2003 Seattle Annual Meeting (November 2–5, 2003)

Paper No. 16
Presentation Time: 11:45 AM

EVOLUTIONARY INERTIA DISCOVERED IN PHANEROZOIC ORIGINATION AND EXTINCTION RATES


LU, Peter, Department of Physics, Harvard Univ, 17 Oxford Street, Cambridge, MA 02138, YOGO, Motohiro, Department of Economics, Harvard Univ, Cambridge, MA 02138 and MARSHALL, Charles R., Department of Invertebrate Paleontology, Museum of Comparative Zoology, Harvard Univ, 26 Oxford Street, Cambridge, MA 02138, plu@fas.harvard.edu

The dynamics of evolution may be approached through quantitative analysis of the fossil record. Here we apply vector autoregression analysis (often used to analyze time-series data in the stock market) to Sepkoski's Phanerozoic compendium of fossil genera. In this analysis we simultaneously calculate autocorrelations within, and cross-correlations between, per-genus origination and extinction rates in successive substages. We find: 1) A strong autocorrelation in the origination rates, and in the extinction rates: the rate of origination (or extinction) in one substage is usually followed by a similar rate in the next. These trends imply the presence of an "evolutionary inertia" in origination and extinction, and suggest that both may be governed by similar forces. We speculate that these forces may result from co-evolutionary dynamics driven by species originations and extinctions, rather than from short-term changes in the external environment. 2) We did not find significant cross-correlations between the origination and extinction rates; the rate of extinction in one substage has little effect on the rate of origination in the following substage, and vice-versa. This indicates that pseudo-extinction does not dominate Sepkoski's compendium. 3) The lack of significant cross-correlation remained when we extended our analysis to include data from the two substages prior to the substage of interest, undermining the significance of the proposed longer-term dynamic of "evolutionary speed limits," an idea partly based on a cross-correlation between origination rates and the extinction rate two substages previous. 4) We were able to explicitly take into account uneven interval lengths (unlike in Fourier analysis where uneven interval durations can complicate interpretation), but our results were unchanged in these analyses.