Image area a country gets, against its true share of the world's land
Globe Mercator Equal Earth

164 to 1 for
a new world map

On 4 September 2026 the UN General Assembly adopted the resolution “Correct the Map”, moved by Togo on behalf of the African member states: 164 votes in favour, one against, six abstentions. It does not ban Mercator; it calls on governments, schools, organisations and technology companies to use equal-area maps wherever relative size matters — naming Equal Earth in particular.

What is at stake can be measured. On the Mercator map, Europe and North America together take up 65 % of the land area shown; their true share is 35 %. The slider fades between them. The colour shows, for every country, how much image area it gets measured against its real share of the world's land — and how that distortion disappears along the way. Next to the two maps stands the globe: the one depiction without any distortion at all, and therefore the yardstick both maps are measured against.

What was decided

The resolution is titled “Correct the Map: Rebalancing global cartographic representation and promoting equitable representation of the world's regions, particularly Africa”. It is not binding — the UN can compel neither Google Maps nor textbook publishers nor national mapping agencies. The United States was the only country to vote against; Estonia, Georgia, Lithuania, Moldova, Serbia and Ukraine abstained.

No particular projection is prescribed; what is asked for is equal area. Equal Earth is named — built in 2018 by Bojan Šavrič, Tom Patterson and Bernhard Jenny, equal-area and designed for recognisable outlines at the same time.

The third button is the control case. On the globe the question of the right projection does not arise: area, shape and angles are all correct, because nothing is forced into a plane. It costs half the world — you only ever see one side. That is exactly the bargain every world map strikes, and the slider shows it in both directions. Drag to turn the globe.

Drag sideways to move the centre of the map. A world map has to be cut open somewhere, and whoever sits in the middle is shown whole while whoever sits at the edge is cut in two. That choice is a second bias on top of the projection, and it is just as arbitrary: the map here starts on Greenwich for no better reason than habit. It is one setting for all three views — on the flat maps it is the central meridian, on the globe it is the longitude facing you — so it survives when you switch projection or let the tour run.

Pinch or scroll to zoom in, on any of the three — the flat maps and the globe alike. Dragging then moves the section: sideways it keeps turning the world, so there is no left or right edge to run into. A double-tap, or a double-click, puts the whole map back in the frame.

Even the globe is not entirely off the hook, and the map says so: its image on a flat screen is a projection again. There the areal scale is the cosine of the distance from the centre of the image — where you look straight down it is exact, at 60° it is halved, at the rim it is zero. That is why the colour is not looked up in a table but measured, in every single frame, from what is actually on the screen. On Mercator this comes out to the same value as the table below, to two decimal places; on the globe it shows the rim compression, and it travels with the rotation.

What it is measured against is not the same thing in both cases. A flat map shows the whole world: the sheet is a fixed stock that gets shared out, and the question is whether a country gets more or less of it than it is due — both directions are possible. The sphere shows one half and has a natural scale, the one at the point you are looking at. There it is undistorted; everywhere else it is too small. The globe shows nothing too large, but almost everything too small — which is why it is pale in the middle and grows steadily stronger towards the rim.

Who gets how much room

Share of the land area shown, Antarctica not counted. Equal Earth and the globe share a column: both are equal-area, one because it was built that way, the other because there is nothing to distort on a sphere. That second column is therefore not just a third opinion — it is the true share.

Biggest change

Countries above 150,000 km², sorted by how their share of the image area changes when switching from Mercator to Equal Earth.

What Mercator is good at

The distortion circles say it most clearly. Every one of them has the same radius on the earth, 800 km. On Mercator each one stays a circle — they merely grow larger towards the north and the south. A circle that stays a circle means: at this spot the stretching is the same in every direction. So angles are preserved, and with them the local shape. On an equal-area projection it is the other way round: every circle is the same size, but sheared into an ellipse — the area is right, the shape is not. There is nothing more to say about the trade, and you see it in a single movement. On the globe the circles are then the same size and round: the control case that shows both are possible at once — just not on a sheet of paper.

Measured, and worth being exact about. On Equal Earth all thirty circles come out to the same area to the last digit — 0.0000 % spread — while their axis ratio runs from 1.23 to 3.16. Even on the equator it is 1.36, so Equal Earth is nowhere conformal; equal area is bought everywhere, not only at the edges. On Mercator the areas grow by 1.00, 1.34, 4.10 from the equator to 60°, against sec²φ of 1.00, 1.33, 4.00 — and the shapes are not perfect circles either: 1.00 on the equator, 1.08 at 30°, 1.25 at 60°. That is not a flaw in the drawing. A Tissot indicatrix is an infinitesimal circle; these are 800 km wide, and across that span Mercator's own scale already changes, so the northern half of each circle is stretched more than the southern half. Drawn at a size you can see, the circles show the projection's second-order behaviour along with the first.

The two flight routes show what that was practically good for. Because angles are correct on Mercator, a line of constant compass course is a straight line there — you lay down a ruler and read off the bearing. In 1569 that was the entire job. The shortest path is a different thing and on Mercator always the arc; both are labelled on the map.

How straight the course line is can be measured — greatest deviation from the straight connection, as a percentage of the route length:

RouteMercator, course lineMercator, shortest pathEqual Earth, shortest path
New York – Lisbon0.00 %10.1 %9.8 %
Frankfurt – Tokyo0.00 %31.4 %17.8 %

Zero per cent, and that for any route whatsoever — not an approximation but the property the projection was built for. The converse does not hold: the shortest path becomes straight on neither projection here, it merely bends less. That would take a gnomonic projection, which in turn cannot show even a full hemisphere in one piece. The price of the straight course line is the area distortion; on a flat map you cannot have both.