In a study that is bound to get tons of press, a paper by van Dam et al (published in Nature, I'll add the exact reference later) claims that long-term periodic changes in Earth's oribital tilt "offers a plausible explanation for the characteristic duration of more or less 2.5 million years of the mean species life span in mammals.” They were looking at rodent species, which is odd as naively you'd assume rodents are pretty much impossible to exterminate!
I shall remain skeptical for now. There have been many claims of periodicity in extinction rates over the years, and for pretty much any time scale you can find a similar astronomical timescale, or make one up (e.g. the orbital period of Nemesis, the hypothesized stellar companion of the Sun).
David Raup even wrote a very convincing (but probably wrong) book exploring whether ALL extinctions might be due to astronomical causes.
Correlation does not imply causation, and at present van Dam et al's results are intriguing but not compelling. Much more evidence will be required to prove such a link between orbital variations and species extinction. Still, this will spur more work by people eager to disprove this hypothesis.
Thursday, October 12, 2006