Human colour vision is a relatively recent acquisition. It is, at most, 63 million years old, and it may be a lot younger. On a genetic level, it is a mess: misalignments and redundancies in the genes that code for our “red” and “green” colour perceptions account for 95 per cent of all variations in human colour vision, and it is quite usual for up to nine genes to cluster together in an attempt to code for these colours. This is why the perception of colours – especially blues and greens – varies so much between individuals. Humans perceive colour through three types of colour-sensitive cell, called cones, but some have four types. Equipped with four receptors instead of three, Mrs M – an English social worker, and the first known human “tetrachromat” – sees rare subtleties of colour. Looking at a rainbow, she can see 10 distinct colours. Most of us only see five. She was the first to be discovered as having this ability, in 1993, and a study in 2004 found that two out of 80 subjects were tetrachromats.

If our eyes did not move – if they simply “drank in” the view before them – we would go blind. Our retinas can only process contrast, and soon become exhausted looking at the same thing for too long. They must tremble constantly in order to bring still objects into view.

Human vision captures only two degrees of the world with any clarity, so we tend to miss things that happen outside our focus of attention – and the more we concentrate, the more extreme our “attention blindness” becomes. This makes us easy prey for psychologists such as Daniel Simons and Christopher Chabris, whose notorious experiment of 1999 asked its viewers to score a three-a-side, 90-second basketball game. Afterwards, the viewers were told to relax, put down their score cards and watch the video again. Only then did the game’s most remarkable feature come to light: the invasion of the court, a few seconds in, by a 7ft-tall pantomime gorilla.

Our eyes stay several steps ahead of us, whatever we happen to be doing. When negotiating a turn in the road, for example, a driver’s eye will provide motor information to his or her arms almost a second before he or she makes any movement. By then, the eyes will already be looking elsewhere. Visually at least, we operate in the world not as it is, but as it existed half a second ago. This raises a not insignificant question: how does the eye know where to direct its gaze next?

The concept of a bionic eye is nothing new. In the 1970s, bio-engineer Paul Bach-y-Rita, now at the University of

Wisconsin-Madison, was turning different parts of the body into eyes. His prototypes were vests containing hundreds of mechanical vibrators. Pixelated images from a low-resolution video camera, worn on a pair of glasses, were translated into mechanical vibrations against the skin of the chest or back. Bach-y-Rita’s volunteers were able to recognise faces using the system. Proof that they could see came when Paul threw balled-up papers at them: they ducked.

Because light behaves differently in water and air, land-adapted human vision is lousy in water. Someone, however, forgot to tell the Moken – gypsies who ply the Burmese archipelago and Thailand’s western coast. Moken children, who spend days diving for clams and sea cucumbers, can see twice as much fine detail underwater as European children. While the pupils of the latter expand underwater, in response to the dimness of the light, Moken pupils shrink to their smallest possible diameter, improving acuity underwater. Mokens also use the lenses of their eyes more, squishing them to the limit of human performance.