What we'd become preoccupied by was the business of mis-sense and non-sense. I mean, sense and nonsense. A very brilliant theory had been produced by Griffith, Crick and Orgel, called the commaless code. Now, the commaless code is very interesting. What you do is you take triplets and you say that, 'Can you find a sequence of triplets which can only be read in one frame, so that every other frame is nonsense?' And you can. And you can prove it gives you exactly 20 again. And this again the magic number comes out. So you can write codes called comma-less codes, where in one frame it makes sense, and every other frame it makes nonsense. And so that gives you 20 sense triplets and 44 mis-sense triplets. And so one of the things we wanted to know was: how frequently is there nonsense in the code? Because, clearly, if a lot of triplets correspond to no nonsense, you should have very high targets for destroying the meaning of a message. You know, most of the mutations should give you nonsense. And certainly… so… but we needed to get away, and so there was the other thing which was called the... which I called the Humpty Dumpty model, which was: how do you get the phase of a message? You start at the beginning and read on in threes till you come to the end; if you remember, that's what Humpty Dumpty tells Alice about how to tell the story and so on. And that model then said that if you changed the phase by adding bases or taking out bases you make a mess of the rest of it. All right? And now putting all of those things together made sense. And so we did a very quick experiment which was formulated on that Saturday afternoon; the experiment was actually done on the Monday. What we realised was there are some proteins which are necessary for the phage. These are these host range mutants. And so what we showed, very quickly, was one: all the existing host range mutants were base analogue revertible. Spontaneous host range mutants were base analogue revertible. And that contrasted with the fact that 90% of the spontaneous mutants in the R2 were not base analogue revertible. And furthermore: we showed we could induce those mutants with base analogues, but we could not induce them with acridines. And we wrote a paper entitled The Theory of Mutagenesis, in which we proposed that there were substitution mutants, for which those were all the ones, and that there were additions and deletions. These were the acridine mutants, producing this drastic effect – and of course now that combined with all these other experiments now made sense.