2008 Joint Meeting of The Geological Society of America, Soil Science Society of America, American Society of Agronomy, Crop Science Society of America, Gulf Coast Association of Geological Societies with the Gulf Coast Section of SEPM

Paper No. 1
Presentation Time: 8:00 AM

Phanerozoic Marine Diversity of Families, Genera, and Species: Did Jack Get It Right the First Time?

ALROY, John, Paleobiology Database, NCEAS, University of California, 735 State Street, Suite 300, Santa Barbara, CA 93101, alroy@nceas.ucsb.edu

The diversity of higher-level taxa such as families might be a good measure of low-level diversity because families are easier to sample, or a bad measure if these data obscure meaningful variation. Sepkoski believed the former. Raw sampled-in-bin (SIB) counts of genera and species in the Paleobiology Database are much more volatile than family counts, even suggesting an exponential post-Paleozoic radiation. Sampling standardization removes much of this volatility and brings all three data sets into close agreement in terms of both overall trajectories and changes between temporal bins. Because the family-level curve changes little, it is a reasonable proxy for all the standardized curves. Nonetheless, a very different pattern is seen when family counts are based on age ranges instead of SIB. There is a much steeper Cambro-Ordovician radiation, a plateau spanning much of the Paleozoic instead of a Devonian through early Permian trough, and an accelerated mid-Mesozoic radiation. Sepkoski's 1984 range-through, family-level curve is quite similar. That said, it is undersampled in the Triassic and now also presents a strong Cretaceous-Cenozoic increase that can be replicated by including collections preserving original aragonite or coming from unlithified sediments. Both his 1997 genus-level data and the new genus-level data show an even more divergent range-through pattern, with a steeper mid-Paleozoic drop and relatively low Mesozoic levels that amplify the artifactual Cretaceous-Cenozoic increase. Therefore, counting methods and sampling methods both matter: the most reliable signal can only be recovered by using SIB counts and either standardizing or working at the family level. Ironically, Sepkoski himself arrived at essentially the same pattern by omitting single-interval genera and truncating age ranges of extant ones at last fossil occurrences. We no longer need such back-of-the-envelope methods, but his conclusion that the fossil record provides "a consistent picture of the history of biodiversity" still stands.