I have a kind of chequered relationship with [Charles] Weismann. As a boy, I guess I read almost everything written by George Bernard Shaw. And you know that Shaw was a Lamarckian and deeply anti-Darwinian in his thinking about evolution. And he treated Weismann as a cruel and ignorant and stupid German pedant who had done this absurd experiment of cutting the tails of mice in order to demonstrate that acquired characters were not inherited. And Shaw says 'Look, you wouldn't expect the tails... taillessness to be inherited, because it isn't something the mouse did of its own volition, it's something that was done to it, and it's only things that the creature does out of its own inner drive that are going to be inherited.' And I grew up with the conviction that Weismann was, you know, was, indeed, an ignorant German pedant. And I was about to say something, I think I actually wrote something quite impolite about Weismann in something I was writing, and I thought 'Well, wait a minute now, I wonder whether it's true that he thought this.' And I went round to the library and started reading his big fat book that he wrote towards the end of his life and discovered that it was totally untrue and, I think that after Darwin, Weismann is my hero as an evolutionary biologist. I mean, first of all, to get rid of the mice story. He has this great thing about how the first time he gave a paper, at a German zoological society, explaining why he thought that acquired characters were not inherited, people leapt up and said 'But this has to be wrong, everybody knows that if you dock the tail of a bitch, she will produce puppies that have no tails, so what you're saying has to be wrong.' And Weismann... indeed, they talked a lot about the inheritance of wounds, and it was claimed that even duelling scars were inherited, fortunately only by the sons but not by the daughters, which was handy. And so Weismann said that he... he had to do the mouse experiment just to get rid of this particular piece of nonsense, but that's not why he thought acquired characters were not inherited at all. And it's... it's fascinating to read him on this. He has two big chapters in the book on why he thinks that they aren't inherited. And he discusses classical examples, like the adaptations of the worker bees cannot possibly be inherited by the inheritance of acquired characters because their ancestors have never had those habits. Or he points out that the... the argument that the whole outer skeleton of an insect - the adult insect - is hardened before it is used. And so there's no way that use could modify the shape of the exoskeleton of an insect so as to become adaptive. And he gives a number of good arguments of that kind. But, at the end of the chapter, he obviously realises he hasn't really quite said what he wants to say. And he says 'But the real problem is,' and I can't remember his exact words, 'that how could it be that some adaption of the body could be transmitted to the germ cells and somehow or other change the germ cells so that when the germ cells form a new individual, they somehow would have acquired the information from the body and can transmit it to the new individual.' And to prove that he really did have this notion of information, he has this extraordinary sentence, which I can more or less quote, though it's an English translation, I'm afraid. He says 'If you came across the case of the inheritance of an acquired character, it's as if a man sent a telegram to China and it arrived translated into Chinese.' And what's fascinating about that is that he does know that heredity is about information. He knows... he has the notion of translation. But also, interestingly, he does use the mechanical analogue of the telegram.