Theory of color vision

While not an official biography, the following is reprinted with the kind permission of the Royal Society.

by F.W. Campbell, F.R.S.†

Land thrived on the challenges of scientific exploration. His favourite scientific interest was the study of human colour vision in complex, real-life scenes. This began with the study of two-colour projection using red and white light. It extended to Mondrian displays - overlapping rectangles of plain colours, after Piet Mondrian (1872-1944) , leader of neoplasticism -- and to computer simulation of real scenes. Later, he worked with David Hubel of Harvard, and Semir Zeki, of University College London, who was the first to realize the physiological implications of the correlation of his colour experiments with Hubel’s neurophysiological findings. A very full account of Land’s contribution to colour vision and, neurophysiological studies is edited by David Ottoson & Semir Zeki (1985).

As a part of his research on making an instant colour film, Land repeated some of James Clerk Maxwell’s experiments. The Maxwell experiment used three black and white transparencies, each taken and projected with a red, green, or blue colour filter. Land loved to experiment. He would study the effect of more red light or less blue light. What happened with different contrast film? To believe any hypothesis Land needed the feeling of certainty that only comes from many probing experiments. At the end of a long evening of experimenting with three projectors, Land decided to go home. A colleague had shut off the blue projector and had put the green filter away, leaving an image of red and white light. Another colleague, Meroè Morse, remarked to Land that the colour was still there. Land replied ‘Oh yes, that’s colour adaptation’. The group went home. At about two o'clock in the morning, Land sat up in bed, saying ‘Colour adaptation, what colour adaptation?’ He got out of bed, went into the laboratory and started three decades of experiments on complex images.

He had assimilated the colour adaptation ideas handed down from Helmholtz (1821-94). He searched for evidence of colour adaptation as the explanation for what he saw. Instead he showed that there was no experimental support for the notion that colour constancy is caused by colour adaptation in complex coloured scenes. Land created a better understanding of how we see the real-world of complex images.

What was special about Land was his passion for doing experiments, along with his quick mind that playfully questioned everything, particularly his own hypotheses. He literally could not sleep when he found an experiment that was trying to tell him something.

I first learnt about Land’s theory of colour vision in 1966 when he gave a demonstration in London. I raised a mild criticism and he said he would have to think about it. About a year later he phoned me in the morning and asked if we could try an experiment in the afternoon. Of course, I agreed. He then spoke to his pilot and asked me to meet him at the local Cambridge airport at 12 noon. We drove to the Physiological Laboratory and a few hours later he said ‘ You are right, I am wrong’. I complimented him on the speed with which he could change his mind and he replied ‘You cannot run a successful company carrying an incorrect fact in your head’.

I returned Edwin to the Cambridge airport and then drove the short distance to my home. I could hear a jet-aircraft taking off and, to my surprise, it flew over my house as it climbed. It then tilted port and starboard twice to say ‘Thank you’ before heading for Heathrow. I suddenly realized that he had planned his itinerary down to the last detail. He must have found out where my home and laboratory were in relation to the airport so that his visit would not inconvenience me. He put his plan into action as soon as he knew he had made a fast time across the Atlantic and could keep his appointment in London. Lastly, he could satisfy his obsession of doing at least one experiment each day. I could almost hear him laugh as he asked his pilot to say ‘Thank you’ or ‘ Au revoir’! The ‘other Cambridge’ had made his day; now back to being Chief Administrator in London.

Land performed thousands of experiments studying two-colour projections. He studied combinations of different filters, different contrast films, positive and negative projections, additivity of stereo projections. He found a 1914 British patent by Fox & Hickey entitled ‘improvements in kinematic apparatus’ which might have contained the gist of Land’s ideas formed many years later.

The experiments on red and white projections fascinated Land. He frequently gave lectures on the subject. His favourite audience was college undergraduates. His lectures were as much on the scientific method as they were on colour. He would not just talk about his experiments, he would bring the experiments to the lecture hall so that the students could do them with him. Around the world, there is an entire generation of scientists in any and every field who recall warmly their fond memories and their great sense of excitement about science that was generated by Land’s lectures on colour.

On 2 November 1972, Land gave the Friday Discourse at the Royal Institution of Great Britain. The experiments used an array of papers with controlled illumination. Later these experiments led to the black and white Mondrian experiments that showed that the same quanta catch at the receptors could generate any level of lightness from white to black in a single field of view. This was followed by the colour Mondrian experiments that showed that the number of quanta caught by any point on the retina could appear in any colour. It demonstrated, in a very dramatic fashion, that the determinant of colour is spatial in nature. It is the relationship of the quanta caught at one point or other points in the field of view that controls the appearance. Lightness and colour are field phenomena, not point phenomena.

Land often quipped that the history of colour vision would have been fundamentally different if Maxwell had studied electromagnetic fields before he invented the colour top - the basis of the science of colorimetry. Colorimetry has extraordinary predictive power for colour matches. But, because all the information used in the calculation comes from a single point in the image, it cannot offer any direct clue as to the appearance of the colour. Both before and after Maxwell, there have been many experiments that pointed out the limitations of single-point colour calculations. If Maxwell had studied colour 30 years later, he might have thought of colour as an array. However, Maxwell did not and Land became the principal proponent of colour as a field phenomena.

To explain the results of various colour constancy experiments, Land proposed the Retinex Theory. He coined the word retinex (made from the words retina and cortex) to de-signate the physiological mechanisms that generate these mathematically independent images: My proposal did not demand that the retinal elements of the same peak sensitivity have to be connected to each other. Instead, somewhere in the retinal-cerebral structure, elements associated with the same wavelength characteristics co-operate to form independent images in terms of lightness.