The title of this paper was deliberately ambiguous. It could be taken to concern either what will happen in evolution in the future, or alternatively the future study of evolution, especially genetics. Professor Dawkins addressed both issues, with the second providing the major focus of the discussion.

Concerning the first issue, future evolutionary developments, Professor Dawkins stressed that it is generally accepted that we cannot extrapolate from past trends. Evolution is not progressive in the way suggested in the later years of the 19th Century and the early years of the 20th Century, whereby humanity was regarded as the apex of the chain of progression. The question was raised: what could be a good reason to extrapolate from past facts about evolution?

Another kind of progression in evolution was noted, that of the `arms race' in nature. If we take the case of predators and prey, we tend to find that development on one side tends to call forth improvements on the other. Evolutionary developments tend to progress in the same direction in this sense. Were it the case that human brain size increase was caused by predators and that the predators still existed then we might be able to extrapolate from this to future increases in brain size. There is, however, no such predator-brain size connection.

Another reason why limited extrapolation might be possible concerns the fact that evolution tends to culminate in the same endpoint in different species in different times. There are remarkable parallels between the evolution and diversity of present big cat species and the diversity of earlier marsupial species.

The same pattern of diversity is likely to appear in different evolutionary scenarios. One example concerns the variety of species of dinosaurs and the similar variety in present mammalian life (the ichthyosaur's similarity to the dolphin is one example). It is the same play with a different cast. Were another extinction level event to occur, 50 million years after this we could expect the same sort of range of animals to inhabit the Earth (the example of an extinct, hippo-sized relative of the guinea pig was used to illustrate this suggestion). This line of thought is followed in Dougal Dixon's `After Man', described by Professor Dawkins as `charmingly speculative'.

To summarise, we might be able to predict the future in very limited, general terms, possibly driven by arms races.

The second, major theme of the talk concerned the future study of evolution. Professor Dawkins started his discussion of this issue by noting that Peter Medawar claimed that it was `not worth arguing' with anyone who failed to recognise that Watson and Crick's work was the greatest scientific achievement of the 20th Century. What changed as a result of their work was that genetics became digital. After Mendel, genes could have been said to be digital in the sense that you either get them from (e.g.) your father or you don't. However, Mendelian genetics still allowed the possibility that individual genes were just chemicals.

Watson and Crick demonstrated that genes were digital through and through, to their very essence. Genetics can be truly seen as a branch of IT. The fact that we can count the number of letters by which gene chains differ would have been inconceivable prior to Watson and Crick's research.

Developments in genetics grew exponentially post-Watson and Crick. Professor Dawkins drew the analogy with the explosive (exponential) progress found within I.T., known as Moore's Law. This states that the processing power of computers doubles every eighteen months. Is there a `Son of Moore's Law' for genetics?

One way of answering this is to use the cost of sequencing one DNA nucleotide base pair as our economic benchmark. In 1965, the cost for one letter of bacteria RNA was £1000; in 1975, it was £10 for one letter of virus code; in 1985, a letter of nematode DNA cost £1, and in 2000 one letter in the human genome project cost around £0.10. Based upon these `back of envelope' figures, the cost-halving time in genetics is 27 months. With the proviso that extrapolating upon these is wildly speculative, this suggests that it will be possible to sequence an individual human genome in 2050 for £100.

If this occurs, it will be possible for any two people to establish the exact relation between them. It will be possible to reconstruct great migrations, journeys, etc., by gene comparison. Doctors will be able to give individually tailored prescriptions based upon your individual genome. For the price of an x-ray now your entire DNA could be made known.

If it becomes that cheap to sequence the genome of an individual human, it will be even cheaper for other species. This could render it economically affordable to plot the unique tree of life--that is, to show how all species are related. This has already thrown up surprises-- for example, it has been shown that hippos are more closely related to whales than to pigs. This was obviously a surprise to zoologists.

`Son of Moore's Law' wiill reveal more such surprises. Sydney Brenner, who coincidentally won the Nobel Prize within a week of this presentation, also did some `crystal ball gazing' in 2000. He suggested three tasks for molecular biology:

(i) compute the protein sequence from the DNA sequence;

(ii) compute the folded (3D) structure of protein from its sequence (this is difficult but not inconceivable);

(iii) compute the embryo from its genes.

Brenner suggested that the third would be possible by 2050. If this is so, we will be able to find out the embryology of a creature, which in turn will give us the behaviour of the adult creature.

It was suggested that by 2050 a Genetic Book of the Dead will be a reality. From the DNA of a creature we will be able to generate an account of the environment of its ancestors. Genes are a coded description of the world in which they were selected in ancient times. Camel DNA, for example, was sculpted by the desert world of ancestors and the oceans of earlier relatives. By 2050 we will be able to read the language of the animal--not only its embryology, but its predators, prey, environment and so on.

What about individual DNA? Will `Jurassic Park' style scenarios be possible? Sadly, individual DNA doesn't last long enough for this. But there will be ways of using the DNA of living creatures to reconstruct ancestors.

Brenner suggests that when the chimp genome is fully known, it will be possible to compare chimp and human DNA to reconstruct the genome of the `missing link' (6 million years ago), the ancestor we share, and even to grow a new missing link. Professor Dawkins noted that we could go further and split the difference to grow new Lucys (3 million years). Clearly, there will be ethical problems with such projects.

As ethicists worry about such matters, geneticists could start work on reconstructing dinosaurs. Focusing on birds, we might be able to advance the genome in an ostrich egg to hatch a `terrible lizard'. Professor Dawkin's greatest regret is that he will not be around in 2050 to shake the hand of a new Lucy.