A few days ago, I wrote about an apparent disagreement between two sets of scientists over the evolution of mammals. I confessed to general confusion as to whether the findings of two different studies actually conflicted with each other. It turns out they did. New Scientist this week contained a short article which nicely summarised the differences:
New Scientist: When did placental and marsupial mammals split?
… According to the fossil record, our ancestors didn't split into modern groups of placental and marsupial mammals until after the dinosaurs bit the dust at the end of the Cretaceous, 65 million years ago. So say John Wible of the Carnegie Museum of Natural History in Pittsburgh, Pennsylvania, and colleagues, who have compared late Cretaceous fossils with modern placental groups…
That bolsters the traditional view of palaeontologists, but flies in the face of molecular studies of genetic divergence of living species, which put the origin of placentals 80 to 140 million years ago… "We're in total discord with the molecular dates," Wible says. He thinks genetic clocks fail to account for the post-Cretaceous burst of mammalian evolution.
Are palaeontologists missing fossils, or do bursts of evolutionary diversification throw off molecular clocks? You have to take both sides seriously, says Rich Cifelli of the Oklahoma Museum of Natural History in Norman.
I have to say, I've always had my doubts about the use of so-called genetic clocks to estimate dates of key evolutionary events. It stands to reason that genetic analyses should be able to give us a very good idea of the sequence in which such events happened, but using them to estimate actual dates for these events seems (to this ill-informed outsider at least) hopeful in the extreme.
The very concept of a genetic clock assumes that genetic mutations occur at a constant rate. This may or may not be the case, but to me it seems a bit too convenient. Physicists use radiocarbon dating and potassium-argon dating to give pretty good estimates of the ages of particular samples (although such techniques are not without their problems), but the biological world is far more messy than the physical one with its precise radioactive half-lives. My gut feeling is that using genetic clocks to provide actual dates for evolutionary events is giving in too much to physics-envy.
For the time-being, I'll side with the palaeontologists, who deal with hard—albeit sparse—physical evidence.
But what the hell do I know? If I turn out to be wrong, I will happily stand corrected.