He started by trying to make histone preparations, again of H3 and H4 and so on and trying to make them on a small scale and trying to aggregate them with small pieces of DNA, which we've got good fibres. But it was... it seemed to be clear that the way the histones were prepared we didn't know if they were native... in a native form or not. So I said to Roger [Kornberg], 'Can you try to look into', because he was a biochemist and he's trained by his father who as I said, was the best biochemist in the world. And Roger was very, very adept at biochemistry, making things. So he began to try to purify the individual histones, that is not just the H3 and H4 together, and the H2 and H2B together. In fact, they do go together as I'll say in a moment. And he... but in the meantime he took lots of X-ray pictures of different fractions interacting with DNA calf thymus. And we could gather all sorts of pics; we could see that on the whole there was some kind of interaction on a 100 angstrom scale that stood up in the X-ray pictures. Now, Wilkins and Luzzati interpreted the 100 angstroms as the pitch of the... as the... somehow the perodicity of a helix because they couldn't think of anything else by helices. And the idea was that the histones, like protamines folded around the DNA, clustered around and made some kind of knobbly structure. So the... but now, so Roger said... I said, we weren't sure about the condition for these histones and we tried different extracting conditions, different salts and all that and also tried to renature them. In the end, he read a paper by a group in South Africa, Van Holt and Van der Westhuizen who were working on sea urchin histones and he noticed that in the paper, they went through a step where the H3 and H4 were together. And the H2A and H2B were together, which were the... but in... in a reasonable form. And he thought that these might be denatured, un-denatured, so he followed it up and he adapted the prep and he made preparation and he made lots of material. And then he began to add them to DNA, both separately and... And together, again using X-ray diffraction because our guide was this 100 angstrom spacing and trying to recover it. It wasn't the best idea but that was all we had at the time, there's no enzymatic activity, the only thing you had was a structural indication. And we knew that if you took X-ray photographs of nuclei as a whole, and people had tried, you tend to get 100 angstrom lines, and also later on you could see 27 angstrom and 55 angstrom lines, that was done later. So he then thought that these might be and since the conditions under which they produce they went... they went on, then... sorry, I've forgotten, Van Holt and Van der Westhuizen went on to separate the histones. And Roger thought that maybe these go together naturally, which had been obvious indications of that. So he began to wonder what these aggregates were and he purified them, he also found that you could under the appropriate salt conditions you could add the H3 and H4 to H2A and H2B and form a really rather tight oligomer. And then he enlisted Jean Thomas, who was working in the biochemistry department, she was an expert protein chemist, which he wasn't. So she... they began to do cross-linking studies and he showed, between the two of them, they showed that H2A, H2B was a dymer, hetrodymer and H3, H4 was a tetramer, which was... so that was the... Now, these were and these looked like naturally existing aggregates. Now, you see, if this is the case, these are quite large and he also showed with Jean that you could put them together to make a histone octamer that is 1 H3 H4 tetramer and 2 H2A H2B dymers making an octamer. And this octamer seemed to be formed under a... and you could associate with DNA, you certainly recreated this 100 angstrom pattern and some other lines. So this was the psychological breakthrough because people had thought that the... that the histones coated the DNA. If they are forming octamers they couldn't possibly do that, they're much too large.