Far out, have you heard about this? The Oklo region in Gabon, Africa contains an ancient uranium deposit that used to be a naturally occurring nuclear fission reactor a couple billion years ago. Scientists know this by analyzing samples of the isotopes mined from the deposit. There was in fact an article recently in Physics Today, a technical magazine of the American Institute of Physics, by some researchers who study Oklo. Their interest is in comparing the decay rate of uranium billions of years ago with the decay rate measured today. Theoretically that rate should be exactly the same over the eons, being a physical constant. But the focus of this article was "hey, here's a chance to check this constant over billions of years - could it drift very slightly in that time?". They estimate the few-billion-years-ago decay rate by analyzing ratios of different isotopes of the uranium left in the deposit (for example U238 vs U235 etc). Since it's an estimation, there's some amount of uncertainty on the ancient decay rate they measure, a "plus or minus" to the value. So the hard part is to try to refine their measurement techniques and hone that uncertainty down small enough to see if a tiny difference in decay rates can be resolved outside the uncertainty. My understanding from the article was it's still inconclusive, ie not fine enough honing yet to notice a difference in the decay rate if there is one. So far then the constant's still a constant - but a pretty clever idea for an experiment I thought.

You may have already seen that the constancy of the decay rates of radioactive elements is actually a popular issue with some religious conspiracy-theorists who take literally the Biblical age of the Earth as only a few thousand years old. Radiometric dating measures the age of old rocks using information about the decay of radioactive elements in them, and the technique assumes that these decay rates are constant over geologic time. But some of these folks try to (baselessly) push the notion that decay rates have not been constant, and in this way allow for a period of fast decay in the past, which would in turn allow rocks "seemingly" a billion years old to only be a thousand years old. Others try to say that radiometric dating is not accurate enough to tell a thousand years from a billion years old. However, this would mean that radiometric dating techniques would have to be off by a factor of a million, when in fact these techniques are known to be consistently accurate to within a percent or so (ie a factor of .01, not 1000000). There's a lot of well-tested theory, and over a hundred years of many many measurements of the decay rates of radioactive elements, behind the constancy of those decay rates. That's where the Oklo thing above comes in, another test of the decay constancy. In any case, the Oklo researchers are looking for the possibility of a really small effect, perhaps a percent or less drift in the rate, which is why the one percent(ish) uncertainty in the measurement is such a big hurdle. And yet, as far as radiometric dating is concerned, one percent would correspond to a difference of a rock being 100 million years old vs 101 million years old, not typically an issue when dating something so old.

Similarly, over the years experimentalists have taken every chance to further test General Relativity to greater accuracy in meticulously careful experiments (and so far those test results have all matched the G.R. theory). As usual there are those out there who try to cite conspiracy and say Einstein was wrong, though for reasons less clear than for the creationists mentioned above. Speaking of relativistic particles, what would've been *really* cool is to have seen the "Cherenkov radiation" in the water naturally surrounding this natural reactor way back when it was going. (A little difficult I suppose, since it was deeply buried - oh yeah, and no people even existed yet!) Cherenkov radiation is the blue glowing light seen in the water around reactors, from neutrons coming out into the water at speeds greater than the speed that light travels in the water, and the result is that cool blue light. It's essentially the light equivalent of a sonic boom. Far out, huh?

Finally, you'll also sometimes come across references to the planned Yucca Mountain nuclear waste repository when reading about these prehistoric natural reactors, saying that "see, they were contained fine after all this time, so why worry about Yucca Mtn". But of course that's in a completely different location, with different geology and different water table environment, a different concentration of waste, and so on. I suppose it's initially encouraging though...

There are some great references and links on the subject of Oklo's natural reactor at Wikipedia's Oklo reactor page.