So what we did was to make large quantities of protein, change the salt conditions, change the temperature and look at the aggregates, either by hydrodynamics using ultra centrifuge or by... and by electron microscopy and so on. And so we produced a phase diagram, the first time anybody had ever done it for a protein, but a very common physical chemistry of small... small molecules. We plotted the different states of aggregation of the TMV protein as a function of pH as a function of temperature. And lo and behold it told us something that had already been observed, been observed by others, that if you put the TMV protein, the so called A protein, the... the proposed or supposed trimer at low pH, low acid conditions, it formed... it formed a helix without any RNA. So the protein determined the protein... a helical array of proteins was determined purely by the properties of the protein; this was only a low pH. However, if you went to high alkali, then, of course, the thing disintegrated and... sorry, made only small aggregates of about roughly the range of a trimer. But in-between these two positions, at about a normal temperature, 20º and about a pH of 7, which is normal, what we saw were ringed shaped objects and they were the discs. The very discs that Reuben had crystallised, now this seemed to... And also, the salt condition was about .15 molar which is normal physiological condition. So I realised then that the disc was an object not adventitious, it was actually... the protein on its own at normal conditions made the disc and didn't make the virus helix. And I realised that it would be... that... it would be... I think these are Don Caspar's words, soon afterwards, it would be... but I realised... they were... it would be a disaster if the protein on its own formed the helix without the RNA. Because what he wanted to do was combine with the RNA and put the RNA into the helical array... and being biological, so it could. So that... so therefore... but the disc which is not... which is a number of rings in the disc was 17, the number of turns in the virus helix at 16 and one third, so it wouldn't require much of a change to start with the disc which would then dislocate and then start the growth of the virus just in the way that we built the Brussels model. And so now what I can't remember is whether I harked back consciously to the Brussels model but I... I must have realised at some point that was what I was just reliving that.