2002 Denver Annual Meeting (October 27-30, 2002)

Paper No. 10
Presentation Time: 10:30 AM

SUDDEN, GRADUAL, OR STEPWISE? EXPANDING THE NULL HYPOTHESIS IN STATISTICAL TESTS OF MASS EXTINCTION


WANG, Steve C., Department of Statistics, Harvard Univ, 1 Oxford St, Cambridge, MA 02138, scwang@stat.harvard.edu

Due to the Signor-Lipps effect, a simultaneous extinction of a collection of taxa may appear gradual in the fossil record. Several methods have been proposed that account for the incompleteness of the fossil record when testing whether a collection of taxa went extinct simultaneously. However, we may be interested in testing other patterns of extinction in addition to a simultaneous extinction of all taxa. For instance, we may hypothesize that an extinction occurred gradually, or in several pulses. Here we introduce a statistical method for testing whether any arbitrary extinction scenario is consistent with observed last occurrences in the fossil record. Using this method, we can determine a confidence region for the “shape” of a hypothesized extinction event.