Genome

Chomsky's conjecture has been brilliantly vindicated in the succeeding decades by lines of evidence from many different disciplines.

9 4 G E N O M E

All converge upon the conclusion that to learn a human language requires, in the words of the psycho-linguist Steven Pinker, a human language instinct. Pinker (who has been called the first linguist capable of writing readable prose) persuasively gathered the strands of evidence for the innateness of language skills. There is first the universality of language. All human people speak languages of comparable grammatical complexity, even those isolated in the highlands of New Guinea since the Stone Age. All people are as consistent and careful in following implicit grammatical rules, even those without education and who speak what are patronisingly thought to be 'slang'

dialects. The rules of inner-city black Ebonics are just as rational as the rules of the Queen's English. To prefer one to another is mere prejudice. For example, to use double negatives ('Don't nobody do this to me . . .') is considered proper in French, but slang in English.

The rule is just as consistently followed in each.

Second, if these rules were learnt by imitation like the vocabulary, then why would four-year-olds who have been happily using the word 'went' for a year or so, suddenly start saying 'goed'? The truth is that although we must teach our children to read and write —

skills for which there is no specialised instinct — they learn to speak by themselves at a much younger age with the least of help from us. No parent uses the word 'goed', yet most children do at some time. No parent explains that the word 'cup' refers to all cup-like objects, not this one particular cup, nor just its handle, nor the material from which it is made, nor the action of pointing to a cup, nor the abstract concept of cupness, nor the size or temperature of cups. A computer that was required to learn language would have to be laboriously equipped with a program that ignored all these foolish options — with an instinct, in other words. Children come pre-programmed, innately constrained to make only certain kinds of guess.

But the most startling evidence for a language instinct comes from a series of natural experiments in which children imposed grammatical rules upon languages that lacked them. In the most famous case, studied by Derek Bickerton, a group of foreign labour-ers brought together on Hawaii in the nineteenth century developed I N S T I N C T 9 5

a pidgin language - a mixture of words and phrases whereby they could communicate with each other. Like most such pidgins, the language lacked consistent grammatical rules and remained both laboriously complex in the way it had to express things and relatively simple in what it could express. But all that changed when for the first time a generation of children learnt the language in their youth.

The pidgin acquired rules of inflection, word order and grammar that made it a far more efficient and effective language - a creole.

In short, as Bickerton concluded, pidgins become Creoles only after they are learnt by a generation of children, who bring instinct to bear on their transformation.

Bickerton's hypothesis has received remarkable support from the study of sign language. In one case, in Nicaragua, special schools for the deaf, established for the first time in the 1980s, led to the invention, de novo, of a whole new language. The schools taught lip-reading with little success, but in the playground the children brought together the various hand signs they used at home and established a crude pidgin language. Within a few years, as younger children learnt this pidgin, it was transformed into a true sign language with all the complexity, economy, efficiency and grammar of a spoken language. Once again, it was children who made the language, a fact that seems to suggest that the language instinct is one that is switched off as the child reaches adulthood. This accounts for our difficulty in learning new languages, or even new accents, as adults. We no longer have the instinct. (It also explains why it is so much harder, even for a child, to learn French in a classroom than on holiday in France: the instinct works on speech that it hears, not rules that it memorises.) A sensitive period during which something can be learnt, and outside which it cannot, is a feature of many animals' instincts. For instance, a chaffinch will only learn the true song of its species if exposed to examples between certain ages. That the same is true of human beings was proved in a brutal way by the true story of Genie, a girl discovered in a Los Angeles apartment aged thirteen. She had been kept in a single sparsely furnished room all her life and deprived of almost all human contact.

9 6 G E N O M E

She had learnt two words, 'Stopit' and 'Nomore'. After her release from this hell she rapidly acquired a larger vocabulary, but she never learnt to handle grammar - she had passed the sensitive period when the instinct is expressed.

Yet even bad ideas take a lot of killing, and the notion that language is a form of culture that can shape the brain, rather than vice versa, has been an inordinate time a-dying. Even though the canonical case histories, like the lack of a concept of time in the Hopi language and hence in Hopi thought, have been exposed as simple frauds, the notion that language is a cause rather than consequence of the human brain's wiring survives in many social sciences. It would be absurd to argue that only Germans can understand the concept of taking pleasure at another's misfortune; and that the rest of us, not having a word for Schadenfreude, find the concept entirely foreign.2

Further evidence for the language instinct comes from many sources, not least from detailed studies of the ways in which children develop language in their second year of life. Irrespective of how much they are spoken to directly, or coached in the use of words, children develop language skills in a predictable order and pattern.

And the tendency to develop language late has been demonstrated by twin studies to be highly heritable. Yet for many people the most persuasive evidence for the language instinct comes from the hard sciences: neurology and genetics. It is hard 'to argue with stroke victims and real genes. The same part of the brain is consistently used for language processing (in most people, on the left side of the brain), even the deaf who 'speak' with their hands, though sign language also uses part of the right hemisphere.3

If a particular one of these parts of the brain is damaged, the effect is known as Broca's aphasia, an inability to use or understand all but the simplest grammar, even though the ability to understand sense remains unaffected. For instance, a Broca's aphasic can easily answer questions such as 'Do you use a hammer for cutting?' but has great difficulty with: 'The lion was killed by the tiger. Which one is dead?' The second question requires sensitivity to the grammar I N S T I N C T 9 7

encoded in word order, which is known by just this one part of the brain. Damage to another area, Wernicke's area, has almost the opposite effect - people with such damage produce a rich but senseless stream of words. It appears as if Broca's area generates speech and Wernicke's area instructs Broca's area what speech to generate. This is not the whole story, for there are other areas active in language processing, notably the insula (which may be the region that malfunctions in dyslexia).4

There are two genetic conditions that affect linguistic ability. One is Williams syndrome, caused by a change in a gene on chromosome 11, in which affected children are very low in general intelligence, but have a vivid, rich and loquacious addiction to using language.

They chatter on, using long words, long sentences and elaborate syntax. If asked to refer to an animal, they are as likely to choose something bizarre like an aardvark as a cat or a dog. They have a heightened ability to learn language but at the expense of sense: they are severely mentally retarded. Their existence seems to undermine the notion, which most of us have at one time or another considered, that reason is a form of silent language.

The other genetic condition has the opposite effect: it lowers linguistic ability without apparently affecting intelligence, or at least not consistently. Known as specific language impairment (SLI), this condition is at the centre of a fierce scientific fight. It is a battleground between the new science of evolutionary psychology and the old social sciences, between genetic explanations of behaviour and environmental ones. And the gene is here on chromosome 7.

That the gene exists is not at issue. Careful analysis of twin studies unambiguously points to a strong heritability for specific language impairment. The condition is not associated with neurological damage during birth, is not associated with linguistically impoverished upbringings, and is not caused by general mental retardation. According to some tests — and depending on how it is defined — the heritability approaches one hundred per cent. That is, identical twins are roughly twice as likely to share the condition as fraternal twins.

That the gene in question is on chromosome 7 is also not in much 9 8 G E N O M E

doubt. In 1997 a team of Oxford-based scientists pinned down a genetic marker on the long arm of chromosome 7, one form of which co-occurs with the condition of S L I . The evidence, though based only on one large English family, was strong and unambiguous.6

So why the battleground? The argument rages about what SLI is. To some it is merely a general problem with the brain that affects many aspects of language-producing ability, including principally the ability to articulate words in the mouth and to hear sounds correctly in the ear. The difficulty the subjects experience with language follow from these sensory problems, according to this theory. To others, this is highly misleading. The sensory and voice problems exist, to be sure, in many victims of the condition, but so does something altogether more intriguing: a genuine problem understanding and using grammar that is quite independent of any sensory deficits. The only thing both sides can agree upon is that it is thoroughly disgrace-ful, simplistic and sensationalist of the media to portray this gene, as they have done, as a 'grammar gene'.

The story centres on a large English family known as the Ks.

There are three generations. A woman with the condition married an unaffected man and had four daughters and one son: all save one daughter were affected and they in turn had between them twenty-four children, ten of whom have the condition. This family has got to know the psychologists well; rival teams besiege them with a battery of tests. It is their blood that led the Oxford team to the gene on chromosome 7. The Oxford team, working with the Institute of Child Health in London, belongs to the 'broad' school of S L I , which argues that the grammar-deficient skills of the K

family members stem from their problems with speech and hearing.

Their principal opponent and the leading advocate of the 'grammar theory' is a Canadian linguist named Myrna Gopnik.

In 1990 Gopnik first suggested that the K family and others like them have a problem knowing the basic rules of English grammar.

It is not that they cannot know the rules, but that they must learn them consciously and by heart, rather than instinctively internalise them. For example, if Gopnik shows somebody a cartoon of an I N S T I N C T 9 9

imaginary creature and with it the words 'This is a Wug', then shows them a picture of two such creatures together with the words 'These are . . .', most people reply, quick as a flash, 'Wugs'. Those with SLI rarely do so, and if they do, it is after careful thought. The English plural rule, that you add an 's' to the end of most words, is one they seem not to know. This does not prevent those with SLI knowing the plural of most words, but they are stumped by novel words that they have not seen before, and they make the mistake of adding 's' to fictitious words that the rest of us would not, such as 'saess'. Gopnik hypothesises that they store English plurals in their minds as separate lexical entries, in the same way that we all store singulars. They do not store the grammatical rule.7

The problem is not, of course, confined to plurals. The past tense, the passive voice, various word-order rules, suffixes, word-combination rules and all the laws of English we each so unconsciously know, give SLI people difficulty, too. When Gopnik first published these findings, after studying the English family, she was immediately and fiercely attacked. It was far more reasonable, said one critic, to conclude that the source of the variable performance problems lay in the language-processing system, rather than the underlying grammar. Grammatical forms like plural and past tense were particularly vulnerable, in English, in individuals with speech defects. It was misleading of Gopnik, said another pair of critics, to neglect to report that the K family has a severe congenital speech disorder, which impairs their words, phonemes, vocabulary and semantic ability as well as their syntax. They had difficulty understanding many other forms of syntactical structure such as reversible passives, post-modified subjects, relative clauses and embedded forms.

These criticisms had a whiff of territoriality about them. The family was not Gopnik's discovery: how dare she assert novel things about them? Moreover, there was some support for her idea in at least part of the criticism: that the disorder applied to all syntactical forms. And to argue that the grammatical difficulty must be caused by the mis-speaking problem, because mis-speaking goes with the grammatical difficulty, was circular.

IOO G E N O M E

Gopnik was not one to give up. She broadened the study to Greek and Japanese people as well, using them for various ingenious experiments designed to show the same phenomena. For example, in Greek, the word 'likos' means wolf. The word 'likanthropos'

means wolfman. The word 'lik', the root of wolf, never appears on its own. Yet most Greek speakers automatically know that they must drop the '-os' to find the root if they wish to combine it with another word that begins with a vowel, like '-anthropos', or drop only the

's', to make 'liko-' if they wish to combine it with a word that begins with a consonant. It sounds a complicated rule, but even to English speakers it is immediately familiar: as Gopnik points out, we use it all the time in new English words like 'technophobia'.

Greek people with SLI cannot manage the rule. They can learn a word like 'likophobia' or 'likanthropos', but they are very bad at recognising that such words have complex structures, built up from different roots and suffixes. As a result, to compensate, they effectively need a larger vocabulary than other people. 'You have to think of them', says Gopnik, 'as people without a native language.' They learn their own tongue in the same laborious way that we, as adults, learn a foreign language, consciously imbibing the rules and words.9

Gopnik acknowledges that some SLI people have low IQ on non-verbal tests, but on the other hand some have above-average I Q . In one pair of fraternal twins, the SLI one had higher non-verbal IQ than the unaffected twin. Gopnik also acknowledges that most SLI people have problems speaking and hearing as well, but she contends that by no means all do and that the coincidence is irrelevant. For instance, people with SLI have no trouble learning the difference between 'ball' and 'bell', yet they frequently say 'fall'

when they mean 'fell' - a grammatical, not a vocabulary difference.

Likewise, they have no difficulty discerning the difference between rhyming words, like 'nose' and 'rose'. Gopnik was furious when one of her opponents described the K family members' speech as

'unintelligible' to outsiders. Having spent many hours with them, talking, eating pizza and attending family celebrations, she says they are perfectly comprehensible. To prove the irrelevance of speaking I N S T I N C T 1 0 1

and hearing difficulties, she has devised written tests, too. For example, consider the following pair of sentences: 'He was very happy last week when he was first.' 'He was very happy last week when he is first.' Most people immediately recognise that the first is grammatical and the second is not. SLI people think they are both acceptable statements. It is hard to conceive how this could be due to a hearing or speaking difficulty.10

None the less, the speaking-and-hearing theorists have not given up. They have recently shown that SLI people have problems with

'sound masking', whereby they fail to notice a pure tone when it is masked by preceding or following noise, unless the tone is forty-five decibels more intense than is detectable to other people. In other words, SLI people have more trouble picking out the subtler sounds of speech from the stream of louder sounds, so they might, for example, miss the '-ed' on the end of a word.

But instead of supporting the view that this explains the entire range of SLI symptoms, including the difficulty with grammatical rules, this lends credence to a much more interesting, evolutionary explanation: that the speech and hearing parts of the brain are next door to the grammar parts and both are damaged by S L I . SLI results from damage to the brain caused in the third trimester of pregnancy by an unusual version of a gene on chromosome 7.

Magnetic-resonance imaging confirms the existence of the brain lesion and the rough location. It occurs, not surprisingly, in one of the two areas devoted to speech and language processing, the areas known as Broca's and Wernicke's areas.

There are two areas in the brains of monkeys that correspond precisely to these areas. The Broca-homologue is used for controlling the muscles of the monkey's face, larynx, tongue and mouth. The Wernicke-homologue is used for recognising sound sequences and the calls of other monkeys. These are exactly the non-linguistic problems that many SLI people have: controlling facial muscles and hearing sounds distinctly. In other words, when ancestral human beings first evolved a language instinct, it grew in the region devoted to sound production and processing. That sound-production and 1 0 2 G E N O M E

processing module remained, with its connections to facial muscles and ears, but the language instinct module grew on top of it, with its innate capacity for imposing the rules of grammar on the vocabulary of sounds used by members of the species. Thus, although no other primate can learn grammatical language at all — and we are indebted to many diligent, sometimes gullible and certainly wishful trainers of chimpanzees and gorillas for thoroughly exhausting all possibilities to the contrary - language is intimately physically connected with sound production and processing. (Yet not too intimately: deaf people redirect the input and output of the language module to the eyes and hands respectively.) A genetic lesion in that part of the brain therefore affects grammatical ability, speech and hearing - all three modules.11

No better proof could be adduced for William James's nineteenth-century conjecture that human beings evolved their complex behaviour by adding instincts to those of their ancestors, not by replacing instincts with learning. James's theory was resurrected in the late 1980s by a group of scientists calling themselves evolutionary psychologists. Prominent among them were the anthropologist John Tooby, the psychologist Leda Cosmides and the psycho-linguist Steven Pinker. Their argument, in a nutshell, is this. The main goal of twentieth-century social science has been to trace the ways in which our behaviour is influenced by the social environment; instead, we could turn the problem on its head and trace the ways in which the social environment is the product of our innate social instincts.

Thus the fact that all people smile at happiness and frown when worried, or that men from all cultures find youthful features sexually attractive in women, may be expressions of instinct, not culture. Or the universality of romantic love and religious belief might imply that these are influenced by instinct more than tradition. Culture, Tooby and Cosmides hypothesised, is the product of individual psychology more than vice versa. Moreover, it has been a gigantic mistake to oppose nature to nurture, because all learning depends on innate capacities to learn and innate constraints upon what is learnt. For instance, it is much easier to teach a monkey (and a man) I N S T I N C T 1 0 3

to fear snakes than it is to teach it to fear flowers. But you still have to teach it. Fear of snakes is an instinct that has to be learnt.12

The 'evolutionary' in evolutionary psychology refers not so much to an interest in descent with modification, nor to the process of natural selection itself - interesting though these are, they are inaccessible to modern study in the case of the human mind, because they happen too slowly - but to the third feature of the Darwinian paradigm: the concept of adaptation. Complex biological organs can be reverse-engineered to discern what they are 'designed' to do, in just the same way that sophisticated machines can be so studied.

Steven Pinker is fond of pulling from his pocket a complicated thing designed for pitting olives to explain the process of reverse engineering. Leda Cosmides prefers a Swiss-army knife to make a similar point. In each case, the machines are meaningless except when described in terms of their particular function: what is this blade for? It would be meaningless to describe the working of a camera without reference to the fact that it is designed for the making of images. In the same way, it is meaningless to describe the human (or animal) eye without mentioning that it is specifically designed for approximately the same purpose.

Pinker and Cosmides both contend that the same applies to the human brain. Its modules, like the different blades of a Swiss-army knife, are most probably designed for particular functions. The alternative, that the brain is equipped with random complexity, from which its different functions fall out as fortunate by-products of the physics of complexity - an idea still favoured by Chomsky - defies all evidence. There is simply nothing to support the conjecture that the more detailed you make a network of microprocessors, the more functions they will acquire. Indeed, the 'connectionist' approach to neural networks, largely misled by the image of the brain as a general-purpose network of neurons and synapses, has tested the idea thoroughly and found it wanting. Pre-programmed design is required for the solving of pre-ordained problems.

There is a particular historical irony here The concept of design in nature was once one of the strongest arguments advanced against 1 0 4 G E N O M E

evolution. Indeed, it was the argument from design that kept evolutionary ideas at bay throughout the first half of the nineteenth century. Its most able exponent, William Paley, famously observed that if you found a stone on the ground, you could conclude little of interest about how it got there. But if you found a watch, you would be forced to conclude that somewhere there was a watchmaker. Thus the exquisite, functional design apparent in living creatures was manifest evidence for God. It was Darwin's genius to use the argument from design just as explicitly but in the service of the opposite conclusion: to show that Paley was wrong. A 'blind watchmaker' (in Richard Dawkins's phrase) called natural selection, acting step by step on the natural variation in the creature's body, over many millions of years and many millions of individuals, could just as easily account for complex adaptation. So successfully has Darwin's hypothesis been supported that complex adaptation is now considered the primary evidence that natural selection has been at work.13

The language instinct that we all possess is plainly one such complex adaptation, beautifully designed for clear and sophisticated communication between individuals. It is easy to conceive how it was advantageous for our ancestors on the plains of Africa to share detailed and precise information with each other at a level of sophistication unavailable to other species. 'Go a short way up that valley and turn left by the tree in front of the pond and you will find the giraffe carcass we just killed. Avoid the brush on the right of the tree that is in fruit, because we saw a lion go in there.' Two sentences pregnant with survival value to the recipient; two tickets for success in the natural-selection lottery, yet wholly incomprehensible without a capacity for understanding grammar, and lots of it.

The evidence that grammar is innate is overwhelming and diverse.

The evidence that a gene somewhere on chromosome 7 usually plays a part in building that instinct in the developing foetus's brain is good, though we have no idea how large a part that gene plays.

Yet most social scientists remain fervently resistant to the idea of genes whose primary effect seems to be to achieve the development I N S T I N C T 1 0 5

of grammar direcdy. As is clear in the case of the gene on chromosome 7, many social scientists prefer to argue, despite much evidence, that the gene's effects on language are mere side-effects of its direct effect on the ability of the brain to understand speech. After a century in which the dominating paradigm has been that instincts are confined to 'animals' and are absent from human beings, this reluctance is not surprising. This whole paradigm collapses once you consider the Jamesian idea that some instincts cannot develop without learnt, outside inputs.

This chapter has followed the arguments of evolutionary psychology, the reverse-engineering of human behaviour to try to understand what particular problems it was selected to solve. Evolutionary psychology is a new and remarkably successful discipline that has brought sweeping new insights to the study of human behaviour in many fields. Behaviour genetics, which was the subject of the chapter on chromosome 6, aims at roughly the same goal.

But the approach to the subject is so different that behaviour genetics and evolutionary psychology are embarked on a collision course.

The problem is this: behaviour genetics seeks variation between individuals and seeks to link that variation to genes. Evolutionary psychology seeks common human behaviour — human universals, features found in every one of us — and seeks to understand how and why such behaviour must have become partly instinctive. It therefore assumes no individual differences exist, at least for important behaviours. This is because natural selection consumes variation: that is its job. If one version of a gene is much better than another, then the better version will soon be universal to the species and the worse version will soon be extinct. Therefore, evolutionary psychology concludes that if behaviour geneticists find a gene with common variation in it, then it may not be a very important gene, merely an auxiliary. Behaviour geneticists retort that every human gene yet investigated turns out to have variants, so there must be something wrong with the argument from evolutionary psychology.

In practice, it may gradually emerge that the disagreement between these two approaches is exaggerated. One studies the genetics of 1 0 6 G E N O M E

common, universal, species-specific features. The other studies the genetics of individual differences. Both are a sort of truth. All human beings have a language instinct, whereas all monkeys do not, but that instinct does not develop equally well in all people. Somebody with SLI is still far more capable of learning language than Washoe, Koko, Nim or any of the other trained chimpanzees and gorillas.

The conclusions of both behaviour genetics and evolutionary psychology remain distinctly unpalatable to many non-scientists, whose main objection is a superficially reasonable argument from incredulity. How can a gene, a stretch of D N A 'letters', cause a behaviour? What conceivable mechanism could link a recipe for a protein with an ability to learn the rule for making the past tense in English? I admit that this seems at first sight a mighty leap, requiring more faith than reason. But it need not be, because the genetics of behaviour is, at root, no different from the genetics of embryonic development. Suppose that each module of the brain grows its adult form by reference to a series of chemical gradients laid down in the developing embryo's head - a sort of chemical road map for neurons. Those chemical gradients could themselves be the product of genetic mechanisms. Hard though it is to imagine genes and proteins that can tell exactly where they are in the embryo, there is no doubting they exist. As I shall reveal when discussing chromosome 12, such genes are one of the most exciting products of modern genetic research. The idea of genes for behaviour is no more strange than the idea of genes for development. Both are mind-boggling, but nature has never found human incomprehension a reason for changing her methods.

C H R O M O S O M E S X A N D Y

C o n f l i c t

Xq28 — Thanks for the genes mom.

T shirt sold in gay and lesbian

bookstores in the mid-1990s

A detour into linguistics has brought us face to face with the startling implications of evolutionary psychology. If it has left you with an unsettling feeling that something else is in control, that your own abilities, linguistic and psychological, were somewhat more instinctively determined than you proudly imagined, then things are about to get a lot worse. The story of this chapter is perhaps the most unexpected in the whole history of genetics. We have got used to thinking of genes as recipes, passively awaiting transcription at the discretion of the collective needs of the whole organism: genes as servants of the body. Here we encounter a different reality. The body is the victim, plaything, battleground and vehicle for the ambitions of genes.

The next largest chromosome after number seven, is called the X chromosome. X is the odd one out, the misfit. Its pair, the chromosome with which it has some affinity of sequence, is not, as I o 8 G E N O M E

in every other case, an identical chromosome, but is the Y chromosome, a tiny and almost inert stub of a genetic afterthought. At least that is the case in male mammals and flies, and in female butterflies and birds. In female mammals or male birds there are instead two X chromosomes, but they are still somewhat eccentric. In every cell in the body, instead of both expressing their genetic message at equal volume, one of the two at random packs itself up into a tight bundle known as a Barr body and remains inert.

The X and Y chromosomes are known as the sex chromosomes for the obvious reason that they determine, with almost perfect predestination, the sex of the body. Everybody gets an X chromosome from his or her mother. But if you inherited a Y chromosome from your father, you are a man; if you inherited an X chromosome from your father, you are a woman. There are rare exceptions, superficially female people with an X and a Y, but they are exceptions that prove the rule. The key masculinising gene on the Y

chromosome is missing or broken in such people.

Most people know this. It does not take much exposure to school biology to come across the X and Y chromosomes. Most people also know that the reason colour-blindness, haemophilia and some other disorders are much more common in men is that these genes are on the X chromosome. Since men have no 'spare' X chromosome, they are much more likely to suffer from these recessive problems than women — as one biologist has put it, the genes on the X chromosome fly without co-pilots in men. But there are things about the X and Y chromosomes most people do not know, disturbing, strange things that have unsettled the very foundations of biology.

It is not often that you find language like this in one of the most sober and serious of all scientific publications, the Philosophical Transactions of the Royal Society: 'The mammalian Y chromosome is thus likely to be engaged in a battle in which it is outgunned by its opponent. A logical consequence is that the Y should run away and hide, shedding any transcribed sequences that are not essential to its function.'1 'A battle', 'outgunned', 'opponent', 'run away'? These C O N F L I C T 1 0 9

are hardly the sort of things we can expect molecules of D N A to do. Yet the same language, a little more technically phrased, appears in another scientific paper about the Y chromosome, entitled 'The enemies within: intergenomic conflict, interlocus contest evolution ( I C E ) , and the intraspecific Red Queen'.2 The paper reads, in part:

'Perpetual I C E between the Y and the rest of the genome can thereby continually erode the genetic quality of the Y via genetic hitchhiking of mildly deleterious mutations. The decay of the Y is due to genetic hitchhiking, but it is the I C E process that acts in a catalytic way to continually drive the male versus female anatagonistic coevolution.' Even if most of this is Greek to you, there are certain words that catch the eye: words like 'enemies' and 'antagonism'.

Then there is a recent textbook on the same material. Its title, quite simply, is "Evolution: the four billion year war'.3 What is going on?

At some point in our past, our ancestors switched from the common reptilian habit of determining sex by the temperature of the egg to determining it genetically. The probable reason for the switch was so that each sex could start training for its special role at conception. In our case, the sex-determining gene made us male and the lack of it left us female, whereas in birds it happened the other way round. The gene soon attracted to its side other genes that benefited males: genes for big muscles, say, or aggressive tendencies.

But because these were not wanted in females — wasting energy they would prefer to spend on offspring - these secondary genes found themselves at an advantage in one sex and at a disadvantage in the other. They are known in the trade as sexually antagonistic genes.

The dilemma was solved when another mutant gene suppressed the normal process of swapping of genetic material between the two paired chromosomes. Now the sexually antagonistic genes could diverge and go their different ways. The version on the Y chromosome could use calcium to make antlers; the version on the X

chromosome could use calcium to make milk. Thus, a pair of middle-si2ed chromosomes, once home to all sorts of 'normal' genes, was hijacked by the process of sex determination and became the sex chromosomes, each attracting different sets of genes. On the Y

1 1 0 G E N O M E

chromosome, genes accumulate that benefit males but are often bad for females; on the X accumulate genes that are good for females and deleterious in males. For instance, there is a newly discovered gene called DAX, found on the X chromosome. A few rare people are born with one X and one Y chromosome, but with two copies of the DAX gene on the X chromosome. The result is, that although such people are genetically male, they develop into normal females.

The reason, it transpires, is that DAX and SKY — the gene on the Y chromosome that makes men into men — are antagonistic to each other. One SRY defeats one DAX, but two DAXes defeat one SRY.4

This outbreak of antagonism between genes is a dangerous situ-ation. Lurching into metaphor, one might begin to discern that the two chromosomes no longer have each other's interests at heart, let alone those of the species as a whole. Or, to put it more correctly, something can be good for the spread of a gene on the X chromosome that actually damages the Y chromosome or vice versa.

Suppose, for instance, that a gene appeared on the X chromosome that specified the recipe for a lethal poison that killed only sperm carrying Y chromosomes. A man with such a gene would have no fewer children than another man. But he would have all daughters and no sons. All of those daughters would carry the new gene, whereas if he had had sons as well, none of them would have carried it. Therefore, the gene is twice as common in the next generation as it would otherwise be. It would spread very rapidly. Such a gene would only cease to spread when it had exterminated so many males that the very survival of the species was in jeopardy and males were at a high premium.5

Far-fetched? Not at all. In the butterfly Acrea encedon, that is exactly what has happened. The sex ratio is ninety-seven per cent female as a result. This is just one of many cases known of this form of evolutionary conflict, known as sex-chromosome drive. Most known instances are confined to insects, but only because scientists have looked more closely at insects. The strange language of conflict used in the remarks I quoted above now begins to make more sense. A C O N F L I C T I I I

piece of simple statistics: because females have two X chromosomes while males have an X and a Y, three-quarters of all sex chromosomes are Xs; one-quarter are Ys. Or, to put it another way, an X

chromosome spends two-thirds of its time in females, and only one-third in males. Therefore, the X chromosome is three times as likely to evolve the ability to take pot shots at the Y as the Y is to evolve the ability to take pot shots at the X. Any gene on the Y

chromosome is vulnerable to attack by a newly evolved driving X

gene. The result has been that the Y chromosome has shed as many genes as possible and shut down the rest, to 'run away and hide' (in the technical jargon used by William Amos of Cambridge University).

So effectively has the human Y chromosome shut down most of its genes that the great bulk of its length consists of non-coding D N A , serving no purpose at all - but giving few targets for the X

chromosome genes to aim at. There is a small region that seems to have slipped across from the X chromosome fairly recently, the so-called pseudo-autosomal region, and then there is one immensely important gene, the SRY gene mentioned above. This gene begins the whole cascade of events that leads to the masculinisation of the embryo. Rarely can a single gene have acquired such power.

Although it only throws a switch, much else follows from that. The genitals grow to look like a penis and testes, the shape and constitution of the body are altered from female (the default in our species, though not in birds and butterflies), and various hormones go to work on the brain. There was a spoof map of the Y chromosome published in the journal Science a few years ago, which purported to have located genes for such stereotypically male traits as flipping between television channels, the ability to remember and tell jokes, an interest in the sports pages of newspapers, an addiction to death and destruction movies and an inability to express affection over the phone - among others. The joke is funny, though, only because we recognise these habits as male, and therefore far from mocking the idea that such habits are genetically determined, the joke reinforces the idea. The only thing wrong with the diagram is that these male behaviours come not from specific genes for each of 1 1 2 G E N O M E

them, but from the general masculinisation of the brain by hormones such as testosterone which results in a tendency to behave this way in the modern environment. Thus, in a sense, many masculine habits are all the products of the SRY gene itself, which sets in train the series of events that lead to the masculinisation of the brain as well as the body.

The SRY gene is peculiar. Its sequence is remarkably consistent between different men: there are virtually no point mutations (i.e., one-letter spelling differences) in the human race. SRY is, in that sense, a variation-free gene that has changed almost not at all since the last common ancestor of all people 200,000 years ago or so. Yet our SRY is very different from that of a chimpanzee, and different again from that of a gorilla: there is, between species, ten times as much variation in this gene as is typical for other genes. Compared with other active (i.e., expressed) genes, SRY is one of the fastest evolving.

How do we explain this paradox? According to William Amos and John Harwood, the answer lies in the process of fleeing and hiding that they call selective sweeps. From time to time, a driving gene appears on the X chromosome that attacks the Y chromosome by recognising the protein made by SRY. At once there is a selective advantage for any rare SRY mutant that is sufficiently different to be unrecognised. This mutant begins to spread at the expense of other males. The driving X chromosome distorts the sex ratio in favour of females but the spread of the new mutant SRY restores the balance. The end result is a brand new SRY gene sequence shared by all members of the species, with little variation. The effect of this sudden burst of evolution (which might happen so quickly as to leave few traces in the evolutionary record) would be to produce SRYs that were very different between species, but very similar within species. If Amos and Harwood are right, at least one such sweep must have occurred since the splitting of chimp ancestors and human ancestors, five to ten million years ago, but before the ancestor common to all modern human beings, 200,000 years ago.6

You may be feeling a little disappointed. The violence and conflict C O N F L I C T 1 1 3

that I promised at the beginning of the chapter turn out to be little more than a detailed piece of molecular evolution. Fear not. I am not finished yet, and I plan to link these molecules to real, human conflict soon enough.

The leading scholar of sexual antagonism is William Rice of the University of California at Santa Cruz and he has completed a remarkable series of experiments to make the point explicit. Let us go back to our putative ancestral creature that has just acquired a distinct Y chromosome and is in the process of shutting down many of the genes on it to escape driving X genes. This nascent Y

chromosome, in Rice's phrase, is now a hotspot for male-benefit genes. Because a Y chromosome will never find itself in a female, it is free to acquire genes that are very bad for females so long as they are at least slightly good for males (if you still thought evolution was about the good of the species, stop thinking so right now). In fruit flies, and for that matter in human beings, male ejaculate consists of sperm cells suspended in a rich soup called the seminal fluid.

Seminal fluid contains proteins, products of genes. Their purpose is entirely unknown, but Rice has a shrewd idea. During fruit-fly sex, those proteins enter the bloodstream of the female and migrate to, among other places, her brain. There they have the effect of reducing the female's sexual appetite and increasing her ovulation rate. Thirty years ago, we would have explained that increase in terms of the good of the species. It is time for the female to stop seeking sexual partners and instead seek a nesting site. The male's seminal fluid redirects her behaviour to that end. You can hear the National Geographic commentary. Nowadays, this information takes on a more sinister aura. The male is trying to manipulate the female into mating with no other males and into laying more eggs for his sperm and he is doing so at the behest of sexually antagonistic genes, probably on the Y chromosome (or switched on by genes on the Y chromosome). The female is under selective pressure to be more and more resistant to such manipulation. The outcome is a stalemate.

Rice did an ingenious experiment to test his idea. For twenty-nine 1 1 4 G E N O M E

generations, he prevented female flies from evolving resistance: he kept a separate strain of females in which no evolutionary change occurred. Meanwhile, he allowed males to generate more and more effective seminal fluid proteins by testing them against more and more resistant females. After twenty-nine generations he brought the two lines together again. The result was a walkover. Male sperm was now so effective at manipulating female behaviour that it was effectively toxic: it could kill the females.7

Rice now believes that sexual antagonism is at work in all sorts of environments. It leaves its signature as rapidly evolving genes. In the shellfish the abalone, for instance, the lysin protein that the sperm uses to bore a hole through the glycoprotein matrix of the egg is encoded by a gene that changes very rapidly (the same is probably true in us), probably because there is an arms race between the lysin and the matrix. Rapid penetration is good for sperm but bad for the egg, because it allows parasites or second sperm through.

Coming slightly closer to home, the placenta is controlled by rapidly evolving genes (and paternal ones, at that). Modern evolutionary theorists, led by David Haig, now think of the placenta as more like a parasitic takeover of the mother's body by paternal genes in the foetus. The placenta tries, against maternal resistance, to control her blood-sugar levels and blood pressure to the benefit of the foetus.8

More on this in the chapter on chromosome 15.

But what about courtship behaviour? The traditional view of the peacock's elaborate tail is that it is a device designed to seduce females and that it is in effect designed by ancestral females' preferences. Rice's colleague, Brett Holland, has a different explanation.

He thinks peacocks did indeed evolve long tails to seduce females, but that they did so because females grew more and more resistant to being so seduced. Males in effect use courtship displays as a substitute for physical coercion and females use discrimination to retain control over their own frequency and timing of mating. This explains a startling result from two species of wolf spiders. One species has tufts of bristles on its forelegs that it uses in courtship.

Shown a video of a male spider displaying, the female will indicate C O N F L I C T 1 1 5

by her behaviour whether the display turns her on. If the videos are altered so that the males' tufts disappear, the female is still just as likely to find the display arousing. But in another species, where there are no tufts, the artificial addition of tufts to males on the video more than doubled the acceptance rate of females. In other words, females gradually evolve so that they are turned off, not on, by the displays of males of their own species. Sexual selection is thus an expression of sexual antagonism between genes for seduction and genes for resistance.9

Rice and Holland come to the disturbing conclusion that the more social and communicative a species is, the more likely it is to suffer from sexually antagonistic genes, because communication between the sexes provides the medium in which sexually antagonistic genes thrive. The most social and communicative species on the planet is humankind. Suddenly it begins to make sense why relations between the human sexes are such a minefield, and why men have such vastly different interpretations of what constitutes sexual harassment from women. Sexual relations are driven not by what is good, in evolutionary terms, for men or for women, but for their chromosomes. The ability to seduce a woman was good for Y chromosomes in the past; the ability to resist seduction by a man was good for X chromosomes in the past.

This kind of conflict between complexes of genes (the Y chromosome being one such complex), does not just apply to sex. Suppose that there is a version of a gene that increases the telling of lies (not a very realistic proposition, but there might be a large set of genes that affect truthfulness indirectly). Such a gene might thrive by making its possessors into successful con-artists. But then suppose there is also a version of a different gene (or set of genes) that improves the detecting of lies, perhaps on a different chromosome.

That gene would thrive to the extent that it enabled its possessors to avoid being taken in by con-artists. The two would evolve antagonistically, each gene encouraging the other, even though it would be quite possible for the same person to possess both. There is between them what Rice and Holland call 'interlocus contest 1 1 6 G E N O M E

evolution', or I C E . Exactly such a competitive process probably did indeed drive the growth of human intelligence over the past three million years. The notion that our brains grew big to help us make tools or start fires on the savannah has long since lost favour.

Instead, most evolutionists believe in the Machiavellian theory —

that bigger brains were needed in an arms race between manipulation and resistance to manipulation. 'The phenomena we refer to as intelligence may be a byproduct of intergenomic conflict between genes mediating offense and defense in the context of language', write Rice and Holland.10

Forgive the digression into intelligence. Let's get back to sex.

Probably one of the most sensational, controversial and hotly disputed genetic discoveries was the announcement by Dean Hamer in 1993 that he had found a gene on the X chromosome that had a powerful influence on sexual orientation, or, as the media quickly called it, 'a gay gene'.11 Hamer's study was one of several published about the same time all pointing towards the conclusion that homosexuality was 'biological' — as opposed to being the consequence of cultural pressure or conscious choice. Some of this work was done by gay men themselves, such as the neuroscientist Simon LeVay of the Salk Institute, keen to establish in the public mind what they were convinced about in their own minds: that homosexuals were

'born that way'. They believed, with some justice, that prejudice would be less against a lifestyle that was not a deliberate 'choice'

but an innate propensity. A genetic cause would also make homosexuality seem less threatening to parents by making it clear that gay role models could not turn youths gay unless they had the propensity already. Indeed conservative intolerance of homosexuality has recently taken to attacking the evidence for its genetic nature.

"We should be careful about accepting the claim that some are "born to be gay", not just because it is untrue, but because it provides leverage to homosexual rights organisations', wrote the Conservative Lady Young in the Daily Telegraph on 29 July 1998.

But however much some of the researchers may have desired a particular outcome, the studies are objective and sound. There is C O N F L I C T 1 1 7

no room for doubt that homosexuality is highly heritable. In one study, for example, among fifty-four gay men who were fraternal twins, there were twelve whose twin was also gay; and among fifty-six gay men who were identical twins, there were twenty-nine whose twin was also gay. Since twins share the same environment, whether they are fraternal or identical, such a result implies that a gene or genes accounts for about half of the tendency for a man to be gay.

A dozen other studies came to a similar conclusion.12

Intrigued, Dean Hamer decided to seek the genes that were involved. He and his colleagues interviewed no families with gay male members and noticed something unusual. Homosexuality seemed to run in the female line. If a man was gay, the most likely other member of the previous generation to be gay was not his father but his mother's brother.

That immediately suggested to Hamer that the gene might be on the X chromosome, the only set of nuclear genes a man inherits exclusively from his mother. By comparing a set of genetic markers between gay men and straight men in the families in his sample, he quickly found a candidate region in Xq2 8, the tip of the long arm of the chromosome. Gay men shared the same version of this marker seventy-five per cent of the time; straight men shared a different version of the marker seventy-five per cent of the time.

Statistically, that ruled out coincidence with ninety-nine per cent confidence. Subsequent results reinforced the effect, and ruled out any connection between the same region and lesbian orientation.

To canny evolutionary biologists, such as Robert Trivers, the suggestion that such a gene might lie on the X chromosome immediately rang a bell. The problem with a gene for sexual orientation is that the version that causes homosexuality would quite quickly become extinct. Yet it is plainly present in the modern population at a significant level. Perhaps four per cent of men are definitively gay (and a smaller percentage bisexual). Since gay men, are, on average, less likely to have children than straight men, the gene would be doomed to have long since dwindled in frequency to vanishing point unless it carried some compensating advantage.

I I 8 G E N O M E

Trivers argued that, because an X chromosome spends twice as much time in women as it does in men, a sexually antagonistic gene that benefited female fertility could survive even if it had twice as large a deleterious effect on male fertility. Suppose, for example, that the gene Hamer had found determined age of puberty in women, or even something like breast size (remember, this is just a thought experiment). Each of those characteristics might affect female fertility. Back in the Middle Ages, large breasts might mean more milk, or might attract a richer husband whose children were less likely to die in infancy. Even if the same version of the same gene reduced male fertility by making sons attracted to other men, such a gene could survive because of the advantage it gave daughters.

Until Hamer's gene itself is found and decoded, the link between homosexuality and sexual antagonism is no more than a wild guess.

Indeed, it remains a possibility that the connection between Xq28

and sexuality is misleading. Michael Bailey's recent research on homosexual pedigrees has failed to find a maternal bias to be a general feature. Other scientists, too, have failed to find Hamer's link with Xq28. At present it looks as if it may have been confined to those families Hamer studied. Hamer himself cautions that until the gene is in the bag, it is a mistake to assume otherwise.14

Besides, there is now a complicating factor: a completely different explanation of homosexuality. It is becoming increasingly clear that sexual orientation correlates with birth order. A man with one or more elder brothers is more likely to be gay than a man with no siblings, only younger siblings, or with one or more elder sisters.

The birth order effect is so strong that each additional elder brother increases the probability of homosexuality by roughly one-third (this can still mean a low probability: an increase from three to four per cent is an increase of thirty-three per cent). The effect has now been reported in Britain, the Netherlands, Canada and the United States, and in many different samples of people.15

For most people, the first thought would be a quasi-Freudian one: that something in the dynamics of growing up in a family with elder brothers might predispose you towards homosexuality. But, C O N F L I C T 1 1 9

as so often, the Freudian reaction is almost certainly the wrong one. (The old Freudian idea that homosexuality was caused by a protective mother and a distant father almost certainly confused cause and effect: the boy's developing effeminate interests repel the father and the mother becomes overprotective in compensation.) The answer probably lies, once more, in the realm of sexual antagonism.

An important clue lies in the fact that there is no such birth-order effect for lesbians, who are randomly distributed within their families. In addition, the number of elder sisters is also irrelevant in predicting male homosexuality. There is something specific to occupying a womb that has already held other males which increases the probability of homosexuality. The best explanation concerns a set of three active genes on the Y chromosome called the H-Y

minor histocompatibility antigens. A similar gene encodes a protein called anti-Mullerian hormone, a substance vital to the masculinisation of the body: it causes the regression of the Mullerian ducts in the male embryo — these being the precursors of the womb and Fallopian tubes. What the three H-Y genes do is not certain. They are not essential for the masculinisation of the genitals, which is achieved by testosterone and anti-Mullerian hormone alone. The significance of this is now beginning to emerge.

The reason these gene products are called antigens is because they are known to provoke a reaction from the immune system of the mother. As a result, the immune reaction is likely to be stronger in successive male pregnancies (female babies do not produce H-Y

antigens, so do not raise the immune reaction). Ray Blanchard, one of those who studies the birth-order effect, argues that the H-Y

antigens' job is to switch on other genes in certain tissues, in particular in the brain - and indeed there is good evidence that this is true in mice. If so, the effect of a strong immune reaction against these proteins from the mother would be partly to prevent the masculinisation of the brain, but not that of the genitals. That in turn might cause them to be attracted to other males, or at least not attracted to females. In an experiment in which baby mice were immunised 1 2 0 G E N O M E

against H-Y antigens, they grew up to be largely incapable of successful mating, compared with controls, though frustratingly the experimenter did not report the reasons why. Likewise, male fruit flies can be irreversibly induced to show only female sexual behaviour by the switching on at a crucial point in development of a gene called 'transformer'.16

People are not mice or flies, and there is plenty of evidence that the sexual differentiation of the human brain continues after birth.

Homosexual men are clearly not, except in very rare cases, 'mental'

women trapped inside 'physical' men. Their brains must have been at least partly masculinised by hormones. It remains possible, however, that they missed some hormone during some early and crucial sensitive period and that this permanently affects some functions, including sexual orientation.

The man who first set in train the ideas that led to sexual antagonism, Bill Hamilton, understood how profoundly it shook our notions of what genes are: 'There had come the realisation', he wrote later,

'that the genome wasn't the monolithic data bank plus executive team devoted to one project - keeping oneself alive, having babies

- that I had hitherto imagined it to be. Instead, it was beginning to seem more a company boardroom, a theatre for a power struggle of egoists and factions.' Hamilton's new understanding of his genes began to affect his understanding of his mind:17

My own conscious and seemingly indivisible self was turning out far from what I had imagined and I need not be so ashamed of my self-pity! I was an ambassador ordered abroad by some fragile coalition, a bearer of conflicting orders from the uneasy masters of a divided empire . . . As I write these words, even so as to be able to write them, I am pretending to a unity that, deep inside myself, I now know does not exist. I am fundamentally mixed, male with female, parent with offspring, warring segments of chromosomes that interlocked in strife millions of years before the River Severn ever saw the Celts and Saxons of Housman's poem ['A Shropshire Lad'].

C O N F L I C T 1 2 1

The idea of genes in conflict with each other, the notion of the genome as a sort of batleefield between parental genes and childhood genes, or between male genes and female genes, is a little-known story outside a small group of evolutionary biologists. Yet it has profoundly shaken the philosophical foundations of biology.

C H R O M O S O M E 8

S e l f - i n t e r e s t

We are survival machines — robot vehicles blindly programmed to preserve the selfish molecules known as genes. This is a truth that still fills me with astonishment.

Richard Dawkins, The Selfish Gene

Instruction manuals that come with new gadgets are notoriously frustrating. They never seem to have the one piece of information you need, they send you round in circles, they leave you high and dry, and they definitely lose something in the translation from Chinese. But at least they do not insert, just when you are getting to the bit that matters, five copies of Schiller's 'Ode to Joy' or a garbled version of a set of instructions for how to saddle a horse.

Nor do they (generally) include five copies of a complete set of instructions for how to build a machine that would copy out just that set of instructions. Nor do they break the actual instructions you seek into twenty-seven different paragraphs interspersed with long pages of irrelevant junk so that even finding the right instructions is a massive task. Yet that is a description of the human S E L F - I N T E R E S T 1 2 3

retinoblastoma gene and, as far as we know, it is typical of human genes: twenty-seven brief paragraphs of sense interrupted by twenty-six long pages of something else.

Mother Nature concealed a dirty little secret in the genome. Each gene is far more complicated than it needs to be, it is broken up into many different 'paragraphs' (called exons) and in between lie long stretches (called introns) of random nonsense and repetitive bursts of wholly irrelevant sense, some of which contain real genes of a completely different (and sinister) kind.

The reason for this textual confusion is that the genome is a book that wrote itself, continually adding, deleting and amending over four billion years. Documents that write themselves have unusual properties. In particular, they are prone to parasitism. Analogies become far-fetched at this point, but try to imagine a writer of instruction manuals who arrives at his computer each morning to find paragraphs of his text clamouring for his attention. The ones that shout loudest bully him into including another five copies of themselves on the next page he writes. The true instructions still have to be there, or the machine will never be assembled, but the manual is full of greedy, parasitic paragraphs taking advantage of the writer's compliance.

Actually, with the advent of email, the analogy is no longer as far-fetched as it once was. Suppose I sent you an email that read: 'Beware, there is a nasty computer virus about; if you open a message with the word "marmalade" in the title, it will erase your hard disk! Please pass this warning on to everybody you can think of.' I made up the bit about the virus; there are, so far as I know, no emails called 'marmalade'

doing the rounds. But I have very effectively hijacked your morning and caused you to send on my warning. My email was the virus.

So far, each chapter of this book has concentrated on a gene or genes, tacitly assuming that they are the things that matter in the genome. Genes, remember, are stretches of D N A that comprise the recipe for proteins. But ninety-seven per cent of our genome does not consist of true genes at all. It consists of a menagerie 1 2 4 G E N O M E

of strange entities called pseudogenes, retropseudogenes, satellites, minisatellites, microsatellites, transposons and retrotransposons: all collectively known as 'junk D N A ' , or sometimes, probably more accurately, as 'selfish D N A ' . Some of these are genes of a special kind, but most are just chunks of D N A that are never transcribed into the language of protein. Since the story of this stuff follows naturally from the tale of sexual conflict related in the last chapter, this chapter will be devoted to junk D N A .

Fortunately this is a good place to tell the story, because I have nothing more particular to say about chromosome 8. That is not to imply that it is a boring chromosome, or that it possesses few genes, just that none of the genes yet found on chromosome 8 has caught my rather impatient attention. (For its size, chromosome 8 has been relatively neglected, and is one of the least mapped chromosomes.) Junk D N A is found on every chromosome. Yet, ironically, junk D N A is the first part of the human genome that has found a real, practical, everyday use in the human world. It has led to D N A fingerprinting.

Genes are protein recipes. But not all protein recipes are desirable.

The commonest protein recipe in the entire human genome is the gene for a protein called reverse transcriptase. Reverse transcriptase is a gene that serves no purpose at all as far as the human body is concerned. If every copy of it were carefully and magically removed from the genome of a person at the moment of conception, the person's health, longevity and happiness would be more likely to be improved than damaged. Reverse transcriptase is vital for a certain kind of parasite. It is an extremely useful — nay essential - part of the genome of the AIDS virus: a crucial contributor to its ability to infect and kill its victims. For human beings, in contrast, the gene is a nuisance and a threat. Yet it is one of the commonest genes in the whole genome. There are several hundred copies of it, possibly thousands, spread about the human chromosomes. This is an astonishing fact, akin to discovering that the commonest use of cars is for getting away from crimes. Why is it there?

A clue comes from what reverse transcriptase does. It takes an S E L F - I N T E R E S T 1 2 5

RNA copy of a gene, copies it back into D N A and stitches it back into the genome. It is a return ticket for a copy of a gene. By this means the A I D S virus can integrate a copy of its own genome into human D N A the better to conceal it, maintain it and get it efficiently copied. A good many of the copies of the reverse transcriptase gene in the human genome are there because recognisable 'retroviruses'

put them there, long ago or even relatively recently. There are several thousand nearly complete viral genomes integrated into the human genome, most of them now inert or missing a crucial gene. These

'human endogenous retroviruses' or Hervs, account for 1.3% of the entire genome. That may not sound like much, but 'proper' genes account for only 3%. If you think being descended from apes is bad for your self-esteem, then get used to the idea that you are also descended from viruses.

But why not cut out the middle man? A viral genome could drop most of the virus's genes and keep just the reverse transcriptase gene. Then this streamlined parasite could give up the laborious business of trying to jump from person to person in spit or during sex, and instead just hitchhike down the generations within its victims' genomes. A true genetic parasite. Such 'retrotransposons' are far commoner even than retroviruses. The commonest of all is a sequence of 'letters' known as a L I N E - 1 . This is a 'paragraph' of D N A , between a thousand and six thousand 'letters' long, that includes a complete recipe for reverse transcriptase near the middle.

L I N E - 1 s are not only very common - there may be 100,000 copies of them in each copy of your genome — but they are also gregarious, so that the paragraph may be repeated several times in succession on the chromosome. They account for a staggering 14.6% of the entire genome, that is, they are nearly five times as common as

'proper' genes. The implications of this are terrifying. L I N E - 1 s have their own return tickets. A single L I N E - 1 can get itself transcribed, make its own reverse transcriptase, use that reverse transcriptase to make a D N A copy of itself and insert that copy anywhere among the genes. This is presumably how there come to be so many copies of L I N E - 1 in the first place. In other words, 1 2 6 G E N O M E

this repetitive 'paragraph' of 'text' is there because it is good at getting itself duplicated - no other reason.

'A flea hath smaller fleas that on him prey; and these have smaller fleas to bite 'em, and so proceed ad infinitum.' If L I N E - 1 s are about, they, too, can be parasitised by sequences that drop the reverse transcriptase gene and use the ones in L I N E - 1 s . Even commoner than L I N E - 1 s are shorter 'paragraphs' called Alus. Each Alu contains between 180 and 280 'letters', and seems to be especially good at using other people's reverse transcriptase to get itself duplicated. The Alu text may be repeated a million times in the human genome - amounting to perhaps ten per cent of the entire 'book'.2

For reasons that are not entirely clear, the typical Alu sequence bears a close resemblance to a real gene, the gene for a part of a protein-making machine called the ribosome. This gene, unusually, has what is called an internal promoter, meaning that the message

' R E A D M E ' is written in a sequence in the middle of the gene.

It is thus an ideal candidate for proliferation, because it carries the signal for its own transcription and does not rely on landing near another such promoter sequence. As a result, each Alu gene is probably a 'pseudogene'. Pseudogenes are, to follow a common analogy, rusting wrecks of genes that have been holed below the waterline by a serious mutation and sunk. They now lie on the bottom of the genomic ocean, gradually growing rustier (that is, accumulating more mutations) until they no longer even resemble the gene they once were. For example, there is a rather nondescript gene on chromosome 9, which, if you take a copy of it and then probe the genome for sequences that resemble this gene, you will find at fourteen locations on eleven chromosomes: fourteen ghostly hulks that have sunk. They were redundant copies that, one after another, mutated and stopped being used. The same may well be true of most genes — that for every working gene, there are a handful of wrecked copies elsewhere in the genome. The interesting thing about this particular set of fourteen is that they have been sought not just in people, but in monkeys, too. Three of the human pseudogenes were sunk after the split between Old-World monkeys and S E L F - I N T E R E S T I 2 7

New-World monkeys. That means, say the scientists breathlessly, they were relieved of their coding functions 'only' around thirty-five million years ago.3

Alus have proliferated wildly, but they too have done so in comparatively recent times. Alus are found only in primates, and are divided into five different families, some of which have appeared only since the chimpanzees and we parted company (that is, within the last five million years). Other animals have different short repetitive 'paragraphs'; mice have ones called B1s.

All this information about L I N E - 1 s and Alus amounts to a major and unexpected discovery. The genome is littered, one might almost say clogged, with the equivalent of computer viruses, selfish, parasitic stretches of letters which exist for the pure and simple reason that they are good at getting themselves duplicated. We are full of digital chain letters and warnings about marmalade. Approximately thirty-five per cent of human D N A consists of various forms of selfish D N A , which means that replicating our genes takes thirty-five per cent more energy than it need. Our genomes badly need worming.

Nobody suspected this. Nobody predicted that when we read the code for life we would find it so riddled with barely controlled examples of selfish exploitation. Yet we should have predicted it, because every other level of life is parasitised. There are worms in animals' guts, bacteria in their blood, viruses in their cells. Why not retrotransposons in their genes? Moreover, by the mid-1970s, it was dawning on many evolutionary biologists, especially those interested in behaviour, that evolution by natural selection was not much about competition between species, not much about competition between groups, not even mostly about competition between individuals, but was about competition between genes using individuals and occasionally societies as their temporary vehicles. For instance, given the choice between a safe, comfortable and long life for the individual or a risky, tiring and dangerous attempt to breed, virtually all animals (and indeed plants) choose the latter. They choose to shorten their odds of death in order to have offspring. Indeed, their bodies are designed with planned obsolescence called ageing that causes 1 2 8 G E N O M E

them to decay after they reach breeding age — or, in the case of squid or Pacific salmon, to die at once. None of this makes any sense unless you view the body as a vehicle for the genes, as a tool used by genes in their competition to perpetuate themselves. The body's survival is secondary to the goal of getting another generation started. If genes are 'selfish replicators' and bodies are their disposable 'vehicles' (in Richard Dawkins's controversial terminology), then it should not be much of a surprise to find some genes that achieve their replication without building their own bodies. Nor should it be a surprise to find that genomes, like bodies, are habitats replete with their own version of ecological competition and co-operation.

Truly, in the 1970s for the first time, evolution became genetic.

To explain the fact that the genome contained huge gene-less regions, two pairs of scientists suggested in 1980 that these regions were replete with selfish sequences whose only function was survival within the genome. 'The search for other explanations may prove', they said, 'if not intellectually sterile, ultimately futile.' For making this bold forecast, they were much mocked at the time. Geneticists were still stuck in the mindset that if something were in the human genome it must serve a human purpose, not a selfish purpose of its own. Genes were just protein recipes. It made no sense to think of them as having goals or dreams. But the suggestion has been spectacularly vindicated. Genes do indeed behave as if they have selfish goals, not consciously, but retrospectively: genes that behave in this way thrive and genes that don't don't.4

A segment of selfish D N A is not just a passenger, whose presence adds to the size of the genome and therefore to the energy cost of copying the genome. Such a segment is also a threat to the integrity of genes. Because selfish D N A is in the habit of jumping from one location to another, or sending copies to new locations, it is apt to land in the middle of working genes, messing them up beyond recognition, and then jumping out again causing the mutation to revert. This was how transposons were first discovered, in the late 1940s, by the far-sighted and much neglected geneticist Barbara McClintock (she was eventually awarded the Nobel prize in 1983).

S E L F - I N T E R E S T 1 2 9

She noticed that mutations in the colour of maize seeds occur in such a manner that can only be explained by mutations jumping into and out of pigment genes.5

In human beings, L I N E - 1 s and Alus have caused mutations by landing in the middle of all sorts of genes. They have caused haemophilia, for instance, by landing in clotting-factor genes. But, for reasons that are not well understood, as a species we are less troubled by D N A parasites than some other species. Approximately 1 in every 700 human mutations is caused by 'jumping genes', whereas in mice nearly ten per cent of mutations are caused by jumping genes. The potential danger posed by jumping genes was dramatically illustrated by a sort of natural experiment in the 1950s in the tiny fruit fly, Drosophila. The fruit fly is the favourite experimental animal for geneticists. The species they study, called Drosophila melanogaster, has been transported all over the world to be bred in laboratories.

It has frequently escaped and has met other, native species of fruit fly. One of these species, called Drosophila willistoni, carries a jumping gene called a P element. Somehow in about 1950, somewhere in South America, perhaps via a blood-sucking mite, Drosophila willistoni's jumping gene entered the Drosophila melanogaster species. (One of the great concerns attached to so-called 'xeno-transplants' of organs from pigs or baboons is that they might unleash a new form of jumping gene upon our species, like the P element of fruit flies.) The P element has since spread like wildfire, so that most fruit flies have the P element, though not those collected from the wild before 1950 and kept in isolation since. The P element is a piece of selfish D N A that shows its presence by disrupting the genes into which it jumps. Gradually, the rest of the genes in the fruit fly's genome have fought back, inventing ways of suppressing the P element's jumping habit. The P elements are settling down as passengers.

Human beings possess nothing so sinister as a P element, at least not at the moment. But a similar element, called 'sleeping beauty', has been found in salmon. Introduced into human cells in the laboratory it thrives, demonstrating cut-and-paste ability. And something similar to the spread of the P element probably happened with 1 3 0 G E N O M E

each of the nine human Alu elements. Each spread through the species, disrupting genes until the other genes asserted their common interest in suppressing it, whereupon it settled down in its present fairly quiescent state. What we see in the human genome is not some rapidly advancing parasitic infection, but the dormant cysts of many past parasites, each of which spread rapidly until the genome found a way of suppressing them, but not excising them.

In this respect (as in others) we seem to be more fortunate than fruit flies. We appear to have a general mechanism for suppressing selfish D N A , at least if you believe a controversial new theory. The suppression mechanism goes by the name of cytosine methylation.

Cytosine is the letter C of the genetic code. Mefhylating it (literally by attaching a methyl group of carbon and hydrogen atoms) prevents it from being transcribed by the reader. Much of the genome spends large chunks of the time in the methylated - blocked - state, or rather most gene promoters do (the parts at the beginning of the gene where transcription starts). It has generally been assumed that methylation serves to switch off genes that are not needed in particular tissues, thus making the brain different from the liver, which is different from the skin and so on. But a rival explanation is gaining ground. Methylation may have almost nothing to do with tissue-specific expression and much to do with suppressing transposons and other intragenomic parasites. Most methylation lies within transposons such as Alu and L I N E - 1 . The new theory holds that during the early development of the embryo, all genes are briefly stripped of any methylation and switched on. This is then followed by a close inspection of the whole genome by molecules whose job is to spot repetitive sequences and close them down with methylation.

In cancer tumours, one of the first things to happen is demethylation of the genes. As a result, the selfish D N A is released from its handcuffs and richly expressed in tumours. Since they are good at messing up other genes, these transposons then make the cancer worse. Methylation, according to this argument, serves to suppress the effect of selfish D N A . 6

L I N E - 1 is generally about 1,400 'letters' long. Alu is generally S E L F - I N T E R E S T 1 3 1

at least 180 'letters' long. There are, however, sequences even shorter than Alu that also accumulate in vast, repetitive stutters. It is perhaps too far-fetched to call these shorter sequences parasites, but they proliferate in roughly the same manner - that is, they are there because they contain a sequence that is good at getting itself duplicated. It is one of these short sequences that has a practical use in forensic and other sciences. Meet the 'hypervariable minisatellite'.

This neat little sequence is found on all the chromosomes; it crops up at more than one thousand locations in the genome. In every case the sequence consists of a single 'phrase', usually about twenty

'letters' long, repeated over and over again many times. The 'word'

can vary according to the location and the individual, but it usually contains the same central 'letters': GGGCAGGAXG (where X can be any 'letter'). The significance of this sequence is that it is very similar to one that is used by bacteria to initiate the swapping of genes with other bacteria of the same species, and it seems to be involved in the encouragement of gene swapping between chromosomes in us as well. It is as if each sequence is a sentence with the words ' S W A P ME A B O U T ' in the middle.

Here is an example of a repetition of a minisatellite: hxckswapmeaboutlopl-hxckswapmeaboutlopl-hxckswapmeaboutlopl-hxckswapmeaboutlopl-

hxckswapmeaboutlopl-hxckswapmeaboutiopl-

hxckswapmeaboutlopl-hxckswapmeaboutlopl-

hxckswapmeaboutlopl-hxckswapmeaboutlopl. Ten repeats in this case. Elsewhere, at each of one thousand locations, there might be fifty or five repeats of the same phrase. Following instructions, the cell starts swapping the phrases with the equivalent series on the other copy of the same chromosome. But in doing so it makes fairly frequent mistakes, adding or subtracting to the number of repeats.

In this way each series of repeats gradually changes length, fast enough so that it is different in every individual, but slowly enough so that people mostly have the same repeat lengths as their parents.

Since there are thousands of series, the result is a unique set of numbers for each individual.

1 3 2 G E N O M E

Alec Jeffreys and his technician Vicky Wilson stumbled on minisatellites in 1984, largely by accident. They were studying how genes evolve by comparing the human gene for the muscle protein myoglobin with its equivalent from seals when they noticed a stretch of repetitious D N A in the middle of the gene. Because each minisatellite shares the same core sequence of twelve letters, but because the number of repeats can vary so much, it is a relatively simple matter to fish out this minisatellite array and compare the size of the array in different individuals. It turns out that the repeat number is so variable that everybody has a unique genetic fingerprint: a string of black marks looking just like a bar code. Jeffreys immediately spotted the significance of what he had found. Neglecting the myoglobin genes that were the target of his study, he started investigating what could be done with unique genetic fingerprints. Because strangers have such different genetic fingerprints, immigration authorities were immediately interested in testing the claims of would-be immigrants that they were close relatives of people already in the country.

Genetic fingerprinting proved that they were generally telling the truth, which eased much misery. But a more dramatic use was to follow soon after.7

On 2 August 1986, a young schoolgirl's body was found in a thorn thicket close to the village of Narborough, in Leicestershire.

Dawn Ashworth, aged fifteen, had been raped and strangled. A week later, the/police arrested a young hospital porter, Richard Buckland, who confessed to the murder. There the matter would have rested.

Buckland would have gone to prison, convicted of the killing. However, the police were anxious to clear up an unsolved case, of a girl named Lynda Mann, also fifteen, also from Narborough, also raped, strangled and left in an open field, but nearly three years before.

The murders were so similar it seemed implausible that they had not been committed by the same man. But Buckland refused to confess to Mann's murder.

Word of Alec Jeffreys's fingerprinting breakthrough had reached the police via the newspapers, and since he worked in Leicester, less than ten miles from Narborough, the local police contacted S E L F - I N T E R E S T 1 3 3

Jeffreys and asked him if he could confirm the guilt of Buckland in the Mann case. He agreed to try. The police supplied him with semen taken from both girls' bodies and a sample of Buckland's blood.

Jeffreys had little difficulty finding various minisatellites in each sample. After more than a week's work the genetic fingerprints were ready. The two semen samples were identical and must have come from the same man. Case closed. But what Jeffreys saw next astonished him. The blood sample had a radically different fingerprint from the semen samples: Buckland was not the murderer.

The Leicestershire police protested heatedly that this was an absurd conclusion and that Jeffreys must have got it wrong. Jeffreys repeated the test and so did the Home Office forensic laboratory, with exactly the same result. Reluctantly, the baffled police withdrew the case against Buckland. For the first time in history a man was exonerated on the basis of his D N A sequences.

But nagging doubts remained. Buckland had, after all, confessed and policemen would find genetic fingerprinting a lot more convincing if it could convict the guilty as well as acquit the innocent. So, five months after Ashworth's death, the police set out to test the blood of 5,500 men in the Narborough area to look for a genetic fingerprint that matched that of the murdering rapist's sperm. No sample matched.

Then a man who worked in a Leicester bakery named Ian Kelly happened to remark to his colleagues that he had taken the blood test even though he lived nowhere near Narborough. He had been asked to do so by another worker in the bakery, Colin Pitchfork, who did live in Narborough. Pitchfork claimed to Kelly that the police were trying to frame him. One of Kelly's colleagues repeated the tale to the police, who arrested Pitchfork. Pitchfork quickly confessed to killing both girls, but this time the confession proved true: the D N A fingerprint of his blood matched that of the semen found on both bodies. He was sentenced on 23 January 1988 to life in prison.

Genetic fingerprinting immediately became one of forensic 1 3 4 G E N O M E

science's most reliable and potent weapons. The Pitchfork case, an extraordinary virtuoso demonstration of the technique, set the tone for years to come: genetic fingerprinting's ability to acquit the innocent, even in the face of what might seem overwhelming evidence of guilt; its ability to flush out the guilty just by the threat of its use; its amazing precision and reliability — if properly used; its reliance on small samples of bodily tissue, even nasal mucus, spit, hair or bone from a long-dead corpse.

Genetic fingerprinting has come a long way in the decade since the Pitchfork case. In Britain alone, by mid-1998 320,000 samples of D N A had been collected by the Forensic Science Service and used to link 28,000 people to crime scenes. Nearly twice as many samples have been used to exonerate innocent people. The technique has been simplified, so that single sites of minisatellites can be used instead of many. Genetic fingerprinting has also been amplified, so that tiny minisatellites or even microsatellites can be used to give unique 'bar codes'. Not only the lengths but the actual sequences of the minisatellite repeats can be analysed to give greater sophistication. Such D N A typing has also been misused or discredited in court, as one might expect when lawyers are involved. (Much of the misuse reflects public naivety with statistics, rather than anything to do with the D N A : nearly four times as many potential jurors will convict if told that a D N A match has a chance probability of 0.1 per cent than if told one in a thousand men match the D N A

- yet they are the same facts.8)

D N A fingerprinting has revolutionised not just forensic science but all sorts of other fields as well. It was used to confirm the identity of the exhumed corpse of Josef Mengele in 1990. It was used to confirm the presidential parenthood of the semen on Monica Lewinsky's dress. It was used to identify the illegitimate descendants of Thomas Jefferson. It has so blossomed in the field of paternity testing, both by officials publicly and by parents privately, that in 1998 a company called Identigene placed billboards by freeways all over America reading: ' W H O ' S T H E F A T H E R ? C A L L

1-800-DNA-TYPE'. They received 300 calls a day asking for their S E L F - I N T E R E S T 1 3 5

$600 tests, both from single mothers trying to demand child-support from the 'fathers' of their children and from suspicious 'fathers'

unsure if their partner's children were all theirs. In more than two-thirds of cases the D N A evidence showed that the mother was telling the truth. It is a moot point whether the offence caused to some fathers by discovering that their partners were unfaithful outweighs the reassurance others receive that their suspicions were unfounded. Britain, predictably, had a fierce media row when the first such private service set up shop: in Britain such medical technologies are supposed to remain the property of the state, not the individual.9

More romantically, the application of genetic fingerprinting to paternity testing has revolutionised our understanding of bird song.

Have you ever noticed that thrushes, robins and warblers continue singing long after they have paired up in spring? This flies in the face of the conventional notion that bird song's principal function is the attraction of a mate. Biologists began DNA-testing birds in the late 1980s, trying to determine which male had fathered which chicks in each nest. They discovered, to their surprise, that in the most monogamous of birds, where just one male and one female faithfully help each other to rear the brood, the female's mate quite often with neighbouring males other than their ostensible 'spouses'.

Cuckoldry and infidelity are much, much commoner than anybody expected (because they are committed in great secrecy). D N A fingerprinting led to an explosion of research into a richly rewarding theory known as sperm competition, which can "explain such trivia as the fact that chimpanzee testicles are four times the size of gorilla testicles, even though chimpanzees are one-quarter the size of gorillas. Male gorillas monopolise their mates, so their sperm meets no competitors; male chimpanzees share their mates, so each needs to produce large quantities of sperm and mate frequently to increase his chances of being the father. It also explains why male birds sing so hard when already 'married'. They are looking for

'affairs'.10

C H R O M O S O M E 9

D i s e a s e

A desperate disease requires a dangerous remedy.

Guy Fawkes

On chromosome 9 lies a very well-known gene: the gene that determines your ABO blood group. Since long before there was D N A fingerprinting, blood groups have appeared in court.

Occasionally, the police get lucky and match the blood of the criminal to blood found at the scene of the crime. Blood grouping presumes innocence. That is to say, a negative result can prove you were not the murderer absolutely, but a positive one can only suggest that you might be the murderer.

Not that this logic had much impact on the California Supreme Court, which in 1946 ruled that Charlie Chaplin was most definitely the father of a certain child despite unambiguous proof from the incompatibility of their blood groups that he could not have been.

But then judges were never very good at science. In paternity suits as well as murder cases, blood grouping, like genetic fingerprinting, or indeed fingerprinting, is the friend of the innocent. In the days of D N A fingerprinting, blood-group forensics is redundant. Blood D I S E A S E 137

groups are much more important in transfusion, though again in a wholly negative way: receiving the wrong blood can be fatal. And blood groups can give us insights into the history of human migrations, though once more they have been almost entirely superseded in this role by other genes. So you might think blood groups are rather dull. You would be wrong. Since 1990 they have found an entirely new role: they promise understanding of how and why our genes are all so different. They hold the key to human polymorphism.

The first and best known of the blood group systems is the A B O

system. First discovered in 1900, this system originally had three different names with confusing consequences: type I blood, according to Moss's nomenclature was the same as type IV blood according to Jansky's nomenclature. Sanity gradually prevailed and the nomenclature adopted by the Viennese discoverer of the blood groups became universal: A, B, AB and O. Karl Landsteiner expressively described the disaster that befell a wrong transfusion thus: 'lytischen und agglutinierenden Wirkungen des Blutserums'. The red cells all stick together. But the relation between the blood groups was not simple. People with type A blood could safely donate to those with A or A B ; those with B could donate to those with B and A B ; those with AB could donate only to those with A B ; and those with O blood could donate to anybody - O is therefore known as the universal donor. Nor was there any obvious geographic or racial reason underlying the different types. Roughly forty per cent of Europeans have type O blood, forty per cent have type A blood, fifteen per cent have type B blood and five per cent have type AB

blood. The proportions are similar in other continents, with the marked exception of the Americas, where the native American population was almost exclusively type O, save for some Canadian tribes, who were very often type A, and Eskimos, who were sometimes type AB or B.

It was not until the 1920s that the genetics of the A B O blood groups fell into place, and not until 1990 that the gene involved came to light. A and B are 'co-dominant' versions of the same gene, O being the 'recessive' form of it. The gene lies on chromosome I 3 8 G E N O M E

9, near the end of the long arm. Its text is 1,062 'letters' long, divided into six short and one long exons ('paragraphs') scattered over several 'pages' - 18,000 letters in all — of the chromosome. It is a medium-sized gene, then, interrupted by five longish introns. The gene is the recipe for galactosyl transferase,1 an enzyme, i.e. a protein with the ability to catalyse a chemical reaction.

The difference between the A gene and the B gene is seven letters out of 1,062, of which three are synonymous or silent: that is, they make no difference to the amino acid chosen in the protein chain.

The four that matter are letters 523, 700, 793 and 800. In people with type A blood these letters read C, G, C, G. In people with type B blood they read G, A, A, C. There are other, rare differences.

A few people have some of the A letters and some of the B letters, and a rare version of the A type exists in which a letter is missing near the end. But these four little differences are sufficient to make the protein sufficiently different to cause an immune reaction to the wrong blood.2

The O group has just a single spelling change compared with A, but instead of a substitution of one letter for another, it is a deletion.

In people with type O blood, the 258th letter, which should read

'G', is missing altogether. The effect of this is far-reaching, because it causes what is known as a reading-shift or frame-shift mutation, which is far more consequential. (Recall that if Francis Crick's ingenious comma-free code of 1957 had been correct, reading-shift mutations would not have existed.) The genetic code is read in three-letter words and has no punctuation. An English sentence written in three-letter words might read something like: the fat cat sat top mat and big dog ran bit cat. Not exactly poetry, I admit, but it will do. Change one letter and it still makes fairly good sense: the fat xat sat top mat and big dog ran bit cat. But delete the same letter instead, and read the remaining letters in groups of three, and you render the whole sentence meaningless: the fat ats att opm ata ndb igd ogr anb itc at. This is what has happened to the A B O

gene in people with the O blood group. Because they lack just one letter fairly early in the message, the whole subsequent message says D I S E A S E 139

something completely different. A different protein is made with different properties. The chemical reaction is not catalysed.

This sounds drastic, but it appears to make no difference at all.

People with type O blood are not noticeably disadvantaged in any walk of life. They are not more likely to get cancer, be bad at sports, have little musical ability or something. In the heyday of eugenics, no politician called for the sterilisation of people with the O blood group. Indeed, the remarkable thing about blood groups, the thing that has made them so useful and so politically neutral, is that they seem to be completely invisible; they correlate with nothing.

But this is where things get interesting. If blood groups are invisible and neutral, then how did they evolve to the present state? Was it pure chance that landed the inhabitants of the Americas with type O blood? At first glance the blood groups seem to be an example of the neutral theory of evolution, promulgated by Motoo Kimura in 1968: the notion that most genetic diversity is there because it makes no difference, not because it has been picked by natural selection for a purpose. Kimura's theory was that mutation pumps a continual stream of mutations that do not affect anything into the gene pool, and that they are gradually purged again by genetic drift

- random change. So there is constant turnover without adaptive significance. Return to earth in a million years and large chunks of the human genome would read differently for entirely neutral reasons.

'Neutralists' and 'selectionists' for a while grew quite exercised about their respective beliefs, and when the dust settled Kimura was left with a respectable following. Much variation does indeed seem to be neutral in its effects. In particular, the closer scientists look at how proteins change, the more they conclude that most changes do not affect the 'active site' where the protein does its chemical tricks. In one protein, there have been 250 genetic changes since the Cambrian age between one group of creatures and another, yet only six of them matter at all.3

But we now know the blood groups are not as neutral as they seem. There is indeed a reason behind them. From the early 1960s, 1 4 0 G E N O M E

it gradually became apparent that there was a connection between blood groups and diarrhoea. Children with type A blood fell victim to certain strains of infant diarrhoea but not to others; children with type B blood fell victim to other strains; and so on. In the late 1980s, people with the O group were discovered to be much more susceptible to infection with cholera. Dozens of studies later, the details grow more distinct. Not only are those people with type O

blood susceptible, but those with A, B and AB differ in their susceptibility. The most resistant people are those with the AB

genotype, followed by A, followed by B. All of these are much more resistant than those with O. So powerful is this resistance in AB

people that they are virtually immune to cholera. It would be irresponsible to say that people with type AB blood can safely drink from a Calcutta sewer — they might get another disease - but it is true that even if these people did pick up the Vibrio bacterium that causes cholera and it settled in their gut, they would not get diarrhoea.

Nobody yet knows how the AB genotype offers protection against this most virulent and lethal of human diseases, but it presents natural selection with an immediate and fascinating problem.

Remember that we each have two copies of each chromosome, so A people are actually AAs, that is they have an A gene on each of their ninth chromosomes, and B people are actually BBs. Now imagine a population with just these three kinds of blood groups: A A , BB and A B . The A gene is better for cholera resistance than the B gene. AA people are therefore likely to have more surviving children than BB people. Therefore the B gene is likely to die out

- that's natural selection. But it doesn't happen like that, because AB people survive best of all. So the healthiest children will be the offspring of A A s and BBs. All their children will be A B , the most cholera-resistant type. But even if an AB mates with another A B , only half their children will be A B ; the rest will be AA and B B , the latter being the most susceptible type. It is a world of strangely fluctuating fortunes. The very combination that is most beneficial in your generation guarantees you some susceptible children.

D I S E A S E 1 4 1

Now imagine what happens if everybody in one town is A A , but a newcomer arrives who is B B . If she can fend off the cholera long enough to breed, she will have AB children, who will be resistant.

In other words, the advantage will always lie with the rare version of the gene, so neither version can become extinct because if it becomes rare, it comes back into fashion. This is known, in the trade, as frequency-dependent selection, and it seems to be one of the commonest reasons that we are all so genetically diverse.

This explains the balance between A and B. But if O blood makes you more susceptible to cholera, then why has natural selection not driven the O mutation extinct? The answer probably lies with a different disease, malaria. People with type O blood seem to be slightly more resistant to malaria than people of other blood groups.

They also seem to be slightly less likely to get cancers of various kinds. This enhanced survival was probably enough to keep the O

version of the gene from disappearing, despite its association with susceptibility to cholera. A rough balance was struck between the three variations on the blood group gene.

The link between disease and mutations was first noticed in the late 1940s by an Oxford graduate student with a Kenyan background, Anthony Allison. He suspected that the frequency of a disease called sickle-cell anaemia in Africa might be connected with the prevalence of malaria. The sickle-cell mutation, which causes blood cells to collapse in the absence of oxygen, is frequently fatal to those with two copies of it, but only mildly harmful to those with just one copy. But those with one copy are largely resistant to malaria. Allison tested the blood of Africans living in malarial areas and found that those with the mutation were far less likely to have the malaria parasite as well. The sickle-cell mutation is especially common in parts of west Africa where malaria has long been endemic, and is common also in African-Americans, some of whose ancestors came from west Africa in the slave ships. Sickle-cell disease is a high price paid today for malaria resistance in the past. Other forms of anaemia, such as the thalassaemia common in various parts of the Mediterranean and south-east Asia, appear to have a similar protective effect 1 4 2 G E N O M E

against malaria, accounting for its presence in regions once infested with the disease.

The haemoglobin gene, where the sickle-cell mutation occurs as just a single-letter change, is not alone in this respect. According to one scientist, it is the tip of an iceberg of genetic resistance to malaria. Up to twelve different genes may vary in their ability to confer resistance to malaria. Nor is malaria alone. At least two genes vary in their ability to confer resistance to tuberculosis, including the gene for the vitamin D receptor, which is also associated with a variability in susceptibility to osteoporosis. 'Naturally', writes Adrian Hill of Oxford University,4 'We can't resist suggesting that natural selection for TB resistance in the recent past may have increased the prevalence of susceptibility genes for osteoporosis.'

Meanwhile, a newly discovered but similar connection links the genetic disease cystic fibrosis with the infectious disease typhoid.

The version of the C F T R gene on chromosome 7 that causes cystic fibrosis — a dangerous disease of the lungs and intestines — protects the body against typhoid, an intestinal disease caused by a Salmonella bacterium. People with just one such version do not get cystic fibrosis, but they are almost immune to the debilitating dysentery and fever caused by typhoid. Typhoid needs the usual version of the C F T R gene to get into the cells it infects; the altered version, missing three D N A letters, is no good to it. By killing those with other versions of the gene, typhoid put natural pressure on the altered version to spread. But because people inheriting two copies of the altered version were lucky to survive at all, the gene could never be very common. Once again, a rare and nasty version of a gene was maintained by disease.5

Approximately one in five people are genetically unable to release the water-soluble form of the A B O blood group proteins into their saliva and other body fluids. These 'non-secretors' are more likely to suffer from various forms of disease, including meningitis, yeast infection and recurrent urinary tract infection. But they are less likely to suffer from influenza or respiratory syncitial virus. Wherever you D I S E A S E 143

look, the reasons behind genetic variability seem to have something to do with infectious disease.6

We have barely scratched the surface of this subject. As they scourged our ancestors, the great epidemic diseases of the past -

plague, measles, smallpox, typhus, influenza, syphilis, typhoid, chicken pox, and others - left behind their imprint on our genes.

Mutations which granted resistance thrived, but that resistance often came at a price, the price varying from severe (sickle-cell anaemia) to theoretical (the inability to receive transfusions of the wrong type of blood).

Indeed, until recently, doctors were in the habit of underestimating the importance of infectious disease. Many diseases that are generally thought to be due to environmental conditions, occupation, diet or pure chance are now beginning to be recognised as the side-effects of chronic infections with little known viruses or bacteria. The most spectacular case is stomach ulcers. Several drug companies grew rich on new drugs intended to fight the symptoms of ulcers, when all that were needed all along were antibiotics. Ulcers are caused by Helicobacter pylori, a bacterium usually acquired in childhood, rather than by rich food, anxiety or misfortune. Likewise, there are strong suggestive links between heart disease and infection with chlamydia or herpes virus, between various forms of arthritis and various viruses, even between depression or schizophrenia and a rare brain virus called Borna disease virus that usually infects horses and cats.

Some of these correlations may prove misleading and in other cases the disease may attract the microbe rather than the other way round.

But it is a proven fact that people vary in their genetic resistance to things like heart disease. Perhaps these genetic variants, too, relate to resistance to infection.7

In a sense the genome is a written record of our pathological past, a medical scripture for each people and race. The prevalence of O blood groups in native Americans may reflect the fact that cholera and other forms of diarrhoea, which are diseases associated with crowded and insanitary conditions, never established themselves in the newly populated continents of the western hemisphere 1 4 4 G E N O M E

before relatively modern times. But then cholera was a rare disease probably confined to the Ganges delta before the 1830s, when it suddenly spread to Europe, the Americas and Africa. We need a better explanation of the puzzling prevalence of the O version of the gene in native Americans, especially given the fact that the blood of ancient pre-Columbian mummies from North America seems quite often to be of the A or B type. It is almost as if the A and B genes were rapidly driven extinct by a different selection pressure unique to the western hemisphere. There are hints that the cause might be syphilis, a disease that seems to be indigenous to the Americas (this is still hotly disputed in medical-history circles, but the fact remains that syphilitic lesions are known in North American skeletons from before 1492, but not in European skeletons from before that date). People with the O version of the gene seem to be less susceptible to syphilis than those with other blood types.8

Now consider a bizarre discovery that would have made little sense before the discovery of the association between susceptibility to cholera and blood groups. If, as a professor, you ask four men and two women each to wear a cotton T-shirt, no deodorant and no perfume, for two nights, then hand these T-shirts to you, you will probably be humoured as a mite kinky. If you then ask a total of 121 men and women to sniff the armpits of these dirty T-shirts and rank them according to attractiveness of smell, you will be considered, to put it mildly, eccentric. But true scientists should not be embarrassable. The result of exactly such an experiment, by Claus Wederkind and Sandra Furi, was the discovery that men and women most prefer (or least dislike) the body odour of members of the opposite sex who are most different from them genetically. Wederkind and Furi looked at M H C genes on chromosome 6, which are the genes involved in the definition of self and the recognition of parasitic intruders by the immune system. They are immensely variable genes. Other things being equal, a female mouse will prefer to mate with a male that has maximally different M H C genes from herself, a fact she discerns by sniffing his urine. It was this discovery that alerted Wederkind and Furi to the possibility that we, too, might D I S E A S E 1 4 5

retain some such ability to choose our mates on the basis of their genes. Only women on the contraceptive pill failed to show a clear preference for different M H C genotypes in male-impregnated T-shirt armpits. But then the pill is known to affect the sense of smell. As Wedekind and Furi put it,9 'No one smells good to everybody; it depends on who is sniffing whom.'

The mouse experiment had always been interpreted in terms of outbreeding: the female mouse tries to find a male from a genetically different population, so that she can have offspring with varied genes and little risk of inbred diseases. But perhaps she - and T-shirt-sniffing people - are actually doing something that makes sense in terms of the blood-group story. Remember that, when making love in a time of cholera, an AA person is best off looking for a BB mate, so that all their children will be cholera-resistant ABs. If the same sort of system applies to other genes and their co-evolution with other diseases - and the M H C complex of genes seems to be the principal site of disease-resistance genes - then the advantage of being sexually attracted to a genetic opposite is obvious.

The Human Genome Project is founded upon a fallacy. There is no such thing as 'the human genome'. Neither in space nor in time can such a definite object be defined. At hundreds of different loci, scattered throughout the twenty-three chromosomes, there are genes that differ from person to person. Nobody can say that the blood group A is 'normal' and O, B and AB are 'abnormal'. So when the Human Genome Project publishes the sequence of the typical human being, what will it publish for the A B O gene on chromosome 9? The project's declared aim is to publish the average or

'consensus' sequence of 200 different people. But this would miss the point in the case of the A B O gene, because it is a crucial part of its function that it should not be the same in everybody. Variation is an inherent and integral part of the human - or indeed any —

genome.

Nor does it make sense to take a snapshot at this particular moment in 1999 and believe that the resulting picture somehow represents a stable and permanent image. Genomes change.

1 4 6 G E N O M E

Different versions of genes rise and fall in popularity driven often by the rise and fall of diseases. There is a regrettable human tendency to exaggerate stability, to believe in equilibrium. In fact the genome is a dynamic, changing scene. There was a time when ecologists believed in 'climax' vegetation - oak forests for England, fir forests for Norway. They have learnt better. Ecology, like genetics, is not about equilibrium states. It is about change, change and change.

Nothing stays the same forever.

The first person who half glimpsed this was probably J. B. S.

Haldane, who tried to find a reason for the abundance of human genetic variation. As early as 1949 he conjectured that genetic variation might owe a good deal to the pressures of parasites. But Haldane's Indian colleague, Suresh Jayakar, rocked the boat in 1970

by suggesting that there need be no stability, and that parasites could cause a perpetual cycling fluctuation in gene frequencies. By the 1980s the torch had passed to the Australian Robert May, who demonstrated that even in the simplest system of a parasite and its host, there might be no equilibrium outcome: that eternal chaotic motion could flow from a deterministic system. May thus became one of the fathers of chaos theory. The baton was picked up by the Briton William Hamilton, who developed mathematical models to explain the evolution of sexual reproduction, models that relied upon a genetic arms race between parasites and their hosts, and which resulted in what Hamilton called 'the permanent unrest of many [genes]'.10

Some time in the 1970s, as happened in physics half a century before, the old world of certainty, stability and determinism in biology fell. In its place we must build a world of fluctuation, change and unpredictability. The genome that we decipher in this generation is but a snapshot of an ever-changing document. There is no definitive edition.

C H R O M O S O M E 1 0

S t r e s s

This is the excellent foppery of the world, that, when we are sick in fortune — often the surfeit of our own behaviour, — we make guilty of our disasters the sun, the moon, and the stars; as if we were villains by necess-ity, fools by heavenly compulsion . . . an admirable eva-sion of whoremaster man, to lay his goatish disposition to the charge of a star. William Shakespeare, King Lear The genome is a scripture in which is written the past history of plagues. The long struggles of our ancestors with malaria and dysentery are recorded in the patterns of human genetic variation. Your chances of avoiding death from malaria are pre-programmed in your genes, and in the genes of the malaria organism. You send out your team of genes to play the match, and so does the malaria parasite.

If their attackers are better than your defenders, they win. Bad luck.

No substitutes allowed.

But it is not like that, is it? Genetic resistance to disease is the last resort. There are all sorts of simpler ways of defeating disease.

Sleep under a mosquito net, drain the swamps, take a pill, spray 1 4 8 G E N O M E

D D T around the village. Eat well, sleep well, avoid stress, keep your immune system in good health and generally maintain a sunny disposition. All of these things are relevant to whether you catch an infection. The genome is not the only battlefield. In the last few chapters I have fallen into the habit of reductionism. I have taken the organism apart to isolate its genes and discern their particular interests. But no gene is an island. Each one exists as part of an enormous confederation called the body. It is time to put the organism back together again. It is time to visit a much more social gene, a gene whose whole function is to integrate some of the many different functions of the body, and a gene whose existence gives the lie to the mind-body dualism that plagues our mental image of the human person. The brain, the body and the genome are locked, all three, in a dance. The genome is as much under the control of the other two as they are controlled by it. That is partly why genetic determinism is such a myth. The switching on and off of human genes can be influenced by conscious or unconscious external action.

Cholesterol — a word pregnant with danger. The cause of heart disease; bad stuff; red meat. You eat it, you die. Nothing could be more wrong than this equation of cholesterol with poison. Cholesterol is an essential ingredient of the body. It lies at the centre of an intricate system of biochemistry and genetics that integrates the whole body. Cholesterol is a small organic compound that is soluble in fat but not in water. The body manufactures most of its cholesterol from sugars in the diet, and could not survive without it. From cholesterol at least five crucial hormones are made, each with a very different task: progesterone, aldosterone, Cortisol, testosterone and oestradiol. Collectively, they are known as the steroids. The relationship between these hormones and the genes of the body is intimate, fascinating and unsettling.

Steroids have been used by living creatures for so long that they probably pre-date the split between plants, animals and fungi. The hormone that triggers the shedding of an insect's skin is a steroid.

So is the enigmatic chemical known in human medicine as vitamin D. Some synthetic, or anabolic, steroids can be manufactured to S T R E S S 149

trick the body into suppressing inflammation, while others can be used for building athletes' muscles. Yet other steroids, derived originally from plants, can mimic human hormones sufficiently well to be used as oral contraceptives. Others still, products of the chemical industry, may be responsible for the ferninisation of male fish in polluted streams and the falling sperm counts of modern men.

There is a gene on chromosome 10 called CYP17. It makes an enzyme, which enables the body to convert cholesterol into Cortisol, testosterone and oestradiol. Without the enzyme, the pathway is blocked and the only hormones that can be made from cholesterol are progesterone and corticosterone. People who lack a working copy of this gene cannot make other sex hormones so they fail to go through puberty; if genetically male, they look like girls.

But put the sex hormones on one side for a moment and consider the other hormone that is made using CYP17. Cortisol. Cortisol is used in virtually every system in the body, a hormone that literally integrates the body and the mind by altering the configuration of the brain. Cortisol interferes with the immune system, changes the sensitivity of the ears, nose and eyes, and alters various bodily functions. When you have a lot of Cortisol coursing through your veins, you are - by definition — under stress. Cortisol and stress are virtually synonymous.

Stress is caused by the outside world, by an impending exam, a recent bereavement, something frightening in the newspaper or the unremitting exhaustion of caring for a person with Alzheimer's disease. Short-term stressors cause an immediate increase in epinephrine and norepinephrine, the hormones that make the heart beat faster, the feet go cold. These hormones prepare the body for 'fight or flight' in an emergency. Stressors that last for longer activate a different pathway that results in a much slower, but more persistent increase in Cortisol. One of Cortisol's most surprising effects is that it suppresses the working of the immune system. It is a remarkable fact that people who have been preparing for an important exam, and have shown the symptoms of stress, are more likely to catch colds and other infections, because one of the effects of Cortisol is to reduce the 1 5 0 G E N O M E

activity, number and lifetime of lymphocytes — white blood cells.

Cortisol does this by switching genes on. It only switches on genes in cells that have Cortisol receptors in them, which have in turn been switched on by some other triggers. The genes that it switches on mostly switch on other genes in turn, and sometimes the genes that they switch on will then switch on other genes and so on. The secondary effects of Cortisol can involve tens, or maybe even hundreds, of genes. But the Cortisol was only made in the first place because a series of genes was switched on in the adrenal cortex to make the enzymes necessary for making Cortisol - among them CYP17. It is a system of mindboggling complexity: if I started to list even the barest outlines of the actual pathways I would bore you to tears. Suffice to say that you cannot produce, regulate and respond to Cortisol without hundreds of genes, nearly all of which work by switching each other on and off. It is a timely lesson that the main purpose of most genes in the human genome is regulating the expression of other genes in the genome.

I promised not to bore you, but let me just take a quick glimpse at one of the effects of Cortisol. In white blood cells Cortisol is almost certainly involved in switching on a gene called TCF, also on chromosome 10, thus enabling TCF to make its own protein, whose job is to suppress the expression of another protein called interleukin 2, and interleukin 2 is a chemical that puts white blood cells on alert to be especially vigilant for germs. So Cortisol suppresses the immune alertness of white blood cells and makes you more susceptible to disease.

The question I want to put in front of you is: who's in charge?

Who ordered all these switches to be set in the right way in the first place, and who decides when to start to let loose the Cortisol?

You could argue that the genes are in charge, because the differentiation of the body into different cell types, each with different genes switched on, was at root a genetic process. But that's misleading, because genes are not the cause of stress. The death of a loved one, or an impending exam do not speak directly to the genes. They are information processed by the brain.

S T R E S S 1 5 1

So the brain is in charge. The hypothalamus of the brain sends out the signal that tells the pituitary gland to release a hormone that tells the adrenal gland to make and secrete Cortisol. The hypothalamus takes its orders from the conscious part of the brain which gets its information from the outside world.

But that's not much of an answer either, because the brain is part of the body. The reason the hypothalamus stimulates the pituitary which stimulates the adrenal cortex is not because the brain decided or learnt that this was a good way to do things. It did not set up the system in such a way that thinking about an impending exam would make you less resistant to catching a cold. Natural selection did that (for reasons I will come back to shortly). And in any case, it is a wholly involuntary and unconscious reaction, which implies that it is the exam, rather than the brain, that is in charge of events.

And if the exam is in charge, then society is to blame, but what is society but a collection of individuals, which brings us back to bodies? Besides, people vary in their susceptibility to stress. Some find impending exams terrifying, others take them in their stride.

What is the difference? Somewhere down the cascade of events that is the production, control and reaction to Cortisol, stress-prone people must have subtly different genes from phlegmatic folk. But who or what controls these genetic differences?

The truth is that nobody is in charge. It is the hardest thing for human beings to get used to, but the world is full of intricate, cleverly designed and interconnected systems that do not have control centres. The economy is such a system. The illusion that economies run better if somebody is put in charge of them - and decides what gets manufactured where and by whom - has done devastating harm to the wealth and health of peoples all over the world, not just in the former Soviet Union, but in the west as well.

From the Roman Empire to the European Union's high-definition television initiative, centralised decisions about what to invest in have been disastrously worse than the decentralised chaos of the market. Economies are not centralised systems; they are markets with decentralised, diffuse controls.

1 5 2 G E N O M E

It is the same with the body. You are not a brain running a body by switching on hormones. Nor are you a body running a genome by switching on hormone receptors. Nor are you a genome running a brain by switching on genes that switch on hormones. You are all of these at once.

Many of the oldest arguments in psychology boil down to misconceptions of this kind. The arguments for and against 'genetic determinism' presuppose that the involvement of the genome places it above and beyond the body. But as we have seen it is the body that switches on genes when it needs them, often in response to a more or less cerebral, or even conscious, reaction to external events.

You can raise your Cortisol levels just by thinking about stressful eventualities - even fictional ones. Likewise, the dispute between those who believe that a certain suffering is purely psychiatric and those who insist it has a physical cause - consider M E , or chronic fatigue syndrome - is missing the point entirely. The brain and the body are part of the same system. If the brain, responding to psychological stress, stimulates the release of Cortisol and Cortisol suppresses the reactivity of the immune system, then a dormant viral infection may well flare up, or a new one catch hold. The symptoms may indeed be physical and the causes psychological. If a disease affects the brain and alters the mood, the causes may be physical and the symptoms psychological.

This topic is known as psychoneuroimmunology, and it is slowly inching its way into fashion, mostly resisted by doctors and mostly hyped by faith healers of one kind or another. But the evidence is real enough. Chronically unhappy nurses have more episodes of cold sores than others who also carry the virus. People with anxious personalities have more outbreaks of genital herpes than sunny optimists. At West Point military academy, the students most likely to catch mononucleosis (glandular fever), and the ones most likely to get a severe illness from it if they do, are the ones who are most anxious and pressured by their work. Those who care for Alzheimer's patients (an especially stressful activity) have fewer disease-fighting T lymphocytes in their blood than expected. Those S T R E S S 153

who lived near Three Mile Island nuclear plant at the time of its accident had more cancers than expected three years later, not because they were exposed to radiation (they weren't), but because their Cortisol levels had risen, reducing the responsiveness of their immune system to cancer cells. Those bereaved by the death of a spouse have a less responsive immune system for several weeks afterwards. Children whose families have been riven by a parental argument in the previous week are more likely to catch viral infections. People with most psychological stress in their past get more colds than people who have led happy lives. And if you find these sorts of studies hard to believe, then most of them have been replicated in some form or another using mice or rats.1

Poor old Rene Descartes usually gets the blame for the dualism that has dominated western thinking and made us all so resistant to the idea that the mind can affect the body and the body can affect the mind, too. He barely deserves the blame for an error we all commit. In any case, the fault is not so much dualism — the notion of a separate mind detached from the material matter of the brain.

There is a far greater fallacy that we all commit, so easily that we never even notice it. We instinctively assume that bodily biochemistry is cause whereas behaviour is effect, an assumption we have taken to a ridiculous extent in considering the impact of genes upon our lives. If genes are involved in behaviour then it is they that are the cause and they that are deemed immutable. This is a mistake made not just by genetic determinists, but by their vociferous opponents, the people who say behaviour is 'not in the genes'; the people who deplore the fatalism and predestination implied, they say, by behaviour genetics. They give too much ground to their opponents by allowing this assumption to stand, for they tacitly admit that if genes are involved at all, then they are at the top of the hierarchy.

They forget that genes need to be switched on, and external events

- or free-willed behaviour — can switch on genes. Far from us lying at the mercy of our omnipotent genes, it is often our genes that lie at the mercy of us. If you go bungee jumping or take a stressful job, or repeatedly imagine a terrible fear, you will raise your Cortisol 1 5 4 G E N O M E

levels, and the Cortisol will dash about the body busy switching on genes. (It is also an indisputable fact that you can trigger activity in the 'happiness centres' of the brain with a deliberate smile, as surely as you trigger a smile with happy thoughts. It really does make you feel better to smile. The physical can be at the beck and call of the behavioural.)

Some of the best insights into the way behaviour alters gene expression come from studies of monkeys. Fortunately for those who believe in evolution, natural selection is an almost ridiculously thrifty designer and once she has hit upon a system of genes and hormones to indicate and respond to stress, she is loath to change it (we are ninety-eight per cent chimpanzees and ninety-four per cent baboons, remember). So the very same hormones work in the very same way in monkeys and switch on the very same genes.

There is a troop of baboons in east Africa whose bloodstream Cortisol levels have been closely studied. When a certain young male baboon attached himself to a new troop, as male baboons of a certain age are wont to do, he became highly aggressive as he fought to establish himself in the hierarchy of his chosen society. The result was a steep increase in the Cortisol concentration in his blood as well as that of his unwilling hosts. As his Cortisol (and testosterone) levels rose, so his lymphocyte count fell. His immune system bore the brunt of his behaviour. At the same time his blood began to contain less and less of the cholesterol bound to high-density lipoprotein (HDL). Such a fall is a classic precursor of furring up of the coronary arteries. Not only was the baboon, by his free-willed behaviour, altering his hormones, and hence the expression of his genes, he was thereby increasing his risk of both infection and coronary artery disease.2

Among monkeys kept in zoos, the ones whose arteries fur up are the ones at the foot of the pecking order. Bullied by their more senior colleagues, they are continuously stressed, their blood is rich in Cortisol, their brains are low in serotonin, their immune systems are permanently depressed and scar tissue builds up on the walls of their coronary arteries. Quite why is still a mystery. Many scientists now S T R E S S 155

believe that coronary disease is at least partly caused by infectious agents, such as chlamydia bacteria and herpes viruses. The effect of stress is to lower immune surveillance of these dormant infections which allows them to flourish. Perhaps, in this sense, heart disease in monkeys is infectious, though stress may play a role as well.

People are very like monkeys. The discovery that monkeys low in the hierarchy get heart disease came soon after the far more startling discovery that British civil servants working in Whitehall also get heart disease in proportion to their lowliness in the bureaucratic pecking order. In a massive, long-term study of 17,000 civil servants, an almost unbelievable conclusion emerged: the status of a person's job was more able to predict their likelihood of a heart attack than obesity, smoking or high blood pressure. Somebody in a low-grade job, such as a janitor, was nearly four times as likely to have a heart attack as a permanent secretary at the top of the heap.

Indeed, even if the permanent secretary was fat, hypertensive or a smoker, he was still less likely to suffer a heart attack at a given age than a thin, non-smoking, low-blood-pressure janitor. Exactly the same result emerged from a similar study of a million employees of the Bell Telephone Company in the 1960s.3

Think about this conclusion for a moment. It undermines almost everything you have ever been told about heart disease. It relegates cholesterol to the margins of the story (high cholesterol is a risk factor, but only in those with genetic predispositions to high cholesterol, and even in these people the beneficial effects of eating less fat are small). It relegates diet, smoking and blood pressure - all the physiological causes so preferred by the medical profession — to secondary causes. It relegates to a footnote the old and largely discredited notion that stress and heart failure come with busy, senior jobs or fast-living personalities: again there is a grain of truth in this fact, but not much. Instead, dwarfing these effects, science now elevates something non-physiological, something strictly related to the outside world: the status of your job. Your heart is at the mercy of your pay grade. What on earth is going on?

The monkeys hold the clue. The lower they are in the pecking 156 G E N O M E

order, the less control they have over their lives. Likewise in the civil service, Cortisol levels rise in response not to the amount of work you do, but to the degree to which you are ordered about by other people. Indeed, you can demonstrate this effect experimentally, just by giving two groups of people the same task to do, but ordering one group to do the task in a set manner and to an imposed schedule.

This externally controlled group of people suffers a greater increase in stress hormones and rise in blood pressure and heart rate than the other group.

Twenty years after the Whitehall study began, it was repeated in a department of the civil service that then began to experience privatisation. At the beginning of the study, the civil servants had no notion of what it meant to lose their jobs. Indeed, when a questionnaire was being piloted for the study, the subjects objected to a question that asked if they feared losing their jobs. It was a meaningless question in the civil service, they explained: at worst they might be transferred to a different department. By 1995 they knew exactly what losing their jobs meant; more than one in three had already experienced it. The effect of privatisation was to give everybody a feeling that their lives were at the mercy of external factors. Not surprisingly, stress followed and with stress came ill health - far more ill health than could be explained by any changes in diet, smoking or drinking.

The fact that heart disease is a symptom of lack of control explains a good deal about its sporadic appearance. It explains why so many people in senior jobs have heart attacks soon after they retire and

'take it easy'. From running offices they often move to lowly and menial jobs (washing dishes, walking the dog) in domestic environments run by their spouses. It explains why people are capable of postponing an illness, even a heart attack, until after a family wedding or a major celebration — until the end of a period of busy work when they are in control of events. (Students also tend to go down with illnesses after periods of acute exam pressure, not during them.) It explains why unemployment and welfare dependency are so good at making people ill. No alpha-male monkey was ever such an S T R E S S 157

intransigent and implacable controller of subordinates' lives as the social services of the state are of people dependent on welfare. It may even explain why modern buildings in which the windows cannot be opened make people sicker than older buildings in which people have more control over their environment.

I am going to repeat myself for emphasis. Far from behaviour being at the mercy of our biology, our biology is often at the mercy of our behaviour.

What is true of Cortisol is also true of other steroid hormones.

Testosterone levels correlate with aggression, but is that because the hormone causes aggression, or because release of the hormone is caused by aggression? In our materialism, we find the first alternative far easier to believe. But in fact, as studies of baboons demonstrate, the second is closer to the truth. The psychological precedes the physical. The mind drives the body, which drives the genome.4

Testosterone is just as good at suppressing the immune system as Cortisol. This explains why, in many species, males catch more diseases and have higher mortality than females. This immune suppression applies not just to the body's resistance to micro-organisms, but to large parasites, too. The warble fly lays its eggs on the skin of deer and cattle; the maggot then burrows into the flesh of the animal before returning to the skin to form a nodule in which to metamorphose into a fly. Reindeer in northern Norway are especially troubled by these parasites, but males noticeably more than females.

On average, by the age of two, a male reindeer has three times as many warble-fly nodules in its skin as a female reindeer, yet castrated males have the same number as females. A similar pattern can be found for many infectious parasites, including, for instance, the protozoan that causes Chagas' disease, the affliction widely believed to explain Charles Darwin's chronic illnesses. Darwin was bitten by the bug that carries Chagas' disease while travelling in Chile and some of his later symptoms fit the disease. If Darwin had been a woman, he might have spent less time feeling sorry for himself.

Yet it is to Darwin that we must turn for enlightenment here.

The fact that testosterone suppresses immune function has been 1 5 8 G E N O M E

seized upon by a cousin of natural selection known as sexual selection and ingeniously exploited. In Darwin's second book on evolution, The descent of man, he put forward the notion that, just as a pigeon breeder can breed pigeons, so a female can breed males. By consistently choosing which males to mate with over many generations, female animals can alter the shape, colour, size or song of males of their species. Indeed, as I described in the chapter on chromosomes X and Y, Darwin suggested that this is exactly what has happened in the case of peacocks. It was not until a century later, in the 1970s and 1980s, that a series of theoretical and experimental studies demonstrated that Darwin was right, and that the tails, plumes, antlers, songs and size of male animals are bred into them by consistent trends of passive or active female choice, generation after generation.

But why? What conceivable benefit can a female derive from picking a male with a long tail or a loud song? Two favourite ideas have dominated the debate, the first being that the female must follow the prevailing fashion lest she have sons that are not themselves attractive to females who follow the prevailing fashion. The second idea, and the one that I propose to consider here, is that the quality of the male's ornament reflects the quality of his genes in some way. In particular, it reflects the quality of his resistance to prevailing infections. He is saying to all who would listen: see how strong I am; I can grow a great tail or sing a great song, because I am not debilitated by malaria, nor infected with worms. And the fact that testosterone suppresses the immune system is actually the greatest possible help in making this an honest message. For the quality of his ornaments depends on the level of testosterone in his blood: the more testosterone he has, the more colourful, large, songful or aggressive he will be. If he can grow a great tail despite lowering his immune defences, yet not catch disease, he must be impressive genetically. It is almost as if the immune system obscures the genes; testosterone parts the veil and allows the female to see directly into the genes.6

This theory is known as the immunocompetence handicap and S T R E S S 159

it depends upon the immune-suppressive effects of testosterone being unavoidable. A male cannot get round the handicap by raising his testosterone levels and not suppressing his immune system. If such a male existed, he would surely be a great success and would leave many offspring behind, because he could grow a long tail with (literally) immunity. Hence, the theory implies that the link between steroids and immune suppression is as fixed, inevitable and important as any in biology.

But this is even more puzzling. Nobody has a good explanation for the link in the first place, let alone its inevitability. Why should bodies be designed so that their immune systems are depressed by steroid hormones? It means that whenever you are stressed by a life event, you become more vulnerable to infection, cancer and heart disease. That is kicking you when you are down. It means that whenever an animal raises its testosterone level to fight its rivals for mates or to enhance its display, it becomes more vulnerable to infection, cancer and heart disease. Why?

Various scientists have struggled with this conundrum, but to little effect. Paul Martin, in his book on psychoneuroimmunology called The sickening mind, discusses two possible explanations and rejects them both. First is the notion that it is all a mistake, and that the links between the immune system and the stress response are accidental by-products of the way some other systems have to be designed. As Martin points out, this is a deeply unsatisfactory explanation for a system full of complex neural and chemical links.

Very, very few parts of the body are accidental, vestigial or func¬

tionless, especially not complex parts. Natural selection would ruthlessly cull links that suppress the immune response if they had no function.

The second explanation, that modern life produces prolonged and unnatural stresses and that in an ancient environment such stresses would have been much shorter-lived, is equally disappointing. Baboons and peacocks live in a state of nature, yet they too - and virtually every other bird and mammal on the planet -

suffer from immune suppression by steroids.

l6o G E N O M E

Martin admits to bafflement. He cannot explain the fact that stress inevitably depresses the immune system. Nor can I. Perhaps, as Michael Davies has suggested, the depression is designed to save energy in times of semi-starvation, a common form of stress before the modern era. Or perhaps the response to Cortisol is a side-effect of the response to testosterone (they are very similar chemicals) and the response to testosterone is deliberately engineered into males by the genes of females the better to sort the fitter — that is more disease resistant - males from the less fit. In other words, the link may be the product of a kind of sexual antagonism like the one discussed in the chapter on chromosomes X and Y. I don't find this explanation convincing, so I challenge you to find a better one.

C H R O M O S O M E 1 1

P e r s o n a l i t y

A man's character is his fate.

Heraditus

The tension between universal characteristics of the human race and particular features of individuals is what the genome is all about.

Somehow the genome is responsible for both the things we share with other people and the things we experience uniquely in ourselves.

We all experience stress; we all experience the elevated Cortisol that goes with it; we all suffer from the immune-suppressive effects thereof. We all have genes switched on and off by external events in this way. But each of us is unique, too. Some people are phlegmatic, some highly strung. Some are anxious, others risk-seeking. Some are confident, others shy. Some are quiet, others loquacious. We call these differences personality, a word that means more than just character.

It means the innate and individual element in character.

To seek out the genes that influence personality, it is time to move from the hormones of the body to the chemicals of the mind

- though the distinction is by no means a hard-and-fast one. On the short arm of chromosome 11, there lies a gene called D4DR.

1 6 2 G E N O M E

It is the recipe for a protein called a dopamine receptor, and it is switched on in cells of certain parts of the brain but not in others.

Its job is to stick out of the membrane of a neuron at the junction with another neuron (known as a synapse), ready to latch on to a small chemical called dopamine. Dopamine is a neurotransmitter, released from the tips of other neurons by an electrical signal. When the dopamine receptor encounters dopamine, it causes its own neuron to discharge an electrical signal of its own. That is the way the brain works: electrical signals that cause chemical signals that cause electrical signals. By using at least fifty different chemical signals, the brain can carry on many different conversations at once: each neurotransmitter stimulates a different set of cells or alters their sensitivity to different chemical messengers. It is misleading to think of a brain as a computer for many reasons, but one of the most obvious is that an electrical switch in a computer is just an electrical switch. A synapse in a brain is an electrical switch embedded in a chemical reactor of great sensitivity.

The presence of an active D4DR gene in a neuron immediately identifies that neuron as a member of one of the brain's dopamine¬

mediated pathways. Dopamine pathways do many things, including controlling the flow of blood through the brain. A shortage of dopamine in the brain causes an indecisive and frozen personality, unable to initiate even the body's own movement. In the extreme form, this is known as Parkinson's disease. Mice with the genes for making dopamine knocked out will starve to death from sheer immobility. If a chemical that closely resembles dopamine (a dopamine agonist, in the jargon) is injected into their brains, they recover their natural arousal. An excess of dopamine in the brain, by contrast, makes a mouse highly exploratory and adventurous. In human beings, excessive dopamine may be the immediate cause of schizophrenia; and some hallucinogenic drugs work by stimulating the dopamine system. A mouse addicted to cocaine so badly that it prefers the drug to food is experiencing the release of dopamine in a part of the brain known as the nucleus acumbens. A rat in which this 'pleasure centre' is stimulated whenever it presses a lever will P E R S O N A L I T Y 163

learn to return to press the lever again and again. But if a dopamine¬

blocking chemical is added to the rat's brain, the rat quickly loses interest in the lever.

In other words, to simplify grossly, dopamine is perhaps the brain's motivation chemical. Too little and the person lacks initiative and motivation. Too much and the person is easily bored and frequently seeks new adventures. Here perhaps lies the root of a difference in personality. As Dean Hamer put it, when he set out to seek the gene for thrill-seeking personalities in the mid-1990s, he was looking for the difference between Lawrence of Arabia and Queen Victoria. Since it takes many different genes to make, control, emit and receive dopamine, let alone to build the brain in the first place, nobody, least of all Hamer, expected to find a single gene controlling exclusively this aspect of personality. Nor did he expect to find that all variation in adventure-seeking is genetic, merely that there would be genetic influences at work among others.

The first genetic difference turned up in Richard Ebstein's laboratory in Jerusalem in the D4DR gene on chromosome 11. D4DR

has a variable repeat sequence in the middle, a minisatellite phrase forty-eight letters in length repeated between two and eleven times.

Most of us have four or seven copies of the sequence, but some people have two, three, five, six, eight, nine, ten or eleven. The larger the number of repeats, the more ineffective is the dopamine receptor at capturing dopamine. A 'long' D4DR gene implies a low responsiveness to dopamine in certain parts of the brain, whereas a 'short' D4DR gene implies a high responsiveness.

Hamer and his colleagues wanted to know if people with the long gene had different personalities from people with the short gene.

This is in effect the opposite procedure from that followed by Robert Plomin on chromosome 6, where he sought to correlate an unknown gene with a known behavioural difference (in I Q ) . Hamer went from the gene to the trait rather than vice versa. He measured the novelty-seeking character of 124 people on a series of set personality tests and then examined their genes.

Bingo. Of the subjects Hamer tested - admittedly not a huge sample 1 6 4 G E N O M E

- people with either one or two long copies of the gene (remember there are two copies of each chromosome in each cell of the adult body, one from each parent) were distinctly more novelty-seeking than people with two short copies of the gene. 'Long' genes were defined as those with six or more repeats of the minisatellite sequence. At first Hamer was worried that he might be looking at what he calls a

'chopstick' gene. The gene for blue eyes is common in people who are bad at using chopsticks, but nobody would dream of suggesting that chopstick skill is genetically determined by the gene for eye colour. It just happens that both blue eyes and chopstick incompetence correlate with non-oriental origin for a blindingly obvious non-genetic reason called culture. Richard Lewontin uses another analogy for this fallacy: the fact that people who are good at knitting tend not to have Y

chromosomes (i.e., they tend to be women) does not imply that knitting is caused by a lack of Y chromosomes.

So, to rule out a spurious correlation of this kind, Hamer repeated the study in the United States with members of one family. Again he found a clear correlation: the novelty-seekers were much more likely to have one or more copy of the long gene. This time the chopstick argument looks increasingly untenable, because any differences within a family are less likely to be cultural ones. The genetic difference may indeed contribute to the personality difference.

The argument goes like this. People with 'long' D4DR genes have low responsiveness to dopamine, so they need to take a more adventurous approach to life to get the same dopamine 'buzz' that short-gened people get from simple things. In search of these buzzes they develop novelty-seeking personalities. Hamer went on to demonstrate a striking example of what it means to be a novelty seeker.

Among heterosexual men, those with the long D4DR genes are six times more likely to have slept with another man than those with the short genes. Among homosexual men, those with the long genes are five times more likely to have slept with a woman than those with the short genes. In both groups, the long-gened people had more sexual partners than the short-gened people.1

We all know people who will try anything, and conversely people P E R S O N A L I T Y 1 6 5

who are set in their ways and reluctant to experiment with something new. Perhaps the first lot have long D4DR genes and the second lot have short ones. It is not quite that simple. Hamer claims to explain no more than four per cent of novelty seeking by reference to this one gene. He estimates that novelty seeking is about forty per cent heritable, and that there are about ten equally important genes whose variation matches the variation in personality. That is just one element in personality, but there are many others, perhaps a dozen. Making the wild assumption that they all involve similar numbers of genes leads to the conclusion that there may be 500

genes that vary in tune with human personalities. These are just the ones that vary. There may be many others that do not normally vary, but if they did would affect personality.

This is the reality of genes for behaviour. Do you see now how unthreatening it is to talk of genetic influences over behaviour? How ridiculous to get carried away by one 'personality gene' among 500?

How absurd to think that, even in a future brave new world, somebody might abort a foetus because one of its personality genes is not up to scratch — and take the risk that on the next conception she would produce a foetus in which two or three other genes were of a kind she does not desire? Do you see now how futile it would be to practise eugenic selection for certain genetic personalities, even if somebody had the power to do so? You would have to check each of 500 genes one by one, deciding in each case to reject those with the 'wrong' gene. At the end you would be left with nobody, not even if you started with a million candidates. We are all of us mutants. The best defence against designer babies is to find more genes and swamp people in too much knowledge.

Meanwhile, the discovery that personality has a strong genetic component can be used in some very non-genetic therapy. When naturally shy baby monkeys are fostered to confident monkey mothers, they quickly outgrow their shyness. It is almost certainly the same with people - the right kind of parenting can alter an innate personality. Curiously, understanding that it is innate seems to help to cure it. One trio of therapists, reading about the new 166 G E N O M E

results emerging from genetics, switched from trying to treat their clients' shyness to trying to make them content with whatever their innate predispositions were. They found that it worked. The clients felt relieved to be told that their personality was a real, innate part of them and not just a bad habit they had got into. 'Paradoxically, depathologising people's fundamental inclinations and giving group members permission to be the way they are seemed to constitute the best insurance that their self-esteem and interpersonal effectiveness would improve.' In other words, telling them they were naturally shy helped them overcome that shyness. Marriage counsellors, too, report good results from encouraging their clients to accept that they cannot change their partners' irritating habits - because they are probably innate - but must find ways to live with them. The parents of a homosexual are generally more accepting when they believe that homosexuality is an immutable part of nature rather than a result of some aspect of their parenting. Far from being a sentence, the realisation of innate personality is often a release.

Suppose you wished to breed a strain of fox or rat that was more tame and less instinctively timid than the average. One way to do so would be to pick the darkest pups in the litter as the stock for breeding the next generation. In a few years you would have tamer, and darker, animals. This curious fact has been known to animal breeders for many years. But in the 1980s it took on a new significance.

It parallels another link between neurochemistry and personality in people. Jerome Kagan, a Harvard psychologist, leading a team of researchers studying shyness or confidence in children, found that he could identify unusually 'inhibited' types as early as four months of age — and fourteen years later could predict how shy or confident those same human beings would be as adults. Upbringing mattered a good deal. But intrinsic personality played just as big a role.

Big deal. Nobody, except perhaps the most die-hard social deter¬

minist, would find an innate component of shyness surprising. But it turned out that the same personality traits correlated with some unexpected other features. Shy adolescents were more likely to be blue-eyed (all the subjects were of European descent), susceptible P E R S O N A L I T Y 1 6 7

to allergies, tall and thin, narrow-faced, to have more heat-generating activity under the right forehead and a faster heartbeat, than the less shy individuals. All of these features are under the control of a particular set of cells in the embryo called the neural crest, from which a particular part of the brain, the amygdala, derives. They also all use the same neurotransmitter, called norepinephrine, a substance very like dopamine. All these features are also characteristic of northern Europeans, Nordic types for the most part. Kagan's argument goes that the Ice Age selected those better able to withstand cold in these parts: people with high metabolic rates. But a high metabolic rate is produced by an active norepinephrine system in the amygdala, and brings with it lots of different baggage - a phlegmatic and shy personality being one aspect and a pale appearance being another. Just as in foxes and rats, shy and suspicious types are paler than bold types.3

If Kagan is right, tall, thin adults with blue eyes are slightly more likely to become anxious when challenged than other people. An up-to-date recruitment consultant might find this handy in his head-hunting. After all, employers already seek to discriminate between personalities. Most job advertisements require candidates with 'good interpersonal skills' - something that is probably partly innate. Yet it would plainly be a repellent world in which we were picked for jobs on the basis of our eye colour. Why? Physical discrimination is so much less acceptable than psychological. Yet psychological discrimination is just chemical discrimination. It is just as material as any other discrimination.

Dopamine and norepinephrine are so-called monoamines. Their close cousin, another monoamine found in the brain, is serotonin, which is also a chemical manifestation of personality. But serotonin is more complicated than dopamine and norepinephrine. It is remarkably hard to pin down its characteristics. If you have unusually high levels of serotonin in your brain you will probably be a compulsive person, given to tidiness and caution, even to the point of being neurotic about it. People with the pathological condition known as obsessive—compulsive disorder can usually alleviate their symptoms l 6 8 G E N O M E

by lowering their serotonin levels. At the other end of the spectrum, people with unusually low serotonin levels in their brains tend to be impulsive. Those who commit impulsive violent crimes, or suicide, are often those with less serotonin.

Prozac works by affecting the serotonin system, though there is still controversy about exactly how it does so. The conventional theory put forward by scientists at Eli Lilly, where the drug was invented, is that Prozac inhibits the reabsorption of serotonin into neurons, and thus increases the amount of serotonin in the brain.

Increased serotonin alleviates anxiety and depression and can turn even fairly ordinary people into optimists. But it remains possible that Prozac has exactly the opposite effect: that it interferes with the responses of neurons to serotonin. There is a gene on chromosome 17, called the serotonin-transporter gene, which varies, not in itself, but in the length of an 'activation sequence' just upstream of the gene — a sort of dimmer switch at the beginning of the gene, in other words, designed to slow down the expression of the gene itself. As with so many mutations, the variation in length is caused by a variable number of repetitions of the same sequence, a twenty-two-letter phrase that is repeated either fourteen or sixteen times.

About one in three of us have two copies of the long sequence, which is marginally worse at switching off its gene. As a result such people have more serotonin transporter, which means that more serotonin gets carried about. These people are much less likely to be neurotic, and slightly more likely to be agreeable than the average person, whatever their sex, race, education or income.

From this, Dean Hamer concludes that serotonin is the chemical that abets, rather than alleviates, anxiety and depression. He calls it the brain's punishment chemical. Yet all sorts of evidence points in the other direction: that you feel better with more serotonin, not less. There is, for instance, a curious link between winter, a desire for snacks, and sleepiness. In some people — probably once more a genetic minority, though no gene version has yet been found that correlates with susceptibility to this condition — the dark evenings of winter lead to a craving for carbohydrate snacks in the late P E R S O N A L I T Y 1 6 9

afternoon. Such people often need more sleep in winter, though they find their sleep less refreshing. The explanation seems to be that the brain starts making melatonin, the hormone that induces sleep, in response to the early evening darkness of winter days.

Melatonin is made from serotonin, so serotonin levels drop as it gets used up in melatonin manufacture. The quickest way to raise serotonin levels again is to send more tryptophan into the brain, because serotonin is made from tryptophan. The quickest way to send more tryptophan into the brain is to secrete insulin from the pancreas, because insulin causes the body to absorb other chemicals similar to tryptophan, thus removing competitors for the channels that take tryptophan into the brain. And the quickest way to secrete insulin is to eat a carbohydrate snack.4

Are you still with me? You eat cookies on winter evenings to cheer yourself up by raising your brain serotonin. The take-home message is that you can alter your serotonin levels by altering your eating habits. Indeed, even drugs and diets designed to lower blood cholesterol can influence serotonin. It is a curious fact that nearly all studies of cholesterol-lowering drugs and diets in ordinary people show an increase in violent death compared with control samples that usually matches the decrease in deaths from heart disease. In all studies put together, cholesterol treatment cut heart attacks by fourteen per cent, but raised violent deaths by an even more significant seventy-eight per cent. Because violent deaths are rarer than heart attacks, the numerical effect roughly cancels out, but violent deaths can sometimes involve innocent bystanders. So treating high cholesterol levels has its dangers. It has been known for twenty years that impulsive, antisocial and depressed people - including prisoners, violent offenders and failed suicides - have generally lower cholesterol levels than the population at large. No wonder Julius Caesar distrusted Cassius's lean and hungry look.

These disturbing facts are usually played down by the medical profession as statistical artefacts, but they are too repeatable for that. In the so-called MrFit trial, in which 351,000 people from seven countries were followed for seven years, people with very low 1 7 0 G E N O M E

cholesterol and people with very high cholesterol proved twice as likely to die at a given age as people with medium cholesterol.

The extra deaths among low-cholesterol people are mainly due to accident, suicide or murder. The twenty-five per cent of men with the lowest cholesterol count are four times as likely to commit suicide as the twenty-five per cent of men with the highest count

- though no such pattern holds with women. This does not mean we should all go back to eating fried eggs. Having low cholesterol, or lowering your cholesterol too far, is highly dangerous for a small minority, just as having high cholesterol and eating high-cholesterol diets is dangerous for a small minority. Low-cholesterol dieting advice should be confined to those who are genetically endowed with too much cholesterol, and not given to everybody.

The link between low cholesterol and violence almost certainly involves serotonin. Monkeys fed on low-cholesterol diets become more aggressive and bad-tempered (even if they are not losing weight), and the cause seems to be a drop in serotonin levels. In Jay Kaplan's laboratory at Bowman Gray Medical School in North Carolina, eight monkeys fed on a low-cholesterol (but high-fat) diet soon had brain serotonin levels that were roughly half as high as those in the brains of nine monkeys fed on a high-cholesterol diet.

They were also forty per cent more likely to take aggressive or antisocial action against a fellow monkey. This was true of both sexes. Indeed, low serotonin is an accurate predictor of aggressive-ness in monkeys, just as it is an accurate predictor of impulsive murder, suicide, fighting or arson in human beings. Does this mean that if every man was forced by law to have his serotonin level displayed on his forehead at all times, we could tell who should be avoided, incarcerated or protected from themselves?5

Fortunately, such a policy is as likely to fail as it is offensive to civil liberties. Serotonin levels are not innate and inflexible. They are themselves the product of social status. The higher your self-esteem and social rank relative to those around you, the higher your serotonin level is. Experiments with monkeys reveals that it is the social behaviour that comes first. Serotonin is richly present in P E R S O N A L I T Y 1 7 1

dominant monkeys and much more dilute in the brains of subordinates. Cause or effect? Almost everybody assumed the chemical was at least partly the cause: it just stands to reason that the dominant behaviour results from the chemical, not vice versa. It turns out to be the reverse: serotonin levels respond to the monkey's perception of its own position in the hierarchy, not vice versa.6

Contrary to what most people think, high rank means low aggressiveness, even in vervet monkeys. The high-ranking individuals are not especially large, fierce or violent. They are good at things like reconciliation and recruiting allies. They are notable for their calm demeanour. They are less impulsive, less likely to misinterpret play-fighting as aggression. Monkeys are not people, of course, but as Michael McGuire of the University of California, Los Angeles, has discovered, any group of people, even children, can immediately spot which of the monkeys in his captive group is the dominant one. Its demeanour and behaviour — what Shelley called the 'sneer of cold command' — are instantly familiar in an anthropomorphic way. There is little doubt that the monkey's mood is set by its high serotonin levels. If you artificially reverse the pecking order so that the monkey is now a subordinate, not only does its serotonin drop, but its behaviour changes, too. Moreover, much the same seems to happen in human beings. In university fraternities, the leading figures are blessed with rich serotonin concentrations which fall if they are deposed. Telling people they have low or high serotonin levels could become a self-fulfilling prophecy.

This is an intriguing reversal of the cartoon picture of biology most people have. The whole serotonin system is about biological determinism. Your chances of becoming a criminal are affected by your brain chemistry. But that does not mean, as it is usually assumed to mean, that your behaviour is socially immutable. Quite the reverse: your brain chemistry is determined by the social signals to which you are exposed. Biology determines behaviour yet is determined by society. I described the same phenomenon in the Cortisol system of the body; here it is again with the serotonin system of the brain. Mood, mind, personality and behaviour are indeed 1 7 2 G E N O M E

socially determined, but that does not mean they are not also biologically determined. Social influences upon behaviour work through the switching on and off of genes.

None the less, it is clear that there are all sorts of innate personality types, and that people vary in the way they respond to social stimuli mediated through neurotransmitters. There are genes that vary the rate of serotonin manufacture, genes that vary the responsiveness of serotonin receptors, genes that make some brain areas respond to serotonin more than others, genes that make some people depressed in winter because of too responsive a melatonin system using up serotonin. And so on and on and on. There is a Dutch family in which the men have been criminals for three generations, and the cause is undoubtedly a gene. The criminal men have an unusual version of a gene on the X chromosome called the monoamine oxidase A gene. Monoamine oxidase is responsible for breaking down serotonin among other chemicals. It is highly probable that their unusual serotonin neurochemistry makes these Dutch men more likely to fall into lives of crime. But this does not make this gene a 'crime gene', except in a very pedestrian sense. For a start, the mutation in question is now considered an 'orphan' mutation, so rare that very few criminals have this version of the gene. The monoamine oxidase gene can explain very little about general criminal behaviour.

But it underscores yet again the fact that what we call personality is to a considerable degree a question of brain chemistry. There are a score of different ways in which this one chemical, serotonin, can be related to innate differences in personality. These are overlaid on the score of different ways that the mind's serotonin system responds to outside influences such as social signals. Some people are more sensitive to some outside signals than others. This is the reality of genes and environments: a maze of complicated interactions between them, not a one-directional determinism. Social behaviour is not some external series of events that takes our minds and bodies by surprise. It is an intimate part of our make-up, and our genes are programmed not only to produce social behaviour, but to respond to it as well.

C H R O M O S O M E 1 2

S e l f - A s s e m b l y

The egg's ordain'd by nature to that end

And is a chicken in potentia

Ben Jonson, The Alchemist

There are human analogies for almost everything in nature. Bats use sonar; the heart is a pump; the eye is a camera; natural selection is trial and error; genes are recipes; the brain is made from wires (known as axons) and switches (synapses); the hormonal system uses feedback control like an oil refinery; the immune system is a counter-espionage agency; bodily growth is like economic growth.

And so, infinitely, on. Although some of these analogies can mislead, we are at least familiar with the kinds of techniques and technologies that Mother Nature employs to solve her various problems and achieve her ingenious designs. We have reinvented most of them ourselves in technological life.

But now we must leave such comfortable terrain behind and step into the unknown. One of the most remarkable, beautiful and bizarre things that Mother Nature achieves without apparent difficulty is something for which we have no human analogy at all: the 1 7 4 G E N O M E

development of a human body from an undifferentiated blob called a fertilised egg. Imagine trying to design a piece of hardware (or software, for that matter) that could do something analogous to this feat. The Pentagon probably tried it, for all I know: 'Good Morning, Mandrake. Your job is to make a bomb that grows itself from a large blob of raw steel and a heap of explosive. You have an unlimited budget and one thousand of the best brains at your disposal in the New Mexico desert. I want to see a prototype by August. Rabbits can do it ten times a month. So it cannot be that hard. Any questions?'

Without the handrail of analogy, it is difficult even to understand Mother Nature's feat. Something, somewhere must be imposing a pattern of increasing detail upon the egg as it grows and develops.

There must be a plan. But unless we are to invoke divine intervention, that imposer of detail must be within the egg itself. And how can the egg make a pattern without starting with one? Little wonder that, in past centuries, there was a natural preference for theories of prefor-mation, so that some people thought they saw within the human sperm a miniature homunculus of a man. Preformation, as even Aristotle spotted, merely postpones the problem, for how did the homunculus get its shape? Later theories were not much better, though our old friend William Bateson came surprisingly close to the right answer when he conjectured that all organisms are made from an orderly series of parts or segments, and coined the term homeosis for it. And there was a vogue in the 1970s for explaining embryology by reference to increasingly sophisticated mathematical geometries, standing waves and other such arcana. Alas for mathematicians, nature's answer turns out, as ever, to be both simpler and much more easily understood, though the details are ferociously intricate. It all revolves around genes, which do indeed contain the plan in digital form. One large cluster of these developmental genes lies close to the middle of chromosome 12. The discovery of these genes and the elucidation of how they work is probably the greatest intellectual prize that modern genetics has won since the code itself was cracked. It was a discovery with two stunning and lucky surprises at its heart.1

As the fertilised egg grows into an embryo, at first it is an undiffer-S E L F - A S S E M B L Y 1 7 5

entiated blob. Then gradually it develops two asymmetries - a head-tail axis and a front-back axis. In fruit flies and toads, these axes are established by the mother, whose cells instruct one end of the embryo to become the head and one part to become the back. But in mice and people the asymmetries develop later and nobody knows quite how. The moment of implantation into the womb seems to be critical.

In fruit flies and toads, these asymmetries are well understood: they consist of gradients in the chemical products of different maternal genes. In mammals, too, the asymmetries are almost certainly chemical. Each cell can, as it were, taste the soup inside itself, feed the information into its hand-held G P S microcomputer and get out a reading: 'you are in the rear half of the body, close to the underside.' Very nice to know where you are.

But knowing where you are is just the beginning. Knowing what you have to do once you are there is a wholly different problem.

Genes that control this process are known as 'homeotic' genes. For instance, our cell, on discovering where it is located, looks this location up in its guidebook and finds the instruction: 'grow a wing', or 'start to become a kidney cell' or something like that. It is not of course literally like this. There are no computers and no guidebooks, just a series of automatic steps in which gene switches on gene which switches on gene. But a guidebook is a handy analogy, none the less, because the great beauty of embryo development, the bit that human beings find so hard to grasp, is that it is a totally decentralised process. Since every cell in the body carries a complete copy of the genome, no cell need wait for instructions from authority; every cell can act on its own information and the signals it receives from its neighbours. We do not organise societies that way: we are obsessed with dragging as many decisions as possible to the centre to be taken by governments. Perhaps we should try.

Fruit flies have been a favourite object of geneticists' studies since the early years of the century, for they breed quickly and easily in the laboratory. It is the humble fruit fly we must thank for the elucidation of many of the basic principles of genetics: the idea that 1 7 6 G E N O M E

genes are linked on chromosomes, or Muller's discovery that genes can be mutated by X-rays. Among the mutant flies thus created, scientists began to find ones that had grown in unusual ways. They had legs where they should have antennae, or wings where they should have small stabilisers called halteres. A certain segment of the body, in other words, had done something appropriate to a different segment of the body. Something had gone wrong with the homeotic genes.

In the late 1970s, two scientists working in Germany named Jani Nusslein-Volhard and Eric Wieschaus set out to find and describe as many such mutant flies as possible. They dosed the flies with chemicals that cause mutations, bred them by the thousand and slowly sorted out all the ones with limbs or wings or other body parts that grew in the wrong places. Gradually they began to see a consistent pattern. There were 'gap' genes that had big effects, defining whole areas of the body, 'pair-rule' genes that subdivided these areas and defined finer details, and 'segment-polarity' genes that subdivided those details by affecting just the front or rear of a small section. The developmental genes seemed, in other words, to act hierarchically, parcelling up the embryo into smaller and smaller sections to create ever more detail.3

This came as a great surprise. Until then, it had been assumed that the parts of the body defined themselves according to their neighbouring parts, not according to some grand genetic plan. But when the fruit-fly genes that had been mutated were pinned down and their sequences read, a further surprise was in store. The result was the first of two almost incredible discoveries, which between them amount to one of the most wonderful additions to knowledge of the twentieth century. The scientists found a cluster of eight homeotic genes lying together on the same chromosome, genes which became known as Hox genes. Nothing strange about that.

What was truly strange was that each of the eight genes affected a different part of the fly and they were lined up in the same order as the part of the fly they affected. The first gene affected the mouth, the second the face, the third the top of the head, the fourth the neck, S E L F - A S S E M B L Y 1 7 7

the fifth the thorax, the sixth the front half of the abdomen, the seventh the rear half of the abdomen, and the eighth various other parts of the abdomen. It was not just that the first genes defined the head end of the fly and the last genes made the rear end of the fly. They were all laid out in order along the chromosome - without exception.

To appreciate how odd this was, you must know how random the order of genes usually is. In this book, I have told the story of the genome in a sort of logical order, picking genes to suit my purpose chapter by chapter. But I have deceived you a little in doing this: there is very little rhyme or reason for where a gene lies.

Sometimes it needs to be close to certain other genes. But it is surely rather literal of Mother Nature to lay these homeotic genes out in the order of their use.

A second surprise was in store. In 1983 a group of scientists working in Walter Gehring's laboratory in Basel discovered something common to all these homeotic genes. They all had the same

'paragraph' of text, 180 'letters' long, within the gene - known as the homeobox. At first, this seemed irrelevant. After all, if it was the same in every gene, it could not tell the fly to grow a leg rather than an antenna. All electrical appliances have plugs, but you cannot tell a toaster from a lamp by looking at the plug. The analogy between a homeobox and a plug is quite close: the homeobox is the bit by which the protein made by the gene attaches to a strand of D N A to switch on or off another gene. All homeotic genes are genes for switching other genes on or off.

But the homeobox none the less enabled geneticists to go looking for other homeotic genes, like a tinker rooting through a pile of junk in search of anything with a plug attached. Gehring's colleague Eddie de Robertis, acting on no more than a hunch, went fishing among the genes of frogs for a 'paragraph' that looked like the homeobox. He found it. When he looked in mice, there it was again: almost exactly the same 180-letter string - the homeobox. Not only that, the mouse also turned out to have clusters of Hox genes (four of them, rather than one) and, in the same way as the fruit fly, the 1 7 8 G E N O M E

genes in the clusters were laid out end-to-end with the head genes first and the tail genes last.

The discovery of mouse—fly homology was bizarre enough, implying as it does that the mechanism of embryonic development requires the genes to be in the same order as the body parts. What was doubly strange was that the mouse genes were recognisably the same genes as the fruit-fly genes. Thus the first gene in the fruit-fly cluster, called lab, is very similar to the first gene in each of three mouse clusters, called a1 , b1 and d1, and the same applies to each of the other genes.4

There are differences, to be sure. Mice have thirty-nine Hox genes altogether, in four clusters, and they have up to five extra Hox genes at the rear end of each cluster that flies do not have. Various genes are missing in each cluster. But the similarity is still mind-blowing.

It was so mind-blowing when it first came to light that few embryolo¬

gists believed it. There was widespread scepticism, and belief that some silly coincidence had been exaggerated. One scientist remembers that on first hearing this news he dismissed it as another of Walter Gehring's wild ideas; it soon dawned on him that Gehring was being serious. John Maddox, editor of the journal Nature, called it 'the most important discovery this year (so far)'. At the level of embryology we are glorified flies. Human beings have exactly the same Hox clusters as mice, and one of them, Cluster C, is right here on chromosome 12.

There were two immediate implications of this breakthrough, one evolutionary and one practical. The evolutionary implication is that we are descended from a common ancestor with flies which used the same way of defining the pattern of the embryo more than 530

million years ago, and that the mechanism was so good that all this dead creature's descendants have hung on to it. Indeed, even more different creatures, such as sea urchins, are now known to use the same gene clusters. Though a fly or a sea urchin may look very different from a person, when compared with, say, a Martian, their embryos are very similar. The incredible conservatism of embryological genetics took everybody by surprise. The practical application was that sud-S E L F - A S S E M B L Y 1 7 9

denly all those decades of hard work on the genes of fruit flies were of huge relevance to human beings. To this day, science knows far more about the genes of fruit flies than it knows about the genes of people. That knowledge was now doubly relevant. It was like being able to shine a bright light on the human genome.

This lesson emerges not just from Hox genes but from all developmental genes. It was once thought, with a trace of hubris, that the head was a vertebrate speciality - that we vertebrates in our superior genius invented a whole set of new genes for building a specially

'encephalised' front end, complete with brain. Now we know that two pairs of genes involved in making a brain in a mouse, Otx (1

and 2) and Emx (1 and 2), are pretty near exact equivalents of two genes that are expressed in the development of the head end of the fruit fly. A gene — called oxymoronically eyeless — that is central to making eyes in the fly is recognisably the same as a gene that is central to making eyes in the mouse: where it is known as pax-6.

What is true of mice is just as true of people. Flies and people are just variations on a theme of how to build a body that was laid down in some worm-like creature in the Cambrian period. They still retain the same genes doing the same job. Of course, there are differences; if there were not, we would look like flies. But the differences are surprisingly subtle.

The exceptions are almost more convincing than the rule. For instance, in flies there are two genes that are crucial to laying down the difference between the back (dorsal) of the body and the front (ventral). One, called decapentaplegic, is dorsalising - i.e., when expressed it makes cells become part of the back. The other, called short gastrulation, is ventralising - it makes cells become part of the belly. In toads, mice and almost certainly in you and me, there are two very similar genes. The 'text' of one, BMP4, reads very like the

'text' of decapentaplegic; the 'text' of the other, chordin, reads very like the text of short gastrulation. But, astonishingly, each of these has the opposite effect in mice that its equivalent has in flies: BMP4 is ventralising, and chordin is dorsalising. This means that arthropods and vertebrates are upside-down versions of each other. Some time 1 8 0 G E N O M E

in the ancient past they had a common ancestor. And one of the descendants of the common ancestor took to walking on its stomach while the other took to walking on its back. We may never know which one was 'the right way up', but we do know that there was a right way up, because we know the dorsalising and ventralising genes predate the split between the two lineages. Pause, for a second, to pay homage to a great Frenchman, Etienne Geoffroy St Hilaire, who first guessed this fact in 1822, from observing the way embryos develop in different animals and from the fact that the central nervous system of an insect lies along its belly while that of a human being lies along its back. His bold conjecture was subjected to much ridicule in the intervening 175 years, and conventional wisdom accreted round a different hypothesis, that the nervous systems of the two kinds of animals were independently evolved. But he was absolutely right.5

Indeed, so close are the similarities between genes that geneticists can now do, almost routinely, an experiment so incredible that it boggles the mind. They can knock out a gene in a fly by deliberately mutating it, replace it by genetic engineering with the equivalent gene from a human being and grow a normal fly. The technique is known as genetic rescue. Human Hox genes can rescue their fly equivalents, as can Otx and Emx genes. Indeed, they work so well that it is often impossible to tell which flies have been rescued with human genes and which with fly genes.6

This is the culminating triumph of the digital hypothesis with which this book began. Genes are just chunks of software that can run on any system: they use the same code and do the same jobs.

Even after 5 30 million years of separation, our computer can recognise a fly's software and vice versa. Indeed, the computer analogy is quite a good one. The time of the Cambrian explosion, between 540 and 520 million years ago, was a time of free experimentation in body design, a bit like the mid-1980s in computer software. It was probably the moment when the first homeotic genes were invented by one lucky species of animal from which we are all descended. This creature was almost certainly a mud-burrowing thing S E L F - A S S E M B L Y l 8 l

known — with delicate contradiction — as the Roundish Flat Worm, or RFW. It was probably just one of many rival body plans, but its descendants inherited the earth or large chunks thereof. Was it the best design, or just the most brilliantly marketed? Who was the Apple of the Cambrian explosion and who the Microsoft?

Let us take a closer look at one of the Hox genes on human chromosome 12. Hox C4 is the genetic equivalent of a gene called dfd in flies, which is expressed in what will become the mouthparts of the adult fly. It is also very close in sequence to its counterparts on other chromosomes, A4, B4 and D4 - and the mouse versions of the same genes: a 4, b4, c4 and d4. In the embryo of a mouse, these genes are expressed in the part that will become the neck: the cervical vertebrae and the spinal cord within them. If you 'knock out' one of these genes by mutation, you find that one or two of the vertebrae of the mouse's neck are affected. But the effect of the knock-out is very specific. It makes the affected vertebrae grow as if they were further forward in the mouse's neck than they are. The Hox 4 genes are needed to make each neck vertebra different from the first neck vertebra. If you knock out two of the Hox 4 genes, more vertebrae are affected, and if you knock out three of the four genes, even more cervical vertebrae are affected. Therefore, the four genes seem to have a sort of cumulative effect. Moving from head to rear, the genes are switched on one after another and each new gene turns that part of the embryo into a more posterior body part. By having four versions of each Hox gene, we and mice have rather more subtle control over the development of our bodies than flies do with just one Hox cluster.

It also becomes clear why we have up to thirteen Hox genes in each cluster rather than eight, as flies do. Vertebrates have post-anal tails, that is spines which go on well past their anuses. Insects do not. The extra Hox genes that mice and people have, which flies do not, are needed for programming the development of the lower back and tail. Since our ancestors, when they became apes, shrank their tails to nothing, these genes are presumably somewhat silent in us compared with their equivalents in mice.

l 8 2 G E N O M E

We are now in a position to face a vital question. Why are the Hox genes laid end-to-end, with the first genes expressed at the head of the animal, in every species so far investigated? There is as yet no definitive answer, but there is an intriguing hint. The foremost gene to be expressed is not only expressed in the foremost part of the body; it is also the first to be expressed. All animals develop from the bow to the stern. So the co-linear expression of the Hox genes follows a temporal sequence, and it is probable that the switching on of each Hox gene somehow switches on the next one in line or allows it to be opened up and read. Moreover, the same is probably true of the animal's evolutionary history. Our ancestors seem to have grown more complicated bodies by lengthening and developing the rear end, not the head end. So the Hox genes replay an ancient evolutionary sequence. In Ernst Haeckel's famous phrase,

'ontogeny recapitulates phylogeny'. The embryo's development occurs in the same sequence as its ancestors' evolution.7

Neat as these tales are, they tell only a fraction of the story. We have given the embryo a pattern - a top—down asymmetry and a bow-stern asymmetry. We have given it a set of genes that get turned on according to a clever sequence of timing and thus are each expressed in a different part of the body. Each Hox compartment has switched on its special Hox gene, which in turn has switched on other genes. The compartment must now differentiate in the appropriate way. It must, for example, grow a limb. The clever part of what happens next is that the same signals are now used to mean different things in different parts of the body. Each compartment knows its location and identity and reacts to the signals accordingly.

Our old friend decapentaplegic is one of the triggers for the development of a leg in one compartment of a fly and a wing in another.

It in turn is triggered by another gene called hedgehog, whose job is to interfere with the proteins that keep decapentaplegic silenced and thus to awaken it. Hedgehog is a so-called segment-polarity gene, which means it is expressed in every segment, but only in the rear half thereof. So if you move a hedgehog-expressing piece of tissue into the anterior half of the wing segment, you get a fly with a sort S E L F - A S S E M B L Y 1 8 3

of mirror-image wing with two front halves fused back to back in the middle and two back halves on the outsides.

It will not surprise you to learn that hedgehog has its equivalents in people and in birds. Three very similar genes, called sonic hedgehog, Indian hedgehog and desert hedgehog, do much the same thing in chicks and people. (I told you geneticists had strange minds: there is now a gene called tiggywinkle and two new gene families called warthog and groundhog. It all started because fruit flies with faulty hedgehog genes had a prickly appearance.) Just as in the fly, the job of sonic hedgehog and its scheming partners is to tell the compartment where the rear half of the limb should be. It is switched on when a blunt limb bud has already formed, telling the limb bud which way is rear. If at the right moment you take a microscopic bead, soak it in sonic hedgehog protein and insert it carefully into the thumb side of the wing bud of a chick embryo for twenty-four hours, the result will be two mirror-image wings fused front half to front half and with two back halves on the outsides — almost precisely the same result as in fruit flies.

The hedgehog genes, in other words, define the front and rear of the wing, and it is Hox genes that then divide it up into digits. The transformation of a simple limb bud into a five-fingered hand happens in every one of us, but it also happened, on a different timescale, when the first tetrapods developed hands from fish fins some time after 400 million years ago. In one of the most satisfying pieces of recent science, palaeontologists studying that ancient transformation have come together with embryologists studying Hox genes and discovered common ground.

The story starts with the discovery in Greenland in 1988 of a fossil called Acanthostega. Half fish and half tetrapod, and dating from 360 million years ago, it surprised everybody by having typical tetrapod limbs with eight-digit hands on the end of them. It was one of several experimental limb designs tried out by the early tetrapods as they crawled through shallow water. Gradually, from other such fossils, it became clear that the hand we all possess developed in a curious way from the fish's fin: by the development of a forward-curving arch 1 8 4 G E N O M E

of bones in the wrist from which digits were flung off towards the rear (little-finger) side. You can still just see this pattern in an X-ray of your own hand. All this was worked out from dry bones of fossils, so imagine the palaeontologists' surprise when they read of the embryologists' discovery that this is exactly how the Hox genes go about their work in the limb. First they set up a gradient of expression curving towards the front of the growing limb, to divide it into separate arm and wrist bones, then they suddenly set up a reverse gradient on the outside of the last bones to throw off the five digits.8