So I consider a sort of, a cell, which is a little bag with molecules inside, so I'm looking at little populations of molecules, and they're going to be able to grow by absorbing stuff from the outside, and then from time to time simply fission; just, the cell will divide into two either because of some surface instability or because it's chopped up by rainstorms or something, so by natural processes the drops divide and then continue to grow and then divide and continue to grow. But there's no replication in the proper sense, there's no exact copying, merely the population just randomly accumulates. And the question is then, could there be an evolution of droplets of this kind without replication, and I think the answer is yes. So that's the model that I have in mind. So it's a system that has the attribute of metabolism; that is, a living process that brings in molecules from the outside.

[Q] These would be primarily proteins rather than...

Nucleic acids. Yes. I mean it's probably neither, it's probably a much more heterogeneous mixture. But anyway, things that eventually evolve into proteins. But the point is that you don't need to have a very low error rate for this to work. So you imagine that the population of molecules has a certain tendency to catalyse its own reproduction, but I make the distinction between reproduction and replication. So you reproduce a population approximately, but you don't replicate the individual molecules. So the population is reproduced by catalysis, by... because the molecules themselves catalyse other molecules to be reproduced, so the population as a whole is being reproduced with a very high error rate, and nevertheless it evolves. And... so you can actually then... I made a little mathematical model of this, and it turned out it did very nicely. There's a certain natural... it's a purely mathematical exercise but it sets a natural scale for the population that... which comes only from natural constants, not constants of physics but constants of mathematics. And it... I mean it goes with essentially the fact that e⁴ - e is the exponential, is 54, which is sort of a reasonable number, and the quality factor for the catalyst has to be at least that good. So there's a natural sort of lower limit on the quality of the catalysts, but that's a very easy limit. 54 is pretty low, because even poor catalysts easily achieve a factor of 100. Biological catalysts, proteins, have quality factors like 10⁴, or 10,000 or higher. So you don't need the modern very efficient proteins to make this work. If you have these very poor catalysts, it turns out that the thing actually can lift itself up statistically by its boot straps and you get an evolution toward better quality. And it turns out that the error rate at this saddle point where you change over from disorder to order is 1/3, which is very satisfactory. So you're dealing with... really with messy mixtures of molecules, which is what you expect at the early stages, so it's totally different from the picture you have with the replicating system where you need error rates like 1 in 1000. So it's a totally unorthodox view but I think it could in fact be right. So I wrote this little book called Origins of Life expounding this point of view and it had an interesting history because it's been totally ignored by biologists and it's been welcomed by physicists and by all kinds of other readers, sort of generally non-expert readers, but made a case which is convincing to everybody except the experts, which is quite normal for revolutionary ideas. I have a... I know my son, George, who is one of my honest critics, says, 'This is what you're going to be remembered for, not the physics', and maybe he's right. Anyhow, I'm happy to say that this book is now going into a 2nd edition, that the Cambridge Press who published it 12 years ago finds there's enough demand, so they're going to publish a new edition next year, and I'm rewriting it for the second edition, putting in some more modern information. So by and large I would say that the experiments are still completely neutral, they... there's no evidence one way or the other whether replication came first or whether metabolism came first, it's still an open question. The biologists all believe that replication came first. I think the reason is clear, because they work with RNA all the time: RNA is such a wonderful experimental material, it's like the hydrogen atom for biology; they can do these beautiful experiments with RNA evolving in a test tube and it... RNA does all sorts of lovely things, it's perfect experimental material. Including acting as a... an enzyme. Right. Now they discovered ribosymes which are these RNA enzymes, which only confirms the orthodox view that RNA was the original stuff. So that's now they talk about the RNA world as having preceded the protein world, which I think is a fantasy, but... it's the party line among biologists. So anyway, I just am quite patiently waiting to see. I think in the end there's a good chance I might turn out to be right, but I probably won't live to see it, and in any case, I don't care.