I mentioned the Dressler experiment of injecting purified protein into a mouse and showing that it became subsequently unable to eliminate the protein immunologically. And that experiment was rapidly, but definitely later, confirmed by Bill Weigle and the Pittsburgh group. So, then I- Dresser had come round to NIMR, National Institute for Medical Research, but he actually had stopped working on tolerance at that stage. I don't quite know what was going on between me and David at that time, because I think I ended up getting a lot of credit which perhaps should've gone to him, but perhaps it shouldn't have gone to him, because I think, he's a moody guy and I think he stopped working on tolerance in part in protest because I started working on it. I don't really know the truth about that. We'd have to ask David. And we wouldn't get a straight answer from him anyway. So, the antigens which David had been using were, and Bill Weigle, were using were foreign immunoglobulins, so I thought I don't want to do exactly the same thing, so I started working with foreign serum albumin. The reason for using these serum proteins was then that was mainly that those were the proteins which you could buy cheaply in bulk. I also worked a little bit with more or less all the other proteins I could lay my hands on, such as lysosime, bacterial lysosime, egg white, ova albumin and so on. And I started in the way that one does, just trying to quantitatively delineate this tolerance, and discovered something which I, at the time, I still think, was pretty interesting. And that was that the amount of protein you needed to give to induce tolerance was pretty much the same for all these different proteins. It was all around one microgram per mouse. With immunoglobulins one microgram into a mouse between one and ten micrograms, was enough to make it tolerant for a long time. But with these other proteins which were eliminated more quickly from the mouse, you had to go on giving them. But that didn't seem to be a very fundamental distinction, so I went on exploring that. So, that discovery was called low-zone tolerance. I called it low-zone tolerance. In contrast to what happened if you gave larger amounts of proteins. If you gave larger amounts of proteins, some of these proteins you can never get tolerance at all, but sometimes, for example, with bovine serum albumin, if you went on injecting it for a time, then the animal would also become tolerant, but you had to give quite a large dosage. It wasn't a matter of doubling or trebling the one microgram, you had to go up tenfold or a hundred-fold, or a thousand-fold. I suppose I wasn't, at the time, very clear what that means. Perhaps I'm still not very clear. But I think that the way to look at that, in retrospect, is that the one microgram doses were tolerising fine, the larger doses were probably turning on the immune response in the same way that the adjuvants had been in David's hands, and then later on in my hands. And, you just needed to give a lot more antigen, for reasons which are not clear, nor I think every will be. There was a lot of argument at the time about what was going on here. Were we looking at what was already suspected, though not securely demonstrated to go on in the thymus, that is the elimination of immunolgically reactive cells or were we generating some kind of suppressor population- and there were people around even at that early time and they became much more numerous in the 1970s, who were in favour of suppression. And I tried to argue against that by showing things like- if you took- and this was actually Roger Taylor's experiment, if you took mice and you made them tolerant with a low dose of protein, then when you look at their recovery, it absolutely depended on the thymus. You had to have a thymus there for them to recover. So we said, well, it can't be suppression, it must be something to do with production of new cells, and that fits the idea of clone elimination. Well, fast forward, if you like, to, what are we, 2004, when Harald von Boehmer has just published a paper showing that if you inject these small doses of protein, not by needle, but by putting in a pump, one of the osmotic pumps which feeds it much slowly, you generate suppression cells very effectively. So, the issue is still open.