Foote, Michael. 2003. Origination and extinction through the Phanerozoic: A new approach. Journal of Geology. Vol: 111. Pages 125-148. (Abstract)
Temporal patterns of origination and extinction are essential components of many paleontological studies, but it has been difficult to obtain accurate rate estimates because the observed record of first and last appearances is distorted by the incompleteness of the fossil record. Here I analyze observed first and last appearances of marine animal and microfossil genera in a way that explicitly takes incompleteness and its variation into consideration. This approach allows estimates of true rates of origination and extinction throughout the Phanerozoic. Substantial support is provided for the proposition that most rate peaks in the raw data are real in the sense that they do not arise as a consequence of temporal variability in the overall quality of the fossil record. Even though the existence of rate anomalies is supported, their timing is nevertheless open to question in many cases. If one assumes that rates of origination and extinction are constant through a given stratigraphic interval, then peaks in revised origination rates tend to be displaced backward and extinction peaks forward relative to the peaks in the raw data. If, however, one assumes a model of pulsed turnover, with true originations concentrated at lower interval boundaries and true extinctions concentrated at upper interval boundaries, the apparent timing of extinction peaks is largely reliable at face value. Thus, whereas rate anomalies may well be real, precisely when they occurred is a question that cannot be answered definitively without independent support for a model of smooth versus pulsed rate variation. The pattern of extinction, particularly the major events, is more faithfully represented in the fossil record than that of origination. There is a tendency for the major extinction events to occur during stages in which the quality of the record is relatively high and for recoveries from extinctions to occur when the record is less complete. These results imply that interpretations of origination and extinction history that depend only on the existence of rate anomalies are fairly robust, whereas interpretations of the timing of events and the temporal covariation between origination and extinction may require substantial revision.