Males over thirty-five are the crucial group; but the data for men are harder to gather than are those for their partners. Husbands and wives tend to share, if nothing else, about the same date of birth so that, in spite of a few aged (and anonymous) Lotharios, the figures for women contain most of the information for men. Aristotle advised that girls marry at eighteen, and men at thirty-seven. Although the age difference between the sexes is less than that husbands do tend to be three to five years older than their wives.

The changes in sexual pattern are most obvious in postwar Europe. In Britain, Poland and Switzerland the proportion of mothers over thirty-five — and hence of fathers over about forty, the group most at risk of mutation — dropped from around twenty per cent in 1950 to less than five per cent in 1985. In that year, just one mother in fifty in what was then East Germany was more than thirty-five years old, a figure probably smaller than any time in history. In Ireland the influence of the Church, and the many young men who spent a period working overseas, meant that until a few years ago the only means of birth control was self-denial. Most Irish people did not marry until their late twenties, or later, and until not long ago almost a third of all mothers {and a higher fraction of fathers) were over the crucial age; more than twice the proportion anywhere else in Europe. The number is now much lower (albeit still above the average). There has been some reversal of the trend over the past two decades, with the numbers of mothers over thirty-five increasing from its low point of around one in twenty.

The general picture remains clear: old mothers (and fathers) are rarer than they have been for much of the past. This is bound to have an effect on the mutation rate. Down's syndrome (ten times more frequent among mothers over forty-five than in teenagers) is three times more common in Pakistan (which has almost no family planning) than in Britain, because Pakistani mothers are older than their British equivalents. From the male point of view, in Britain the mutation rate in men is about one and a half times that expected if all fathers were less than thirty, but in Pakistan it is three times this low figure. At the moment, at least, it looks as if the human mutation rate is on the way down. Whether this will continue is not certain, but it puts fears about a new race of mutated monsters into context.

Mutation is the fuel of evolution but, as far as can be seen, evolution rarely runs out of steam. Natural selection, though, is its engine and, like most engines, often speeds up and slows down to face changing circumstances.

Selection is an elusive process and it is more difficult to forecast what its future might be. Nature is always liable to come up — as it has so often before — with a nasty shock. The emergence of the AIDS virus shows the risk of this happening again. Even so, in the western world at least, some of the greatest challenges have gone, because of the control of infectious disease. Once such a disease has disappeared the future of the genes that combat it will change. Cypriots carry the inherited anaemia beta-thalassaemia because it defended their ancestors against malaria. That illness has now disappeared from the island — as, in time, will thalassaemia, with the incidence of carriers dropping by as much as one per cent per generation. In time, and given success in public health, the same will happen to the many other genes that resist the infection elsewhere in the world. Soon, they will remain only as mute and waning witnesses to an ancient past.

Life has also got better for babies. They are important; changes in the survival of adults — essential to individuals as they may be — are of not much relevance to selection because they kill those who have already passed on their genes. What counts for evolution is death before reproduction. What happens to the rest of us is, more or less, beside the point. The history of one inherited character, the weight of newborns, shows just how effective improved conditions at an early age can be in reducing the action of natural selection.

At birth, it pays to be average. Underweight babies, needless to say, survive less well than do others. However, babies heavier than normal are also more likely to die in the first few weeks. In the 1930s about half the babies who died in their first year did so because they were above or below the ideal weight. A difference of just one pound had a large effect. Since some of the variation in this character is genetic, natural selection was at work against genes for extreme birth weight as it had been, no doubt, since our species began.

Now, such selection has almost gone. Improved care means that only very underweight babies, or those much larger than average, are at risk. The intensity of selection has gone down by more than two thirds since the 1950s. Nowadays there is little risk in being a baby of even a kilogram above or below the mean weight. What was once a powerful agent of evolution is on the way out.

Improved child-care has also changed the ratio of the sexes at the age when people begin to choose a mate. At birth, there are slightly more boys than girls. Boys once had less chance of surviving the hazards of childhood, which led to an almost exact balance of the sexes in the late teens. Now, boys survive almost as well as girls do, so that in future there will be a slight but noticeable excess of young men looking for a mate. If (and many dispute the idea) the differences between men and women in size, or in appearance, are driven by sexual selection, perhaps in years to come that aspect of our evolution will (unlike most of it) advance, to give a generation of taller, hairier and more libidinous males.

There are better ways of looking at the future of selection than just to multiply examples of how it works. Natural selection acts only on differences. If everyone lived to adulthood, found a partner and had the same number of children (whether that number was one, two, or ten) it could not operate. We do not need to know what genes are involved to estimate how important selection might be. Simple changes in the pattern of birth and death reveal its actions in the past, the present and, perhaps, in the future.

In affluent countries, the differences between families in how many people survive have much decreased. This much reduces the power of the evolutionary engine. Ten thousand — even two hundred — years ago, the struggle for existence meant a lot. Skeletons from cave cemeteries show that few lived to be more than twenty. If ancient fertility was like that of modern tribal groups each female had about eight children, most of whom died young. For nine tenths of human evolution, society was like a village school, with lots of infants, plenty of teenagers and a few — probably harassed — adult survivors. Almost every death was potential raw material for selection as it involved someone young enough still to have a hope of passing on their genes. Now, ninety-eight our of every hundred new-born British babies live to the age of eighteen, so that selection acting through the deaths of the young once its main mode of operation has almost disappeared.

Not until the past few years have humans lived as long as they are able. For the first time in history, most people die old, perhaps as old as biology allows. Life expectancy has risen from forty-seven to seventy-six years in the past century. Progress has now stopped, for some social classes at least. In the USA in 1979 a white woman of sixty-five could expect to five for another eighteen and a half years. In 1999, the figure was almost the same. In Britain, even if all infectious diseases and all accidental deaths were to be eliminated by government decree, average life expectancy would go up by only a little more than a year. There is still room for progress because of class differences in health. A baby born to an unskilled worker in Britain can expect to live for eight years less than one born to a professional person, a difference which, to our national shame, was until recently increasing. In spite of the effects of class, the prospects for any dramatic improvement in longevity are dim. George Bernard Shaw was wrong. We will not go back to Methuselah.