears on modern county-by-county election maps of the United States, where large, thinly populated counties fill most of the map. During the 2016 presidential campaign, The Washington Post, FiveThirtyEight and The Wall Street Journal all published block cartograms of the Electoral College, which size each state by its electoral votes rather than its land. Mark Monmonier’s 1991 book How to Lie with Maps gives particular attention to how colour on choropleth maps can mislead, and sets out to promote a healthy scepticism about maps.
Three questions to ask of any map, from Kong to Agloe
The six cases point to a short checklist for any map. Ask what the projection keeps and what it gives up. Ask where the data came from, and whether it was simply copied from an older source. Ask whether colour is being spread over land when the subject is people. Our guide to how map projections distort reality covers the first question in depth, the Mercator vs Peters projection comparison follows the 1970s argument in full, and the long story of how mapmakers got things right and wrong is in our history of cartography.