In this chapter I will give a very brief summary of how our biological ‘vision machinery’ is constructed and how it works, particularly with regard to colour vision. The eye and the brain belong to the most complex organs in the body and the whole book might easily have been filled with only these subjects.

Our three main kinds of vision

Traditionally, humans are considered to have five senses: sight, hearing, touch, smell and taste (more modalities might be added, like pain, equilibrium, etc.). All the traditional five are essential for a normal life, but the one that the majority of people would be most unwilling to lose is probably sight; this, in spite of the fact that a blind person probably becomes less socially isolated than somebody who is completely deaf. Compared to many other species of mammals, humans, apes and monkeys are unusually dependent on vision in their normal behaviour and interactions. From an evolutionary point of view this is understandable: a good sense of vision is essential for climbing and moving around in the trees of jungles and savannahs. Consequently, a large proportion of our brain is devoted to various kinds of visual analysis. What are we seeing and how? What is the relationship between colour vision and our other visual capacities? How important are the colours?

The great importance of our visual capacities is further underlined by the fact that we may be said to have three different kinds of vision, two for use in daylight and a third one for darkness:

1. Our luminosity vision (vision of achromatic lightness contrast, i.e. ‘blackand- white’ vision). In normal daylight, our eyes and brain are extremely well equipped to discover the edges of contrast between darker and lighter regions in the visual field. This capacity is the most essential one for our visual orientation and for recognizing structures and objects in the surroundings. In the central portion of the visual field, such functions may concern very small details (Figure 3.2).

2. Our colour vision. This kind of vision is also used in normal daylight, and we are well equipped for discovering differences in colour and edges of colour contrast.

For a better understanding our own visual capacities and our perception of colours, it is highly interesting to know what properties such functions have in other animals and how they emerged during evolution. Furthermore, for the general understanding of the behaviour and ecology of animals, it is in itself essential to consider the functions of their visual systems, including their colour vision.

Methods for comparative studies of (the capability for) colour vision

There are two main techniques available for studying colour vision capabilities in animals:

1. Behavioural experimental methods can be used for finding out how well an animal might distinguish different hues and levels of saturation. This is usually done such that the animal gets some attractive reward (piece of food or drink) if it makes a correct choice between a number of different colours. For instance, the food might be hidden behind one out of several panels, all painted in suitably different colours. If the animal pushes or taps on the correct panel, it receives the reward. When testing for the recognition of colour hues, it is important to take care that the animal cannot use any other clues, i.e. the hue must be recognized independently of the lightness and saturation of the target colour. Such behavioural methods are time consuming, partly because they typically require an extended period of training before the animal learns what it is supposed to do. However, with patience and resolve, these techniques may deliver very detailed information about many aspects of the animal's sensory perceptions, e.g. which wavelengths and wavelength combinations may be distinguished from each other. Furthermore, with suitable adaptations, these methods are applicable to practically any species of animal; they have, for instance, been successfully used for studies of honey bees.

2. Physiological and anatomical/biochemical experimental methods can be used to investigate the functional properties of the sensory machinery available for colour vision. For instance, the reactions of the eye and its receptor cells may be measured while being stimulated with light of different wavelengths. With regard to colour vision it is then of particular interest to find out how many different types of receptor cells there are in the eye (e.g. kinds of cones and rods) and in roughly which relative proportions they occur.

Methods for the testing of colour vision became needed when it was discovered, at the end of the eighteenth century, that a considerable percentage of the population was ‘colour blind’. In the earliest investigations, the testing material often consisted of various colours of cloth. One such test, 20 ribbons in different colours, was used by John Dalton, who asked his participants about the colours of these items in daylight as compared to in candle light (cf. Section 5.1). In the 1870s, Frithiof Holmgren developed a test with skeins of coloured wool (see), using procedures based on colour matching rather than naming. A large number of variously coloured skeins were mixed in a heap. The examiner chose a small number of ‘master skeins’ and then asked the participant to search the heap for other skeins with a similar colour.

Nowadays deficient colour vision is commonly tested with several techniques used in parallel. For a rapid and easy initial diagnosis, one of the many ‘coloured-dot tests’ may be used (Section A.1). For persons found to be colour blind in such a test, it will often be appropriate to add some other procedure for evaluating the degree of colour vision deviation. This might involve a general colour-sorting task, like the D15 test (Section A.2). For certain kinds of occupations, potential employees will be evaluated using colour-identification tasks similar to those encountered in a work situation (e.g. coloured light signals like those used at sea; Section A.3). In scientific contexts, it is important to be able to identify both the degree and the type of deviation; this is best done using a specialized instrument called an anomaloscope (Section A.4). Still other procedures might be added in the future (see Sections A.5 and A.6).

Pseudoisochromatic plates

Ishihara

One of the best and most famous examples of this type of test was developed in Japan by Shinobu Ishihara (1879–1963), a professor of eye diseases at the University of Tokyo. The first edition of his test was published in 1917, around 100 years ago. The test is so well made that it is continuously reprinted, and it remains one of the most frequently used tests for colour blindness.

This is a book about many different aspects of colours, how they arise and how one might see and experience them. When writing this book, my first source of inspiration was my own visual system: I belong to the rather large minority with an inherited red-green blindness. It has often astonished me that most people know so little about what this sensory constitution means, in spite of the fact that, in our part of the world, it affects more than 4% of the total population. Thus, I started my writing enterprise as a book about colour blindness, but the project gradually expanded to become a more general survey of matters concerning colour. The description includes an account of the physical and physiological mechanisms of colours and colour vision in humans and other animals, which comes naturally to me because I worked in neurophysiological research for many years (albeit on subjects other than colour vision).

Colours often give us a very direct and immediate kind of sensory experience and one might therefore be inclined to think that the nature of the phenomenon is simple and straightforward. This is, however, not the case: colour vision is a highly complicated and multidimensional subject matter. For many people, colours are an important source of enjoyment in everyday life, in nature and in various expressions of art and culture (true also for red-green colour-blind persons). Publications about colour often mainly deal with their various aesthetic qualities. In 1819, Keats published his very long poem, Lamia, which includes a few famous lines suggesting that the rainbow might lose its colourful beauty if one knows too much about it:

However, it might equally well be argued that the unweaving of a rainbow does not make its colours and beauty less impressive but rather the opposite: the more one knows about a subject the more interesting and captivating it usually becomes. According to some interpretations of Keats’ poem, the author himself and his contemporary colleagues might even have agreed on this point, provided that one does not lose one's sense of wonder when confronted with the many complexities of human perceptions and the natural world.

Our eyes are often said to be ‘camera-like’ (or, rather, our cameras are ‘eye-like’): using a lens, a picture of the surrounding world is projected onto the reverse side of the eye/camera, where it is somehow recorded and retained for later use. In our eyes, the recording and further processing happens using receptors and nerve cells. In analogue cameras, the recording is done using a light-sensitive film. In digital cameras, the film is replaced by a light-sensitive electronic sensor and the further processing is done using computer techniques. There are many interesting similarities and differences between our visual system (eyes and brain) and the functional properties of modern digital cameras. In this Appendix, I will concentrate on colour-associated aspects.

Sensor and colour analysis

The light sensor of a digital camera contains a huge number of light-sensitive picture elements (pixels), often about 10 million or more. There are three types of elements, each one reacting to a different range of wavelengths: R pixels for red, G pixels for green, B pixels for blue/violet; this wavelength selectivity is typically produced by placing a different optical filter on top of each kind of sensor element. The signals from the R, G and B pixels are analysed using an RGB model (cf. Plate 2.6). With methods similar to those of the CIE system, the RGB-signals may be used to calculate the colour to be produced in different portions of the picture. The RGB-models are directly derived from studies of our visual system, and these digital camera techniques may indeed seem to be very eye-like. However, there are also several evident differences between eyes and the RGB-organization of digital cameras.

Camera sensors often have their RGB-pixels arranged according to a so-called Bayer pattern (Plate D.1) with 25% R, 50% G and 25% B pixels; the name of this pattern comes from Bryce Bayer, who worked at Kodak in the 1970s. This pixel pattern gives the camera sensor a maximal light sensitivity within the green region of the spectrum (cf. eye sensitivity ‘In daylight’, Figure 4.4). However, in the retina the proportions of the various light-sensitive elements are quite different from those of the Bayer pattern (Section 4.3.1); for instance, the eye has only about 5% S cones (cf. 25% B pixels). Furthermore, the spatial distribution of the retinal cones is very heterogeneous and complex (Section 4.3.2).

In my childhood, the surrounding world was rather sparsely coloured. In the 1930s to 1950s, photographs and movies were typically in black and white and there were no TVs or computers and, hence, no associated colour monitors. Typewriters wrote in black (sometimes also in red), and important sections of a text were not highlighted but simply underlined with a pencil. The telephone was immobile and typically black. Books about the visual arts were largely illustrated in greyish halftones, perhaps supplemented with a small number of expensive colour plates. Since then, colours have exploded into almost all sections of modern life, at work as well as in leisure activities: colour TV, computers with colour monitors and colour printers, digital cameras with colour sensors, even telephones take colour snapshots, and magazines and newspapers are illustrated in full colour. This present-day ubiquity of colour probably does not signal a sudden change in human culture and preferences but rather it reflects the power of modern electronic and chemical technology, allowing the innate human fascination with colour to become more fully expressed.

In this introductory chapter, I write a little about the practical importance and use of colours, their names, history and cultural significance, things one might discuss without much knowledge about the colour mechanisms. For instance, how many different colours can one actually see? How can they be described and labelled?

Numbers and dimensions of colours: hue, colourfulness, brightness

Laptop manuals often mention the mind-boggling number of about 16 million colours that may be seen on the monitor screen. However, this only means that, technically, the screen can produce about 16 million different light qualities. It does not tell you how many of these qualities the human eye can distinguish from each other, seeing them as different colours; many of the 16 million colour settings will actually look the same to a human observer. Still, the total number of distinguishable colours is very large. For the analysis of this complex landscape, it is practical to consider its three main dimensions (Plate 1.1):

1. The hue of the colour, the property that we label with terms such as red, yellow, green, blue, etc.

2. The colourfulness of the colour (also often referred to as its saturation, purity or chroma), i.e. how strongly the hue is expressed. The colourfulness varies with admixtures of white/grey/black: the more white/grey/black one adds to a colour, […]

Wavelength and oscillation frequency of light

There is an inverse proportional relation between the wavelength (λ) and frequency (f) of light according to the formula: f = c/λ, in which c is the speed of light in a vacuum (about 300 000 km/s). For visible light, wavelength is usually measured in nanometres (nm) and the frequency in terahertz (THz). There are 109 nm per metre (i.e. 1000 million) and a frequency of 1 THz means that there are 1012 oscillations per second (i.e. a million million). The oscillation frequency of light in terahertz is obtained by dividing 300 000 by the wavelength in nanometres. Visible light has wavelengths of about 380–730 nm, which thus corresponds to oscillation frequencies of (300 000/380) to (300 000/730) = 789 to 411 THz (cf. Plate 2.4).

Light refraction

The famous speed of light of about 300 000 km/s is only valid in a vacuum, and it is then alike for all wavelengths. However, when passing through various materials the speed of light goes down, and this slowing is more marked for shortwave than for more longwave light. This slowing-down causes a light beam to change its direction (to refract) when passing an oblique borderline between two different translucent materials, e.g. at the border between air and glass (or air and water). The refraction is greater the shorter the wavelength. Hence, as was investigated by Newton in his famous experiment, light may be separated into differently coloured components by letting it pass through a glass prism (Plate 2.1b). The degree of slowing-down and the resulting refraction is different for different materials, and a measure of this property is given by the index of refraction, which is equal to the ratio between the speed of light in a vacuum and the speed in the target material. For instance, the index of refraction is 1.0003 for air, 1.33 for water, between about 1.47 and 2.04 for different kinds of glass, and unusually high for a glittering material like diamond (2.417).

Quality of lighting

Practically everything we see in the outside world depends on light reflected from various objects. For colour vision, both the quality of the light (its wavelength composition) and its intensity are of great importance. In very weak light we see no colours because we are then using only our rods.

Colour interactions within a visual image

In the preceding chapter, results were described indicating that one might predict the perceived colour of a light if its wavelength composition is known. However, this is mainly valid for small and isolated portions of our central field of vision, e.g. for the observation of a small rotating colour top against a neutral background, or for Newton's prismatic spectrum displayed in a dark room. Our visual surroundings are typically much more complex; normally our eyes will simultaneously encounter a multitude of objects, differing in colour and lightness. Within such a complex scene, the colours perceived for each one of the various details will partly depend on local wavelength compositions, but partly it also depends on complex interactions between the various components of the scene. While we are looking, the brain uses a complex set of rules for correcting and adjusting all the colours before we perceive them.

Simultaneous contrast

In experiments concerning colour perception, colour samples are often used in the shape of small discs or pieces of paper which are shown against a neutral background. By such means, each colour may be judged on its own, in a state of relative isolation from other colours. It is important to realize that the same colour sample, having the same wavelength composition of its reflected light, might give quite a different colour perception in other surroundings or with another background (Plate 3.1). This concerns all the main dimensions of the colour: its hue, saturation and lightness. One of the common kinds of interaction concerns the ‘strengthening’ effects of adjacent opposites. A grey paper looks whiter against a black background. A yellow paper looks even more yellow against a blue background, and a blue colour becomes stronger against yellow. An interesting and paradoxical aspect of these interactions concerns the behaviour of complementary colours: they strengthen each other for adjacent surfaces, but they may instead sum and become very unsaturated if the two samples are small and close together (e.g. for minor stones in a mosaic, for threads in a woven tissue; cf. definition of complementary colours, Section 2.7).

One of the early scientists studying these colour interactions was the unusually long-lived French biochemist M.E. Chevreul (1786–1889). He is considered one of the founders of modern organic chemistry, and his work in this field led to a renewal of the soap and candle industry.

More or less quantitative coloristic specifications are important in many branches of industry, controlling whether various fabricated items have the specified and/or required colour. Such determinations might concern, for instance, consumer articles like food, cosmetics, textiles, clothes, furniture, car seats, etc. There are at least three alternative procedures for the specification of colours: (1) direct comparisons to samples from various kinds of colour charts; (2) measurements using a (tri-stimulus) colorimeter; (3) measurements using a spectrophotometer.

Colour charts and systems

In the first chapter of this book, I dealt with the rather limited everyday vocabulary used for describing the confusing multitude of possible colours (Sections 1.1, 1.2; Plate 1.2). However, it is sometimes necessary to be able to use a more differentiated terminology for defining precisely which colours one sees or wants to see. As a basis for such definitions, the colours need somehow to be ordered into a system.

From ancient times, various philosophers and scientists have tried to analyse how all the many colours arise and how they are linked and interacting. In the course of history, hundreds of such surveys have been assembled; more than 160 schemes of this kind were summarized and reviewed in a recently published book. Such lists and overviews have been made with different main purposes in mind:

1. Analytical surveys: colours ordered in relation to scientific findings concerning the physical/physiological basis for colour differences. This includes the systematic ordering of colours as seen in the solar spectrum (Plates 2.1b, 2.4), often arranged in circles with complementary colours opposite to each other (Plates 2.1c, 3.4b, 3.5b). Such analytical schemes form the basis for many of the more practically oriented surveys.

2. Practical surveys: lists and schemes providing easily used definitions for great numbers of different colours (Section B.1.1).

3. Measurement schemes: quantitative systems for the graphical analysis and summary of measured colour parameters (Section B.1.2).

These various strategies are not mutually exclusive, and several colour maps have been made with more than one of these purposes in mind. Besides systematic surveys of the kinds 1–3, there are also other kinds of summaries in which colours are ordered according to various aesthetic, emotional or other subjective/introspective aspects. Examples of such popular summaries are continually published in great numbers of books and articles, e.g. concerning clothes, gardens, home decoration, religion and spirituality, etc.

This chapter will deal with the peripheral mechanisms of colour: which are the characteristics of the physical phenomena that our eyes and brains interpret as ‘colours’? As a starting point, it is evident that colours in our surroundings are dependent on light: in darkness we generally see no colours, and the colour of an object can clearly be influenced by the characteristics of the illumination. Thus, for an analysis of colours and colour vision, it is necessary to know more about how varying characteristics of light might influence our sense organs and visual perception. This is a complex subject and the knowledge and insights concerning these matters have been gradually growing during a long period of time. The key concepts concerned are illustrated in a very concrete manner, I think, by considering the progress in this field historically, including a description of some crucial early experiments; this will be done in part of the present chapter.

It is, in this context, important to keep remembering that colours are experiences of conscious brains; as Newton pointed out ‘the rays are not coloured’. Colours do not, in themselves, have an independent existence but we see particular colours when lights of appropriate characteristics enter our eyes. Thus, when we talk about a light as being ‘blue’ or ‘yellow’, this is a simplified way of saying that the light in question has characteristics that most human observers would perceive and interpret as being ‘blue’ or ‘yellow’.

Light and darkness: early thoughts concerning the nature of colours

In the ancient world, the sense of vision was interpreted using models which now seem highly exotic. It was believed that, somehow, portions of the observed objects physically entered the eyes and gave rise to the perception. It was sometimes suggested that some kind of rays emerged from the eye itself to cause the correct capture of the necessary ‘layers’ of surrounding objects. Colours were believed to emerge from mutual interactions between light and darkness. Nowadays, darkness is essentially thought of as an absence of light. In the ancient world, darkness was instead believed to have a positive existence of its own, like some kind of black soup that might be mixed, in various proportions, with the white soup of light.

Plate 5.1 shows a number of coloured dots. For people with normal colour vision, some of the dots form the digit ‘6’, which may be seen thanks to the colour difference between these dots and the surrounding ones. However, for about 8% of all European men and and about 0.4% of the women these dots have about the same colour as many of the other ones and no numbers are seen. This picture is taken from a sensitive test for the identification of defective colour vision.

People with a deviant kind of colour vision are usually called ‘colour blind’, and in this chapter I will describe the most common type of this deviation, the inborn so-called ‘red-green blindness’, a visual constitution which is often diagnosed with test pictures like Plate 5.1. The terminology used for these varieties of colour vision is unfortunate and often leads to misunderstandings: only a very few of the so-called ‘colour-blind’ individuals are actually seeing the world without colour (Chapter 6), and this is certainly not the case for red-green blind people. Such persons see most or all of the reddish or greeinsh hues as chromatic colour, i.e. they are actually not at all ‘red-green blind’; they do, however, have difficulties in distinguishing many of the red-green hues and nuances from each other, a property that might be labelled colour uncertainty or colour ambivalence rather than colour blindness. Unfortunately, the terms ‘colour blindness’ and ‘red-green blindness’ are so firmly established that I feel forced to use them also in this book. However, it should be stressed at the outset that, as I see it, red-green blindness should not to be regarded as a disease, but rather as one of the many ways in which people might differ from each other in their inborn constitution.

John Dalton and his brother

In the world of a couple of hundred years ago, there were few (if any) colour-coded signals and the colours perceived were then a rather personal and introspective affair: people were not yet required to react to particular colours in rapid and highly specific manners. For such reasons, it took humanity a surprisingly long time to discover that people differed in their ability to detect and identify colour differences. One of the first investigators in this field was the prominent English physicist/chemist John Dalton (Figure 5.1a).