Disease may say more about human diversity. Even the ABO blood groups system might result from its actions. The A and B variants differ in just seven bases in the thousand or so that code for them; O has a single DNA base missing part-way down the message, which scrambles all the text from there onwards and removes part of the cell-surface structure coded for by this gene. AB individuals have some protection against childhood diarrhoea and, more important, against cholera, while those with O are more susceptible to that infection (but might be more resistant to malaria). Other genes, too, seem to be associated with resistance. Perhaps ancient illnesses explain a lot of our diversity. Nevertheless, plenty of infections have gone for ever. Optimists claim that the conquest of disease, cold and starvation means that natural selection has come to an end. If evolution has one rule, it is to expect the unexpected. New pestilences may appear and cause as much damage as malaria, or those that seem near extinction will stage a resurgence, as has malaria itself.

The history of the battle against disease says useful things about natural selection. Far from designing a simple and effective protection, whenever a straw appears, it is clutched at. Selection acted like a handyman rather than a craftsman. Its products often seem badly, not to say extravagantly, planned and roughly made. If man is indeed made in God's image, malaria does not say much for divine engineering. This haphazard approach has its strengths. Used by engineers or computer programmers it can make subtle and unexpected things. The logic of selection is that of the living world: to produce a complicated design without a designer.

Natural selection has never in its three-billion-year history produced a wheel, let alone a work of art; although it has managed to generate eyes, brains and other organs of great complexity. This is because of its grearest weakness, its plodding approach. A wheel, or a watch, needs some long-term ideas. To make either demands an intellectual leap that is beyond evolution. Natural selection has superb tactics, but no strategy — but tactics, if pursued without thought for cost, can get to places which no strategist would dream of.

Nineteen hundred and six was a successful year for the Bronx Zoo. A new exhibit was pulling in the crowds. An African Pygmy — Ota Benga by name — was in the same cage as an orang-utan. The exhibit caused an uproar, not because it was a shameful spectacle, but because it promoted the idea of evolution, that apes and humans were related. After a time, Ota Benga was released, in part as a result of his habit of shooting arrows at those who mocked him. He moved to Virginia, where he committed suicide a few years later.

The Bronx Zoo view of human evolution was once widespread. Linnaeus himself, who first classified animals and plants, put the idea well in 1754: 'All living things, plants, animals and even mankind themselves, form one chain of universal being from the beginning to the end of the world.' Many still see evolution as a smooth progress, a seamless transition from the primaeval slime to New Labour. Linnaeus recognised several distinct varieties among our own species. As well as the yellow, melancholic and flexible asiaticus there was europaeus, white, ruddy and muscular; americanus red, choleric and erect; and afer, black, phlegmatic and indolent.

The groups of humanity were at different stages. Africans were at the bottom, close to the apes, Asians somewhere in between, and white Europeans — needless to say — at the top. Victorian writers did not hesitate to make the idea clear. Robert Chambers, who wrote an influential book on evolution fifteen years before Darwin, claimed that 'Our brain passes through the characters in which it appears in the Negro, Malay, American and Mongolian nations, and finally is Caucasian. The leading characters, in short, of the various races of mankind, are simply representatives of particular stages in the development of the highest or Caucasian type. The Mongolian is an arrested infant, newly born.'

The theory that races are different has a long and ignoble history that has brought misery and death in its wake. It reached into medicine. Most people have seen children with Down's Syndrome, which is due to an error in their chromosomes. This was called by its discoverer, Langdon Down, 'Mongolism¹ in his 1H66 paper 'Observation on an Ethnic Classification of Idiots' for what seemed to him a good scientific reason — these children had slipped a couple of rungs down the evolutionary ladder to resemble a lower form of life, the Mongols. A Japanese friend once told me that in his country the same condition is called Englishism. The idea is ridiculous. Down's Syndrome is due to a mistake in the transmission of a particular chromosome which is found in all groups of humankind and even in chimpanzees.

The history of race illustrates, more than anything else, the limitations of biology. Biologists have been talking — or shouting — about race for years. Ignorance and confidence have gone together. Politicians take scientists less seriously than scientists do, but the story of scientific racism, as it was known, is a grim one.

I have always felt a certain compassion for those whose ability to despise their fellow men is limited by the colour of their victim's skin. Genetics has — and should have — nothing to do with judgements about the value of one's fellow beings. In this sense, the biology of race has no relevance to racism, which is always happy to bend any scientific fact to its perverse ends. The genes do show that there are no separate groups within humanity. This may be reassuring, but should be beside the point. To depend on DNA to define morals is dangerous. Science evolves. It learns more, and theories alter. Our views on human biology have changed and may change again. The same should not be true of attitudes to human rights. Where biology stops and principles begin must not be forgotten.

Humankind can be divided into groups in many ways; by culture, by language and by race — which usually means by skin colour. Each division depends to some extent on prejudice and, because they do not overlap, can lead to confusion. In 1987, a secretary from Virginia sued her employer for discrimination as she was black. She lost the case on the grounds that, as she had red hair, she must be white. She then worked for a black employer and, undaunted by her earlier experience, sued him for picking on her as she was white. She lost again as the court found that she could not be white as she had been to a black school.

Nations, too, differ in how they define their racial affinity. In South Africa just one African ancestor, even in the distant past, once meant ejection from the white race. In Haiti, in contrast. Papa Doc proclaimed his country to be a white one, as almost everyone — dark though their skin might be — had a European ancestor somewhere. Other countries developed fine distinctions based on colour. Latin America once recognised more than twenty races. The offspring of a Spaniard and an Indian was a mestizo, that of a mestizo and a Spaniard a castizo, a Spaniard and a negro a mulatto, a mulatto and a Spaniard a morisco, a morisco and a Spaniard an albino, an albino and a Spaniard a torna atras and so on in a lengthy, hair-splitting and subjective series. Races were supposed to be distinct because they descend from different ancestors. Ham, Shem and Japhet, the sons of Noah, were popular candidates. Anthropology began with the- search for perfect examples of each lineage, for racial types. Africans, Europeans and Asians were seen as separate versions of humankind. Perhaps, its students thought, every race was once a pure and unpolluted line, secure in its ancestral homeland. Only in modern times was that purity sullied by interbreeding. Race mixture was against nature (exceptions were allowed in emergency, as when Saints Cosima and Damian, with divine help, transplanted a black leg onto a white patient).