So then we began looking at electron microscope pictures of various virus particles... plant viruses. Now, we knew what we were looking for but we didn't know... we didn't understand what we were looking at. And I came to realise, what we were looking at was the... was the image projected in... projected in, from three dimensions to two dimensions. Now, people never talked about this. And I remember a man in Oxford, the reader in something or other, a well-known biologist saying to... you know, that they couldn't understand complicated electron micrographs. And I suggested that we actually tilt the specimens... and that came about, tilting the specimen, it was originally, I don't know if you remember, John, I got hold of a... of a... metallurgist stage in electron, which allowed you to tilt the specimens because I thought we should be able to take stereos. And because I can't see with the one eye, I asked people to see what they could see. And nobody could see any stereographic images. And I realised we... there wasn't a stereographic image because that depends upon surface features. And what we were doing was, they were projection, and some way or other the penny dropped and realised that we were looking at a projection. And so the...
[Q] And not just a... a footprint.
Not just a footprint like X-ray. And so it was quite easy to work out the theory of that. And in fact, taking a series of tilts at different angles is a parallel like collecting X-ray data in a series of two dimensional projections using a camera. I don't think that was explicit in my mind, but really, it was... but before we did this we had been already building models of viruses to explain the pictures purely because we had the theory of how the shells were built and we could explain, for example, as shown there, shown on the slide how you could get very different appearances, very different appearances if you put, say, 180 units into 60 triangles, shown at the top left. And depending on how you cluster the units, we get them clustering in groups of... three, groups of three... that would look like the top right. If you cluster them in groups of six and five they look like that and then you would have on the bottom right clusters of groups of two.
