Measuring and Comparing Extinction Events: Reconsidering Diversity Crises and Concepts.

To understand our present diversity crisis, it is natural to look to past crises for parallels and indicators. This is difficult because the present crisis is unlike the "Big Five" of the past: it is mostly terrestrial (with an increasing marine component), involves widespread habitat destruction and alteration of climate, and is largely anthropogenic, with confounding effects of differences in loss of diversity among continents and the difficulty of separating anthropogenic extinctions from natural Pleistocene and post-Pleistocene extinctions. In contrast, the "Big Five" crises of the geologic record are mainly marine (in the first two, no land vertebrates existed), and because marine taxa outnumber terrestrial taxa by a margin of about 25:1, global analyses of diversity crises have tended to lump together all phyla and environments. As a result, terrestrial evidence has been "swamped" statistically by the marine data. Both synchroneity and causality of terrestrial and marine events have usually been assumed, but without decisive data. Terrestrial vertebrate faunas do not seem to have been suddenly and catastrophically affected at the ends of the Permian, the Triassic, and the Cretaceous; rather, the pattern generally seems to be of steady turnover and replacement of groups and sometimes of slow decline. Here I suggest a revision of the concept of "mass extinction," which has no definitional limits on the application of the term with respect to duration, geography, ecology, or taxa affected. Unusual drops in taxonomic diversity have traditionally focused on increases in extinction rates, with scarce consideration of origination rates and their interplay with extinction rates. Analyses of hypothesized diversity crises should be operationally and situationally defined and statistically normalized through the histories of taxa and biotas, and should explicitly include both origination and extinction rates. The term "mass extinctions" would be usefully replaced by "diversity crises." These parameters require not absolute numerical (or percentage) limits but situational ones.