The second crystal we tried to work on was selenium cyanide of vitamin B12, which Lester Smith prepared at our request, which had very beautiful sort of green/red pleochroism, and... but unfortunately the selenium atom, contrary to the expectations of the chemists, was directly attached to the cobalt and lay too close to it on a line of x = nought which meant that it only contributed to part of the structure factors.

However, we got the machine, the old punch card machine, very slowly but in time to calculate a three-dimensional electron density map in which we placed the moved atoms in this part of the structure and then we made it calculate another map with more of the atoms, used in phasing calculation. This was the one that was shown at Paris. Then this one, the chorrin ring itself, appeared and not as a porphyrin but with these two atoms, the ring directly in contact, and while this was happening to the selenium compound it was mostly being worked on by John Robertson. Crystals of the red fragment were got in Cambridge by a young Australian called Cannon, wasn't he?

[Q] Cannon is right.

Yes. And this was obviously the most urgent compound to be examined. We called it the red fragment. It consisted of the whole B12 molecule minus the known part which was the benzimidazole and phosphate intruder group, so that it was a smaller molecule to work on all together. And... but we did the same thing with it and we found the cobalt positions with the Patterson's, calculated three-dimensional maps with just the cobalt phasing and then we placed what atoms we could from this map, which in fact were about 26. Now, at that stage our career was really saved because it would have taken us a long time to calculate the structure factors in three dimensions of the 26 atoms that we thought we knew at that stage and get to the next stage and so on. But at that stage Ken Trueblood walked into the lab...

[Q] Ken Trueblood was from?

... with his mother from UCLA. He had been, when I was visiting America the year or two before, in Pauling's lab and he had calculated a three-dimensional electron density map. Then, with their punched card set up which they had made the year or two before, it was shown off as a particularly beautiful object because they'd coloured the atoms differently according to their chemical nature, oxygen, nitrogen, carbon and so on, and of course they'd done the drawing particularly well. But by this time he'd given up the idea of using punched cards and the methods that had been useful for threonine and made friends with a machine, a new electronic machine which was in the hands of the Department of Agriculture. It was called SWAC, which SWAC stood for South Western Agricultural Computer, and he said if we would like it, he would [unclear] the programmes necessary and he would put some of our calculations on it free and gratis and for nothing and he would just do them and send them to us to look at and that seemed too good an offer to miss.

Actually, we were slightly cagey. We hadn't the B12 calculations quite ready to give over but we had what should have been the final map for calciferol ready and so we gave that to him to do. When he got that perfectly right we thought he was a reliable person to give the B12 to. He had a very good group of young men who were his computing assistants with... whom one of them, Bob Sparks, has gone on being a sort of a major influence in computing and making diffractometers for X-ray analysis ever since. But another of them, whom I was rather fond of, was Dick Prosen, died - I don't really remember what from - towards the end of the analysis.

Anyway, then began a very exciting period in which we would send him a list of the co-ordinates of the atoms that we wanted calculated on, and he would send us back the list of structure factors and then the three-dimensional map. And of course these would just come out on computer tapes so they were not transparent but, but just...

[Q] Sheets of paper?

Yes. And Jenny transferred these to tracing paper and drew them to scale so that they should fit one over the other in this sequence and we had the first of the red fragment three-dimensional maps showing the corrin nucleus absolutely clearly.