Our theme for the month of June is “Top Ten.”
The Imago Mundi, a fifth-century BCE Babylonian illustration accompanied by cuneiform script labels on a clay tablet, is considered the earliest world map. If you saw it, you might not think of it as a map of the world; for one, it is symbolic, drawn in neat circles and lines rather than an exact replica of actual landforms. For another, its contents do not contain the “world,” but the Euphrates, the cities of Babylon and Susa, Assyria, and geological landmarks like a mountain, a swamp, and an ocean. The world was much smaller for the map’s creators, much more locally bound. Besides trade routes and inferences of what may lie in other lands, the Imago Mundi captures the entire Babylonian world.
Other societies created maps of different forms. Petroglyphs on Mohave Rock depict the Colorado River and eight mountain ranges. Polynesian peoples mapped islands, currents, and wind patterns using grids of palm strips and shells or stones. Early Han Dynasty silk maps detailed principalities, military sites, and geographical features. The very oldest known map, found in the modern Czechia, is a mammoth tusk carved with mountains, rivers, and valleys.
In the ancient world, different societies’ maps converged, presenting more comprehensive representations of northern Africa and southern Eurasia. Zhang He’s Navigation Map, a set of Ming dynasty military charts, represented Persia, Arabia, southern Asia, and east Africa. In 1500, the Juan de la Cosa Map first indicated parts of the Americas in a map that also included Eurasia and Africa; two years later, the Cantino planisphere provided a more comprehensive outline of the southern American coast. (A fifteenth-century document called the Vinland map may have been the first European map to include parts of the Americas; however, the authenticity of the Vinland is highly disputed.) The Hendrik Hondius map, published in 1630, was the first substantially available map to include parts of Australia. The earliest map that looked more or less like a modern world map—certainly less around the north pole and the nonexistent Antarctica—was Samuel Dunn’s A General Map of the World, or Terraqueous Globe, published in 1794 (pictured above).
In the second century, Marinus of Tyre introduced the concept of the map projection, a mathematical and visual tool for accurately representing the surface of a sphere in a flat format. His equirectangular projection is still used today, though it does not appear here (beyond a nod in #5). Ptolemy, who approved of the equirectangular projection for local use, thought it too inaccurate to represent the entire world. He developed three new projections, and mapmakers haven’t given it a rest ever since.
As mathematicians, artists, and geographers have followed Ptolemy’s lead, many map projections have missed the mark. Some are worthy of critique (Mercator, HEALPix), others utterly confusing (Collignon, anything retroazimuthal), and some demand outright shaming (Gall-Peters, gnomonic).
But some are works of art.
A favorite among flat-earthers, a blatant distortion of the southern hemisphere, and the choice projection for the emblem of the United Nations. The azimuthal equidistant is flawed—it creates an alarming portrayal of Antarctica, for a start—but I delight in it nonetheless. It dates back to eleventh-century Iranian scholar Abu Rayhan al-Biruni, though its earliest form would look little like the image above. An azimuthal equidistant projection can center about any point on the Earth, with the edge of the map spreading around the antipodal (opposite) point of the center. In the commonly-used version of the projection, the map is designed about the North Pole and stretches to the South Pole, with distortion fanning the southern hemisphere.
By Tobias Jung [CC BY-SA 4.0], via map-projections.net
9. Ginzburg V
Sometimes you have to defer to a higher authority, even if that authority says that a vaguely Grecian urn-shaped projection is the best thing after a globe. In a presentation at the twentieth International Cartographic Conference in 2001, Charles University professor Richard Capek provided a list of the 100 best map projections according to a complex formula for evaluating distortion. Though some on this list (#3) did not yet exist for Capek’s review, it’s fair to say that his number one choice should have a spot in any top ten list.
The Van der Grinten has a few variations, most of which differ primarily in how much it looks like Alaska is fleeing the bondage of the Americas. In terms of distortion, it’s not a great map; it’s basically the Mercator in a circle. The National Geographic Society even used the Van der Grinten as its preferred projection for sixty-six years, although it is useful neither for navigation nor for presenting land masses in equal or relative area. As is the apparent trend with circular projections, Antarctica looks entirely absurd, as if it might soon swallow up the remaining continents. (The truncated version of the Van der Grinten removes the top and bottom of the circle around the eightieth parallel, creating an only slightly less alarming Antarctica.)
Now things are starting to get interesting. The “Lee conformal world in a tetrahedron” is a delightful, mathematically complex projection. It preserves area with almost no distortion of land masses, and while a bit peculiar to view on its own, it can be tessellated to create an array of complete maps so (nearly) perfect it draws a tear to my eye. Also, look at Antarctica: so complete. So respected.
6. Nicolosi Globular
As the azimuthal equidistant projection can be split in two for an equal-area capture of both hemispheres, al-Biruni is considered the originator of this projection, too. Its name comes from Giovanni Battista Nicolosi, who developed a corollary to al-Biruni’s projection—whether or not Nicolosi knew of the predecessor—as a revision of another map. Once a widely used projection (see Terraqueous Globe!), we rarely use it today, which feels almost unfathomable: its distortion is minimal, isolated to the edges of each circle and most evident in northwest Africa and northeast Russia. (A similar projection, the Mollweide Hemispheres, resolves the distortion even more cleanly.) It does, however, separate New Zealand from Australia, which is a minor cartographic sin.
“Tripel” refers to creator Oswald Winkel’s triple-pronged approach to reducing distortion, which considered area, direction, and distance. It’s also known as “Winkel III,” and it is the arithmetic mean of Marinus of Tyre’s equirectangular projection (expanded, of course, to globe size) and the Aitoff, an ovular projection from 1889. Capek ranked it in ninth place, with a distortion score just above the minimum value to be considered ideal in his evaluation. The Winkel tripel is quite widespread, having become the National Geographic Society’s projection of choice in 1998. (Winkel tripel replaced Robinson, which Capek ranked fourth.)
4. Peirce Quincuncial
When tessellated, the Lee conformal map is absolutely superior to the Peirce quincuncial, but as a standalone map, this one is more coherent. Despite slight but obvious distortions of northwest Africa, the eastern coast of Brazil, and the Malay Archipelago, most distortion in this map falls in the oceans, minimally disturbing land masses. For a projection first designed in 1879, that’s quite a feat.
By Tobias Jung [CC BY-SA 4.0], via map-projections.net
3. Natural Earth II
The Natural Earth series can be considered both an updated Robinson and a precursor to the Equal Earth projection developed in 2018. The latter is an equal-area map, but both Natural Earth and Robinson are compromise projections, meaning some distance and land area sacrifices were made to produce a map with less distortion overall. The Equal Earth projection is mathematically excellent, but its vertical distortion is off-putting. While Natural Earth II maintains horizontal distortion at the northern and southern extremes, for a north-south oriented map with no significant disruptions to the oceans, it’s an excellent option.
It’s beautiful, isn’t it?
Other versions of this map include Antarctica separately, in its own mini-map between the lower wings of the butterfly. I prefer that version, but images of it are uncommon. The Waterman Butterfly was developed from Bernard Cahill’s “Butterfly Map,” an octahedral projection that also paved the way for the Cahill-Keyes projection (shaped like the butterfly, with the wings each rotated ninety degrees) and the decidedly less glamorous Cahill-Concialdi Bat. All technicalities aside, this map is stunningly, almost unbelievably gorgeous.
The Dymaxion, sometimes known as the Fuller Projection, has little distortion of either land masses and oceans, preserves all land masses, and folds into an icosahedron. Like all maps, the Dymaxion can represent the spherical earth from any orientation, meaning it also produces a reasonably comprehensive map of earth’s ocean. Since its inception, Dymaxion and the Cahill family of maps have been at odds, with ardent supporters of each penning flurried op-eds and studied comparisons in favor of their preference. I waver between the two as well, but at the end of the day, proportionally and practically, Dymaxion is simply superior.
Or, you know, just use a globe.