Once, the only way to measure the rate of new mutations to haemophilia (or any other inherited illness) was to count the sufferers, estimate the damage done to their chances of passing on the error and work out from this how often it must happen. Technology has changed everything. Now it is possible to compare the genes of haemophiliac boys with those of parents and grandparents to see when the mutation took place.

If the mother of such a boy already has the haemophilia mutation on one of her two X chromosomes, then she must herself have inherited it and the damage must have occurred at some time in the past. If she has not, then her son's new genetic accident happened when the egg from which he developed was formed within her own body. In a survey of a British families with sons with haemophilia B (whose gene, that for Factor IX, is 33,000 bases long) many different mutations were found, most unique to one family. Eighty per cent of the mothers of affected boys had themselves inherited a mutation. However, in most cases the damaged gene was not present in their own father (the grandfather of the patient). In other words, the error in the DNA must have taken place when his grandparental sperm was being formed.

A quick calculation of the number of new mutations against the size of the British population gives a rare for the haemophilia B gene of about eight in a million. The difference in the incidence of changes between grandfathers and their daughters suggest that the rate is nine rimes higher in males than in females. The sex difference is easy to explain. There are many more chances for things to go wrong in men (who — unlike women — produce their sex cells throughout life, rather than making a store of them early on, and hence have many more DNA replications in the germ line than do females). For some genes the rate of mutation among males is fifty times higher than in the opposite sex. Men, it seems, are the source of most of evolution's raw material.

Most people with severe forms of haemophilia have each suffered a different genetic error. Such mistakes happen in a parent's sex cells and disappear at once because the child dies young. Those with milder disease often share the same change in their DNA; an error that took place long ago and has spread to many people. The shared mutation is a clue that these individuals descend from a common ancestor. The non-functional DNA in and around the haemophilia gene is full of changes which appear to have no effect at all and have passed down through hundreds of generations. Near the gene itself is a region with many repeats of the same message. The number of copies often goes up and down, but its high error rate seems to do no damage.

All this hints that mutation is an active process, with plenty of churning round within the DNA. This new fluidity once alarmed geneticists as it violates the idea of gene as particle (admittedly a particle which sometimes makes mistakes) which used to be central to their lives. So powerful is the legacy of Mendel that his followers have sometimes been reluctant to accept results which do not fit. This is very true of some of the new and bizarre aspects of mutation.

Scientists, in general, despise doctors, loi in.my years, physicians reported a strange genetical effect called 'anticipation*. The malign effects of some inherited disi-.isi's seemed to show themselves at a younger age with mcli generation that passed. The effect was named eighty years ago by an enthusiastic eugenical doctor called Mott. He thought that it presaged the inevitable degeneration of society: 'The law of anticipation of the insane represents… rotten twigs continually dropping off the tree of life.' Later geneticists were resistant to the idea and it disappeared from view. In fact it represents a new kind of mutation, an inherited error which gets worse as the generations succeed and which is now known to be common.

The process is seen in a disease called the fragile X syndrome, the most important single cause of inborn mental impairment, with symptoms that range from mild to crippling. Many children diagnosed as autistic have in fact a minor form of this illness. At first sight its inheritance is odd, as in some families it is found in just one person, whereas others have dozens of affected members. Boys tend to suffer more damage than do girls, with mental retardation and, sometimes, a characteristic face with large ears, and heart problems.

Although both males and females are affected, the gene is sex-linked. Males never pass it on to their sons, but girls whose mothers have it are carriers, and some may be affected. Fragile-X is one of the few genetic diseases in which the damage can be seen down the microscope, for near the end of every affected X chromosome is a small constriction which looks as if it might be about to break. About one woman in two hundred and fifty has one or other X chromosome damaged in this way. Many show no symptoms at all and neither do their children. Others have signs of the disease, as do their offspring, while a fraction albeit themselves normal may have children with the illness.

The mutation is a multiplication of a three-letter DNA repeat — C, G and G — within a gene. Its protein helps form the connections made as the young brain begins to respond to experiences from the outside world. The damage is not in the coding section of the gene, but in its on-off switch. The mutation is flexible. Most people have thirty or fewer repeats, and some have as few as half a dozen. When the number creeps above fifty or so, children are in danger and may find it hard to speak or to read and as it rises over two hundred (and severely affected individuals have more than a thousand repeats) the full symptoms set in.

The strangest aspect of fragile X — and, we now know, of many other mutations — is that the number of repeats (and the amount of damage) changes from generation to generation. The daughter of a mother who has fragile X is more likely to have an affected child than was her own parent although (or so it seems) she has passed on exactly the same gene. Each generation, the number of copies changes, going up when it is transmitted through a female, but staying the same or decreasing when a man passes on the damaged chromosome.

One form of muscular dystrophy also shows more virulent effects as the generations succeed. Again, a repeated sequence is involved. The pedigrees of one group of the children with the disease shows that all shared an ancestor who b'ved in the seventeenth century. He was healthy; as were, for two hundred years, his descendants; but suddenly some, distant relatives though they now are, began to suffer from dystrophy. More copies of a DNA repeat within the crucial gene had been made each generation. Once a critical number was reached the symptoms appear. Each generation, more and more appear, and the effects of the damaged gene become more severe as it passes down the lineage. Huntington's Disease, too, is duo to a repeal of the three DNA letters CACi. Each triplet codes Inr.1 single amino acid, which shoulders itself into tin* inuklle ul tin¹ hunt-ingtin molecule. Some people liavi- li-wcr than uti copies, some more than a hundred. Once more tluui.ihoui thirty-five copies are made, the symptoms of the disease emerge, and the more repeats, the earlier they do so. Those with fifty are in danger of illness while still in their twenties. Half a dozen other diseases of the nervous and muscular system are due to such three-letter intruders. Why nerves should be so prone to them is not certain, although the tendency of such augmented proteins to form great clumps in the cell may have something to do with it.