The Face on Mars: Viking 1 photographs Cydonia (1976)

Summer 1976. NASA's Viking 1 orbiter was photographing the northern latitudes of Mars. Its task was practical. Find a landing site for Viking 2.

On July 25 it passed over Cydonia, on the boundary between the cratered southern highlands and the northern plains. The range was 1,873 kilometres. The sun stood about 20 degrees above the horizon. The camera took frame 035A72.

The Viking team was working through the images for likely landing sites. Among the eroded hills was one that resembled a face. A mesa about 1.5 kilometres across. Eyes. A nose. A mouth.

JPL released the picture on July 31, 1976, as press release P-17384. The caption was sceptical from the start. The huge rock formation, it said, "is formed by shadows giving the illusion of eyes, nose and mouth." Viking chief scientist Gerald Soffen called it "a trick of light and shadow." A later NASA caption added that bit errors formed part of one of the "eyes" and "nostrils."

In 1977 two engineers, Vincent DiPietro and Gregory Molenaar, came across the image. They found a misfiled second frame, 070A13, taken 35 orbits later under a different sun angle. To them, the face was still there. They computer-enhanced both and in 1982 published Unusual Martian Surface Features, which also pointed to a pyramid and a grid-like pattern some saw as a city.

Richard C. Hoagland went further. His 1987 book The Monuments of Mars: A City on the Edge of Forever argued that an ancient civilisation built the Face, the nearby D&M Pyramid and a "City." Mark J. Carlotto used shape-from-shading in Applied Optics in 1988 and argued the Face held up in three dimensions. Stanley V. McDaniel's 1993 report said it "could well have been artificially created." Hoagland accused NASA of withholding data. Some in the same circles tied the August 1993 loss of Mars Observer to Cydonia.

April 5, 1998. Mars Global Surveyor passed over at 4.3 metres per pixel. Malin Space Science Systems processed the image the next morning and sent it to JPL for immediate internet release. It showed an eroded butte or mesa. Hoagland said the contrast enhancement "removed too much data."

So the cameras returned. Mars Global Surveyor imaged it again in April 2001, at 1.56 metres per pixel. Jim Garvin, NASA's lead scientist for Mars and the Moon, explained why: "We felt this was important to taxpayers." Mars Odyssey followed in 2002. ESA's Mars Express came in 2006, in stereo and colour. Then HiRISE, on April 5, 2007, at about 30 centimetres per pixel. Its caption: "an eroded mesa made famous by its similarity to a human face."

The scientific consensus is pareidolia. "The human brain is wired to identify faces, even when they aren't there," wrote the Planetary Society. Low resolution, a low sun and bit errors did the rest. Carl Sagan made the case in The Demon-Haunted World. Skeptoid pointed to the law of large numbers: among thousands of mesas, one that looks like a face is to be expected. ESA read the landform as a remnant massif shaped by landslides. In 2023 HiRISE recreated the Viking lighting from a terrain model. The face appears only in low sunset light.

Nobody ever claimed the 1976 frame was faked. The argument was always about what it showed. Along the way the Face reached the Weekly World News, The X-Files, Futurama and Mission to Mars. The picture was taken to find a landing site. It ended up as a lesson in how much a face depends on the light.