Plate58

The Face at Cydonia

Viking 1, frame 35A72, 25 July 1976

A mesa in the northern plains of Mars, photographed at low sun, at forty-three metres per pixel, with a data dropout that fell where a nostril would be.

Twenty-five years later, at 1.5 metres per pixel, it is a hill.

The standard telling makes this a story about credulity corrected by resolution. It is a better story about the fact that the face was in the data. At forty-three metres per pixel, a face is what that arrangement of light is. The perception was not a mistake about the mesa; it was a correct reading of the frame, and the frame was all anyone had.

Increase resolution and the face leaves. Increase it further — down to grain, down to the mineral — and the hill leaves too. There is no resolution at which the object resolves into what it finally is. There are only scales, each with its own inhabitants, and the face lived at one of them, briefly, the way a species lives in a climate.

It is not there now because the light changed. Nothing on Mars moved.