In the second part of our series looking forward to the 21st century, Britain's foremost fertility expert explains how the science of human reproduction may transform the way we live
In 1677, Anton van Leeuwenhoek used the newly invented microscope to examine animal spermatozoa. His observations soon led to descriptions of human sperm, seen swimming in fresh seminal fluid. But, with their home-ground lenses, microscopists of that time imagined that they could see a homunculus inside the sperm head; and artists drew pictures of what they thought they saw. There is a fine woodcut by Hartsoeker, in the Wellcome collection in London, of a perfectly formed tiny man with over-sized head, curled up in a foetal position, ready to hatch from the sperm head.

In the 18th century, an obscure rabbi, Elijah Pinhas ben Meir, used these findings as ethical justification against the practice of masturbation: "So the sperm contain little men, fashioned in man's image. Now we can see how the words of the old sages who wrote the Talmud 1,500 years ago were true. Clearly, destruction of the seed is in effect killing the homunculus and is akin to murder." Had sperm really contained little men, the destruction of semen would indeed have been like murder. But the rabbi's polemic was based on flawed observations, on bad science. Ethical values must be based on the truth, on the best factual assessment of the natural world that we have. And that is still one of science's greatest challenges in the coming century.

There is hardly an area of science where such challenges are greater than in human reproduction. Only a fool believes he can predict what will happen, even in his own field, with any accuracy. And at the turn of the millennium, when every newspaper is importuning reluctant "experts" to write of their view of the future, it seems foolhardy to offer my own predictions. But the subject of human reproduction is a microcosm for many of the ethical problems raised by science. And society's attitude to such problems in future and their various solutions will have a profound influence on scientific events. It is already largely forgotten that in vitro fertilisation (IVF), the "test-tube babies" advance pioneered by Drs Edwards and Steptoe, was almost banned before it really began. Had Parliament accepted the arguments of some right-wingers in 1985, many families, now happily complete, would not exist today.

Humans are among the most naturally infertile animals, but the reasons for that have only just started to surface. One of the most unexpected facts is that human existence is even more fragile than we thought. Only about one in every five human embryos seems capable of implantation and development into a full-grown baby. The majority perish - but why is mysterious. A significant proportion, at least one-quarter of all human embryos, is demonstrably malformed. Such embryos have abnormal numbers of chromosomes, which usually, perhaps fortunately, makes them incapable of independent life. Nearly all such embryos die. A few - around 1 per cent - survive. They emerge from the womb with severe abnormalities, such as Down's syndrome.

It seems that Nature herself is capable of efficient antenatal diagnosis, aborting those embryos she considers unfit. Nature practises eugenics much more frequently and effectively than those doctors called upon by their patients to undertake pregnancy termination, who are vilified by "right to life" groups in consequence.

Using IVF techniques, it is already possible to examine embryos at the stage of development when they are often destroyed by nature. Such pre- implantation diagnosis enables doctors to treat families at risk of having a child with a fatal gene disorder. Initially, DNA tests were used to establish the sex of an embryo in those families who had lost children from some sex-linked disorders, such as muscular dystrophy. Screening for cystic fibrosis and one or two other fatal genetic disorders is now also just possible. Within a few years of completion of the human genome project, we should be able to screen for a great number of the 6,000 or so single-gene defects that regularly cripple, then kill babies and children. Additionally, there will be a number of inherited cancers - a particularly malignant form of breast cancer mainly affecting young women; a familial bowel cancer that destroys whole families; a type of brain cancer mostly affecting children - that will be preventable by those methods.

The majority of people in this country have little problem with such treatments. They feel human life cannot be regarded as starting at conception, but rather that it is part of a continuum. Fertilised eggs are not thought by most people to be little babies, as, unlike Rabbi Elijah, most people have accurate scientific evidence on which to base their ethical attitudes. That, surely, is why contraception is not regarded as a sin by most Christians in Britain.

This Millennium sees us on the brink of an even more important advance. In the first four or five days after fertilisation, cells taken from the embryo could be manipulated so that they develop into embryonic stem cells. Stem cells could be kept in cultures where they would multiply as a continuously dividing resource. Thereafter, with appropriate treatments, cells taken from such cultures could be made to develop into any of the different cell types in the body.

So-called tissue engineering could be one of the greatest therapeutic advances of the next century. The value of liver cells to a patient dying of hepatic failure; neurons to a sufferer crippled by degenerative brain disease; bone marrow to a leukaemia victim; hormone-secreting cells to a severe diabetic; or cardiac muscle for a patient at risk of death after a severe heart attack would be incalculable. One problem is that, though slow progress is being made, we do not fully understand how to direct embryonic stem cells to grow along a particular path of differentiation. Various growth factors can be added to the embryonic cell cultures from mice, but procedures remain hit-and-miss.

One tissue-engineering process with considerable potential is so-called therapeutic cloning. The word "cloning" is really a misnomer but it has stuck because the process utilises somewhat similar techniques to those used to produce Dolly the sheep. The nucleus from an adult differentiated cell of the desired type could be removed and placed in the embryonic stem cell. With a new command and control centre, that cell could then generate a multitude of cells of the desired type. Thus a baby suffering from 50 per cent burns might be grafted with new skin, all generated from embryonic tissue initially treated with a transplanted nucleus from a tiny sample of skin. If the nucleus was from the burnt baby's own healthy skin, any skin tissue generated would have the huge advantage of being immunologically compatible - it would not be rejected.

It has been said that it would require a great number of human embryos to produce a wide variety of tissues. That is not true. Once stem cell cultures were established they would be self-replicating, a banked resource, readily available for many treatments. But the main argument used against this work, and which caused the Government to be reluctant to allow human research in this field, is that the use of any embryonic cells may be morally unacceptable to a majority of the public.

Surely that attitude is ethically wrong. Embryos at this stage of development perish during IVF and are destroyed by different methods of contraception. More important, Nature herself destroys many of the embryos that could be used to generate tissues to save countless lives. Which harks back to Rabbi Elijah. An understanding of embryology leads to the view that fertilised eggs are not human beings. True, they have the potential to be. But the moral imperative is surely to save existing life wherever we can. Perhaps that potential may be justifiably directed to saving the lives of existing people. Rather than delay, we should be trying to help as many dying patients as possible, just as soon as it is technically feasible.

So, reproductive science in the next century - indeed all science - will depend largely on society's evaluation of the ethics of its application. There will be increasingly difficult choices to make, particularly in biology because of its importance in treating disease. Such choices must depend on the clearest understanding of their scientific basis and a proper evaluation of how technologies may protect and enhance human lives. Without knowing how society will choose, predicting is pointless.

Tomorrow: David Thomson on 21st-century cinema