IOWA CITY, Iowa -- A University of Iowa researcher and his colleague have confirmed that researchers studying the evolution of ancient marine life, while overlooking much relevant data, have the record straight.
That's important, according to Jonathan Adrain, assistant professor in the UI department of geoscience, and Stephen Westrop of the University of Oklahoma, because such data provide the underpinnings for many scientific theories, including some that examine the mass extinction of the dinosaurs.
In a paper published in the July 7 issue of Science, Adrain and Westrop note that paleobiologists record data on the number of organisms living in specific geological time frames by reading published research papers, rather than by making direct field observations. This practice has raised a question about the accuracy of such "library-based" research. The two authors say that they were able to answer that question by comparing their own database, based on expert taxonomic knowledge of fossil trilobites (extinct marine arthropods), to an equivalent portion of a well-respected, library-based database (the late J.John Sepkoski's widely used global genus database).
They found that upwards of 70 percent of the entries in the global database are inaccurate, however the error is randomly distributed across geological time frames and doesn't introduce bias into figures on the comparative numbers of organisms that lived in one epoch versus another.
"Basically, what we found was that, yes, the fears were true -- there's a whole bunch of error in the taxic paleobiologists' databases, but, no, the error doesn't seem to matter very much," Adrain says. "The overall patterns that the error-filled Sepkoski data yield are nearly identical to those that our own data yield. Why? Because the error is randomly distributed -- so even a ton of it will not bias the underlying pattern; the error would have to be driven or skewed in a particular direction to really w
'"/>
Contact: Gary Galluzzo
gary-galluzzo@uiowa.edu
319-384-0009
University of Iowa
6-Jul-2000