Jablonski tests morphological data against molecular data

During a seminar at another institution several years ago, University paleontologist David Jablonski fielded a hostile question: Why bother classifying organisms according to their physical appearance, let alone analyze their evolutionary dynamics, when molecular techniques had already invalidated that approach?

With more than a few heads in the audience nodding their agreement, Jablonski, the William Kenan Jr. Professor in Geophysical Sciences and the College, saw more work to be done.

The question launched him into a rigorous study that has culminated in a new approach to reconciling the supposed conflict between fossil and molecular data in evolutionary studies.

For more than two decades, debate has waxed and waned between biologists and paleontologists about the reliability of their different methods. Until now, attention has focused on the dramatically different evolutionary history of certain lineages as determined by fossils or by genetics.

Scientists using molecular techniques assert that the field of genetics more accurately determines evolutionary relationships than does a comparison of the kinds of physical characteristics—the morphology—preserved in fossils. But how inaccurate, really, were those aspects of form used to infer evolution? Jablonski and the University of Michigan’s John Finarelli have published the first quantitative assessment of these assumed discrepancies in the Proceedings of the National Academy of Sciences, Online Early Edition of April 27-May 1.

They compared the molecular data to data based on the kinds of features used to distinguish fossil lineages for 228 mammal and 197 mollusk lineages at the genus level (both wolves and dogs belong to the genus Canis, for example).

No matter how they looked at it, the lineages defined by their morphological features “showed an imperfect but very good fit to the molecular data,” Jablonski said. The fits were generally far better than random. The few exceptions included freshwater clams, “a complete disaster,” he said.

Jablonski and Finarelli (Ph.D.,’07) then decided to push their luck. It was all very well to have a “good” fit between morphology and molecules, but was it good enough to ask evolutionary questions rigorously? They looked at the fits again, but this time focused on geographic range and body size—two key features for many questions in evolution, ecology and conservation biology. The result: a “spectacularly robust” match between the morphological and molecular data.

Jablonski interprets the results as good news for evolutionary studies. The work backs up a huge range of analyses among living and fossil animals, from trends in increasing body size in mammal lineages, to the dramatic ups and downs of diversity reported in the fossil record of evolutionary bursts and mass extinctions.

“Our study also points the way toward new partnerships with molecular biology, as we straighten out the mismatches that we did find,” he said.