But I was interested in looking further at this, getting finer and finer limits on the fine-structure constant. It's a long story; I'll tell only a couple of things. The... the next step was looking at the rhenium decay which is even more sensitive to the fine-structure constant. Rhenium 187 decays to osmium 187 and the decay energy is only 7 kilovolts, so it's a case where the energy of the two nuclei - it's not just the nuclei, it's the atoms, the nuclei including all the bound electrons - are so very, very close to being equal that if you change the Coulomb binding of the nucleus by a tiny amount, you'll shift the balance drastically and so the osmium would become unstable, or the decay rate would be twice as large, either one way or the other. So it meant that the decay rate of this rhenium is even more sensitive indicator of the fine-structure constant than uranium. It turns out it goes with the eighteen thousandth power of the fine-structure constant. And by looking at rhenium and osmium in the meteorites and various other places which have long histories - many of these meteorites, in fact, are 4 billion years old, and so they are very good indicators for the rhenium decay - you can show that the rhenium decay has not varied by more than 10% over 4 billion years, and multiplying that by 18,000, you get a much more precise limit on the fine-structure constant variation, which turns out to be something like, I think, 1 part in 10¹⁵ per year, so way beyond what Teller proposed. Anyhow, the last chapter was opened by Alex Shlyakhter, who unfortunately is dying of cancer as we speak. He's a brilliant young man, and he's... Russian. Was a student in Leningrad about 15 years ago, and he had this brilliant idea that you could get an even more sensitive test of the fine-structure constant variation by looking at the fossil fission reactors in Gabon, in Equatorial Africa. These fission reactors are natural fission reactors which actually were running about 2 billion years ago, when uranium 235 was much more abundant than it is today. In those days, 2 billion years ago, the uranium in the ground had 3.5 percent uranium 235, which is just about what we use nowadays for reactor fuel, so any lump of uranium could be a reactor in those times, and so... that was what actually happened. In several places in that particular ore body, uranium went critical and it was operating as a fission reactor for about a hundred thousand years, and the fission products are still there. And the French have published all this; two books were published describing in great detail what they found. It's a fascinating subject and it turns out that one of the fission products, is samarium 149, is heavily depleted in these fossil reactors. They find only 2% as much samarium 149 as you would in fresh fission products. So something happened to the samarium 149, and what happened was that it was converted by the neutron flux to samarium 150, and you find the samarium 150 instead of 149. But that proves that 2 billion years ago samarium 149 had an enormous fission cross... I mean an enormous capture cross section for slow neutrons.