The Mollweide projection is an equal-area world map drawn as a 2:1 ellipse, published in 1805 by the Leipzig mathematician and astronomer Karl Brandan Mollweide. Every region keeps its true share of the Earth’s surface, so Africa correctly appears about 14 times the size of Greenland, but shapes are squeezed and sheared toward the rim of the oval. The equator and central meridian are straight lines at right angles, the parallels are straight and bunch together toward the poles, and the meridians at 90° east and west form a perfect circle. Reinvented in 1857 by the French physicist Jacques Babinet as the “homalographic” projection, it is now a standard choice for maps of global distributions and for whole-sky maps such as NASA’s WMAP image of the cosmic microwave background.
Karl Mollweide’s 1805 ellipse and the trick that keeps area true
Mollweide (1774–1825) taught in Halle and Leipzig and is also remembered in trigonometry for Mollweide’s formula. His map projection belongs to the pseudocylindrical family: like a cylindrical map it has straight, horizontal parallels, but unlike one its meridians curve inward to meet at the poles. On the Mollweide the equator is drawn twice as long as the central meridian, which gives the 2:1 outline, and meridians are equally spaced along the equator.
What keeps area true is the spacing of the parallels. Mollweide set them so that the share of the ellipse lying between the equator and any parallel equals the share of the globe lying between the equator and that latitude. That requires an auxiliary angle, θ, defined by the equation 2θ + sin 2θ = π sin φ, where φ is latitude. It has no simple closed-form solution and is normally solved by Newton–Raphson iteration, which converges quickly except near the poles. Once θ is known, the vertical position is √2·R·sin θ and the horizontal position scales with cos θ. The whole map then covers 4πR², exactly the area of the generating globe.
The result is that parallels are pushed apart near the equator and crowded near the poles. Half of each hemisphere’s area lies between the equator and 30° latitude, so on the Mollweide the 30th parallel sits about 40% of the way from the equator to the pole, not a third of the way as it would on a map with even spacing.
The projection travelled under several names. Jacques Babinet reinvented and popularised it in 1857, calling it the homalographic projection, a name meaning an equal-area map. The variant spelling homolographic spread through nineteenth-century star atlases, and the map is also catalogued as the Babinet projection or simply the elliptical projection. The geometry never changes with the label: on a globe of radius R, the map runs from −2√2R to +2√2R horizontally and from −√2R to +√2R vertically.
Mollweide vs Hammer (1892) and Aitoff (1889): three ovals that are not the same
Put a Mollweide beside a Hammer map and most people cannot tell them apart. Both are 2:1 ellipses of the whole world. The giveaway is the parallels: straight on the Mollweide, gently curved on the Hammer. Ernst Hammer described his projection in 1892, deliberately keeping the Mollweide’s outline but trying to reduce the severe distortion in the outer meridians. He built it from the equatorial form of the Lambert azimuthal equal-area projection, halving longitudes, projecting, then stretching the result sideways into a 2:1 ellipse. Like the Mollweide it is equal-area, but its curved parallels were Hammer’s way of easing the shearing at the left and right edges.
Hammer was openly building on a method published three years earlier by David Aitoff, who in 1889 applied the same halve-and-stretch trick to the azimuthal equidistant projection. The Aitoff looks almost identical to the Hammer but is not equal-area, which is why the Hammer saw more use. The two are confused so often that some authors still label Hammer maps “Aitoff”, even though Hammer carefully credited his predecessor. Later variants pushed the idea further: in 1953 William Briesemeister rotated the Hammer to centre it on 45° N and 10° E and squashed it to a 7:4 ratio, to present the land masses more centrally and with less distortion.
So the three ovals divide neatly. Mollweide: equal-area, straight parallels, worst shearing at the rim. Hammer: equal-area, curved parallels, gentler rim. Aitoff: curved parallels, not equal-area, a compromise between shape and size.
Mollweide vs the sinusoidal and Goode’s 1923 homolosine
The sinusoidal projection is the Mollweide’s older, sharper-edged cousin. Jean Cossin of Dieppe used it on a world map in 1570, and it is also called the Sanson–Flamsteed projection after later users. It too is equal-area and pseudocylindrical with straight parallels, but its parallels are evenly spaced and true to length, and its poles are points, so the outline is a pointed lemon rather than an oval. The equator and central meridian carry no distortion at all, but the outer meridians curve so sharply that continents near the edge lean badly. The Mollweide also distorts its rim, just less severely. The sinusoidal still has a scientific afterlife: NASA’s MODIS land team built its integerized sinusoidal grid on it.
In 1923 the American geographer John Paul Goode stitched the two together. His homolosine projection uses the sinusoidal from the equator to 40°44′11.8″ north and south, where the scales of the two projections match, and the top slices of a Mollweide above that, which leaves a slight kink in each meridian at the seam. The name fuses “homolographic”, the old name for the Mollweide, with “sinusoidal”. Because both parent projections are equal-area, so is the hybrid. Goode then interrupted the map through the oceans, giving the “orange-peel” look that was common on world maps in the 1960s. The homolosine grew out of his 1916 experiments in cutting up the Mollweide itself.
Interruption is also available on the Mollweide alone. A sinusoidal-interrupted Mollweide replaces the single central meridian with lobes, each with its own half-meridian; a parallel-interrupted version joins several small ellipses at the equator. Oblique versions rotate the grid so the worst distortion falls on the oceans and the continents stay truer to form.
Mollweide vs Eckert IV, Robinson (1963) and Equal Earth (2018)
The other common look-alikes flatten the poles into lines. Max Eckert published six projections in 1906; his Eckert IV is equal-area like the Mollweide, but its poles are lines half the length of the equator and its outer meridians are semi-ellipses, which gives a rounded rectangle instead of a pure oval and less crowding at high latitudes.
The Robinson, created by Arthur H. Robinson in 1963 for Rand McNally, also has polar lines, but it is a compromise projection that is neither equal-area nor conformal; Robinson tuned it by eye until it looked right and derived the numbers afterwards. The National Geographic Society used it for world maps from 1988 until 1998, when it moved to the Winkel tripel. The Equal Earth projection, designed in 2018 by Bojan Šavrič, Bernhard Jenny and Tom Patterson, borrows the Robinson’s look but keeps areas exact. Its authors wrote it after Boston Public Schools adopted the Gall–Peters projection in March 2017 and they found no existing equal-area map they liked.
| Projection | Year | Equal-area? | Outline and poles |
|---|---|---|---|
| Sinusoidal | used by 1570 | Yes | Pointed; poles are points |
| Mollweide | 1805 | Yes | 2:1 ellipse; poles are points; straight parallels |
| Aitoff | 1889 | No | 2:1 ellipse; curved parallels |
| Hammer | 1892 | Yes | 2:1 ellipse; curved parallels |
| Eckert IV | 1906 | Yes | Rounded rectangle; poles are lines |
| Goode homolosine | 1923 | Yes | Usually interrupted lobes |
| Robinson | 1963 | No | Rounded rectangle; poles are lines |
| Equal Earth | 2018 | Yes | Rounded rectangle; poles are lines |
Against this field the Mollweide’s trade-off is clear. Its pointed poles keep polar regions small and honest in area, but Alaska, eastern Russia and New Zealand, which sit near the rim on a map centred on Greenwich, come out visibly slanted. Projections with polar lines spread the high latitudes out and look calmer, at the cost of a less elegant outline.
Where the Mollweide wins: WMAP’s sky map and global data
The Mollweide’s natural home is anywhere the eye should compare quantities by area. On a Mercator map Greenland looks as big as Africa; on a Mollweide, Africa is correctly shown about 14 times larger, and Greenland is about the size of the Democratic Republic of the Congo, as it is in reality. That makes it a common base for maps of population, forest cover, climate zones and other global distributions, alongside the Winkel tripel and Robinson for general reference.
It is just as useful for the sky. A whole-sky map is also a sphere seen from the inside, and astronomers want every patch of sky to keep its true size. The elliptical image of the cosmic microwave background released from NASA’s Wilkinson Microwave Anisotropy Probe is a Mollweide projection. The final, nine-year WMAP data release came on 29 December 2012 and mapped temperature fluctuations of about ±200 microkelvins across the whole sky.
The projection has also seeded others: besides Goode’s homolosine, the van der Grinten and the Boggs eumorphic were inspired by it. Its weakness is the one every equal-area map shares: you cannot keep both area and shape on a flat sheet, and the Mollweide spends its shape budget at the rim. For the opposite trade-off, where shapes stay true and areas balloon, see Mercator vs Peters; for the compromise that National Geographic once chose, read about the Robinson projection; and for the basics behind all of them, start with what a map projection is.
