Colour-blindness marks the extreme of a system of normal variation in perception. When asked to mix red and green light until they march a standard orange colour, people divide into two groups that differ in the hue of the red light chosen. There arc two distinct receptors for red, differing in a single change in the DNA. About sixty per cent of Europeans have one form, forty per cent the other. Both groups are normal (in the sense that they are aware of no handicap) but one sees the world through rather more rose-tinted spectacles than the other. The contrast is small but noticeable. If two men with different red receptors were to choose jacket and trousers for Father Christmas there would be a perceptible clash between upper and lower halves.

In the 1930s, a manufacturer of ice trays was surprised to receive complaints that his trays made ice taste bitter. This baffled the entrepreneur as the ice tasted just like ice to him, but was a hint of inherited differences in the ability to taste. To some, a trace of a substance used in the manufacturing process is intolerable, while to others a concentration a thousand times greater has no taste at all. Much of the difference depends on just one gene which exists in two forms. That observation, the ability or otherwise to perceive a substance, now called PROP, was the key to a new universe <>l taste. Genetic 'supertasters' are very sensitive to the hops in beer, to pungent vegetables like broccoli, to sugar and to spices, while non-tasters scarcely notice them. Half the population of India cannot taste the chemical at all, but just one African in thirty is unable to perceive it. Students of my day thought it witty to make tea containing PROP to see the bafflement of those who could drink it and those who could not. Today's undergraduates have more sense.

As truffle-hunters know, scent and taste are related. There is genetic variation in the ability to snu-ll, among other things, sweat, musk, hydrogen cy.imdi- and (hi' odour of frcesias. Many animals comniiiiiu.iU' vvnh c.kIi oiIkt through the nose. Female mice can smell noi only who a male is, but how close a relative he might be. Humans also have an odorous identity, as police clogs find it more difficult to separate the trails of identical twins (who have all their genes in common) than those of unrelated people. Man has more scent glands than does any other primate, perhaps as a remnant of some uniqueness in smell which has lost its importance in a world full of sight. The tie between sex and scent in ourselves is made by a rare inborn disease that both prevents the growth of the sex organs and abolishes the sense of smell, suggesting that the two systems share a common pathway of development in the early embryo.

Variation in the way we look, see, smell and taste is but a tiny part of the universe of difference. The genes that enable mice to recognise each other by scent are part of a larger system of identifying outsiders. The threat of infection means that every creature is always in conflict with the external world. The immune system determines what should be kept out. It differentiates 'self from 'not-self' and makes protective antibodies that interact with antigens (chemical clues on a native or foreign molecule) to define whether any substance is acceptable. The millions of antibodies each recognises a single antigen. Cells bear antigens of their own that, with great precision, separate each individual from his fellows. Antigens are a hint of the mass of uniqueness beneath the bland surface of the human race.

When blood from two people is mixed, it may turn into a sticky mess. The process is controlled by a system of antigens called the blood groups. Only certain combinations can mix successfully. Some groups, ABO and Rhesus for example, are familiar, while others, such as Duffy and Kell, are less so. Because o[their importance in transfusion, millions of people li.ivc been tested. A dozen systems are screened on a routine basis and each comes in a number of forms. This miijiII sum pie of genes generates plenty of diversity. The chances ot two Englishmen having the same combination of.til twelve blood groups is only about one in three thousand. Of an Englishman and a Welshman it is even less;imt <>l.in Unglish person and an African less again.

Since the discovery of the blood groups and other cues on the surfaces of cells, there has been a technical revolution. Like the stone age revolution a thousand centuries ago, it depends on simple tools that can be used in many ways. The DNA of different people can now be compared letter by letter, to test how unique we are. The Human Genome Diversity Project is a spin-off from the main mapping effort which has tested thousands of people. On the average, and depending on what piece of the DNA is tested, two people differ in about one or two DNA letters per thousand; that is, in about three to six million places in the whole inherited message. Some of the differences involve changes in single bases (single nucleotide polymorphisms, or 'snips' as they are called), some in the number of short repeals (it particular sequences ('microsatellites¹ and 'mini-s.ilrllKcs'] and some turn on the presence or absence of hip. nl riHilnle DNA that leapt into a particular place in the genome long ago. Blood groups show how improbable it is that two will be the same when a mere twelve variable systems are used. The chance that they both have the same sequence of letters in the whole genetic alphabet is one in hundreds of billions. Genetics has made individuals of us all. It disproves Plato's myth of the absolute, that there exists one ideal form of human being, with rare flaws that lead to inborn disease.

Variation helps us to understand where we fit in our own family tree, in the pedigree of humankind, and in the world of life. Relatives ait more likely to xh.ne genes because they have an ancestor in lommou. As,tll genes descend from a carrier long dead they can be used to test kinship, however distant that might be. The more variants two people share the more they are related. This logic can be used to sort out any pattern of affinity.

This detective work is easiest when close — or identical — relatives are involved. The US Army tests the fit of dead bodies to their previous owners by storing DNA samples from soldiers in the hope of identifying their corpses after death. DNA can also say a lot about the immediate family. Once, immigration officers faced with applicants for entry often refused to believe that a child was the offspring of the woman who claimed it. Comparison of the genes of mother and child almost always showed that the mother was telling the truth. Our society being what it is, the tests are now less used than they were. However, not all families are what they seem. Attempts to match the genes of parents and offspring in Britain or the United States reveal quite a high incidence of false paternity. Many children have a combination of genes which cannot be generated from those of their supposed parents. Often, they show that the biological father is not the male who is married to the biological mother. In middle class society about one birth in twenty is of this kind.

Such detective work can skip generations. During the Argentinian military dictatorship of the 1970s and 1980s thousands of people disappeared. Most were murdered. Some of the victims were pregnant women who were killed after they had given birth. Their children were stolen by military families. When civilian rule was restored, a group of mothers of the murdered women began to search for their grandchildren, whose DNA was compared with rhose who claimed to be their parents. The message passed in the genes enabled more than fifty children to be restored to their biological families, two generations on.