All the action in gastrin is in the tetra, the terminal... four amino acids – Trp-Met-Asp-Phe – and it's amidated. And, so when we started here in '88, that's what we knew, and the question was: how can we go with that knowledge to make an antagonist? Well, we... in our initial hiring of people here, a lot of it was done through Imperial College and Barrett Collingen who had replaced me came as part of that, but we also took on a modeller who was exceptionally bright. And, anyway, he found in the literature that the infrared structure... pattern, for tetragastrin could only happen if the two aromatic rings of the indol of the Trp and the phenyl were about 4-5Å apart. Now, what Biola did was to show that they could only be in that position if that little peptide was wound into a tight helix, a 3₁₀-helix, it was called. And so, the chemist started off then to say, 'Can we get two aromatics held together', and we knew that it had dipolar structure in a dipole, and again, a bit like the first histamine story, we went on... we got leads right from the beginning in peptides. We could make peptide analogues, but all of them had partial agonist activity which, this appears on all of these programmes, so that wasn't what we were looking for. And it took us, really, a long time – about five years – before... and what happened, I mentioned Barrett Collingen who was our Head of Chemistry during this time, and he suddenly remembered a... an experiment he had done as an undergraduate at University College.