Bambach et al. take a new look at Sepkoski's data base, using the version which tried to compile data at the generic level rather than the family level. In the book I used Sepkoski's own analysis based on the family-level version, which is more stable. Now it is 20 years later, and this new analysis may give us more insight.

Bambach et al. have looked at the data in a slightly different way. First, they place most emphasis on the diversity of life at the end of a time interval, rather than the total diversity that comes from adding together all the fossils recorded during the length of an interval. Obviously, that removes some bias that comes from having unequal lengths of intervals.

Second, in their new analysis Bambach et al. emphasize the balance between origination and extinction. A massive extinction may be masked if it is accompanied by a massive replacement of the newly extinct genera; and a small extinction may look unrealistically large if the rate of origination is abnormally low. Obviously, all sudden changes in diversity need explanation; but we and they are concerned here with the question of what constitutes a "mass extinction" and what caused each one.

Third, Bambach et al. worry that the rates of origination and extinction have not been uniform throughout the 500+ million years of the Phanerozoic. Rates in the Mesozoic, say, are not the same as in the Cambrian, and a fair analysis would consider Mesozoic events within the context of the biological world of the Mesozoic.

With these changes in analysis, it looks as if there are only three "mass extinctions": the end-Ordovician, the end-Permian, and the end-Cretaceous. At these times, origination rates were "normal", but extinctions were overwhelmingly severe, resulting in a massive drop in diversity, a "mass extinction". In contrast, at the end-Devonian and the end-Triassic, extinction rates were higher than normal, and origination rates were significantly low: the result was a severe drop in diversity, but not for the same reason or (likely) from the same cause as in the "Big THree" mass extinctions.

As one might have expected (but no-one did), the Cambrian is a time when diversity is difficult to analyse. Stratigraphy is difficult so far back in time, fossil preservation is poor except in a few lagerstätten, and fewer paleontologists study these older rocks. Diversity was low (compared with later periods), so fluctuations in origination rates and extinction rates tended to yield much larger relative swings in apparent diversity than at later times. Was the biology and ecology different, or is all this the result of preservational bias? Cambrian life and Cambrian ecology were likely different, and we should not (yet) worry too much about the rather anomalous numbers that come out of the Cambrian data. Most likely, this period will require some separate consideration because Cambrian life was so different.

The end-Devonian extinction probably relates to the extinction of reef faunas world-wide, and although that needs more research and explanation, it does not constitute a global "mass extinction".

Overall, then, we are ahead as the result of this new appraisal. Bambach et al. "tweaked" the data more than Sepkoski did, but they lay out their reasoning and their methodology, and their results look sound. It is a convincing study. I have not yet marked the Sepkoski papers as "superseded", but I likely will by the time the next edition rolls around (if that day ever comes!).

First drafted, January 18, 2005

Return to Notes for Chapter 6