A palaeontologist ponders how biodiversity is spread across the vertebrate tree of life.

Why do some biological groups burst at the seams with many different species, whereas others, despite their deep evolutionary heritage, contain only a handful of members? Many of my old vertebrate-biology textbooks are rife with qualitative scenarios, peddled with surprising degrees of confidence, that explain how species-rich branches can chalk up their success to key evolutionary 'innovations' and how less-diverse ones haven't kept up with changing conditions. What you won't find are details of how these exceptional groups might be identified in the first place.

Michael Alfaro of the University of California, Los Angeles, and his colleagues have now quantified this black art (Proc. Natl Acad. Sci. USA 106, 13410–13414; 2009). They marry statistically explicit models with fossil-calibrated evolutionary trees and counts of living species to ask a basic, but surprisingly unanswered question: precisely which branches of the vertebrate family tree are more or less species-rich than expected given their age?

The authors identify nine groups that show substantial changes from the background tempo of vertebrate evolution: 'living fossils' such as lungfishes are characterized by lower-than-predicted diversity, whereas other branches, such as the perch-like fishes and a subset of mammals, contain vastly more species than expected.

As a palaeontologist, I am intrigued that three of the exceptionally diverse radiations are thought (although not without controversy) to have proliferated following the mass extinction that killed off the dinosaurs, hinting at the far-reaching consequences of this event in structuring the modern vertebrate fauna. Most importantly, these authors establish a clear quantitative framework that can be used to test all those textbook stories. I'm confident that in a few years, my students will learn a much more nuanced picture of vertebrate diversification than I ever did, one that will trace its own roots back to studies such as this.

Discuss this paper at http://blogs.nature.com/nature/journalclub