Larry Gilbert is a great non-publisher, and I'm not sure he's ever published his PhD work - has he? I've never seen it. I've always had to quote it as an anecdote. And indeed, for quite a time after I came back to Sussex and started talking about game theory, Larry Gilbert was known as John's imaginary biologist, they refused to believe that he existed until I actually produced him in the flesh and he said, 'No, no, it's all quite true.' I'll tell you one story about Nick, I think Nick Davis probably thought up his experiments on speckled wood butterflies without knowing about Larry Gilbert's work, but it was essentially a similar ownership game. But I do remember Nick once saying at a meeting that he'd like to tell the meeting, it was a small meeting, what John Maynard Smith had done for field workers, and it went as follows, he said: 'You know, that in the old days, what we had to do if we were a graduate student doing field work, we had to go and demonstrate that strategy A was fitter than strategy B, to show that selection was operating. And if you looked at a thousand cases and there was no significant difference, your supervisor told you to go out and measure another thousand until it was significant.' But what I'd done was to demonstrate that at an ESS the two strategies should be equally fit, and so frankly, the fewer you looked at, the more likely they were to fit the theory, so you didn't have to look at thousands, you just looked at a few. I don't think he meant this very seriously, because, as you know, he's a very meticulous experimenter. He was one of the first people in behaviour to really cotton on to game theory as being something he could use to explain what his beasts were doing.