We began to make two kinds of libraries, of archives of zinc fingers. At the beginning, we did two times three, putting in various extender linkers and so on. And they bound with picomolar affinity. But what we found was they didn't discriminate so well, if you had a target of... a target of eighteen base pairs, if you change one-base pairs, it didn't discriminate so well because you had to have even higher sensitivity. Now, I had thought at one time that we should put in, just link individual fingers and have spaces between them. That seems a lot of work, but as I said a moment ago, the... we were making two-finger constructs anyway because of the way of getting in the cross-strand interaction. So I suggested that we ought to distribute the strain, and Yen Choo said, 'Why don't we just combine three lots of two fingers to produce six fingers?' And so, we did that and Michael Moore developed the correct linkers for that and that's when I talked about gylcine serine, things of that sort. And it turned out that these had a great advantage. They bound with picomolar affinity but they were much more sensitive to any mutations or omissions in the DNA sequence and so that, if you... if you introduce a single mutation they loose anything between ten to 100 fold... loss of affinity and sometimes 200 fold. So... so we decided in the end that, and Sangamo's libraries are mostly made now, combining sets of two fingers. So we tend to make twos, fours, sixes and even eights. We did make some nines here, that is three times three, but they were just laboratory exercises and Sangamo, not Sangamo, Gendaq made a few of those, but they didn't really give much gain. You see, because once you got 18 base pairs, it's pretty well unique in the human genome, six base, nine base pairs would occur several hundred thousand times, 18 base pairs would occur an only a very small number of times, in the genomes which is random, as you can calculate that. But the... of course genomes aren't random so you've got to have some reservations about that.