Logo IMG
HOME > PAST ISSUE > July-August 2005 > Article Detail


Life Cycles

Are there periodic booms and busts in the diversity of life on Earth? Hear a tale of fossils and Fourier transforms

Brian Hayes

A Catalog of Life

The search for patterns in the history of life builds on the labor of generations of paleontologists who went out in the field to dig up fossils. It also owes a debt to one paleontologist who went into the library to dig up thousands of records of fossil discoveries. John Sepkoski began this research while he was still a student of Stephen Jay Gould in the 1970s; by the time of Sepkoski's death in 1999 (at age 50), his database had grown to include more than 36,000 genera of marine organisms. The compendium was published in 2002, both as a weighty tome and as a CD-ROM.

The Sepkoski database has a simple structure. For each genus, he lists the oldest and the youngest geological layers in which at least one member of the genus is reported to appear. For example, the genus Tellinimera carries the notation "K (Camp-l) - T (Dani), signifying that these bivalve molluscs are first observed in the lower substage of the Campanian stage of the Cretaceous period (which is abbreviated K to avoid confusion with the Cambrian and the Carboniferous); the last appearances are in the Danian stage of the Tertiary period. (Thus Tellinimera was one of the lucky survivors of the K-T catastrophe, the extinction that's famous for doing in the dinosaurs.)

Even though the Sepkoski compendium is available on CD-ROM, getting it into a form suitable for further analysis is more than a routine clerical chore. I did some preliminary reformatting with the search-and-replace functions of a text editor, then wrote a small program to do further processing, and finally imported the result into a database manager. What's maddening about such a conversion process is that even tiny typographical inconsistencies in the text—a misplaced hyphen, an extra tab character—can totally derail the operation. Other kinds of errors turn up, too. For example, I found a few dozen entries where Sepkoski apparently recorded the same genus twice. Such minor oversights are hardly a surprise in a document that took decades to compile, and which the author never had a chance to review and revise before publication. In any case, for statistical purposes the database needn't be perfect; random errors might blur a genuine periodic signal, but they are unlikely to generate a spurious one.

The database gives the dates of fossils in terms of geologic periods, epochs, stages and so on; for studies of periodicity, these layers of the stratigraphic column have to be assigned dates and durations in calendar years. As it happens, a new calibration of the geologic sequence, assembled by the International Commission on Stratigraphy, has just been published (in a tome even weightier than the Sepkoski compendium). Based on radio-isotope measurements, paleomagnetism and evidence of astronomical cycles, Geologic Time Scale 2004, or GTS2004, gives dates for strata as far back as the beginning of the Cambrian period—which according to GTS2004 was 542 million years ago.

The Sepkoski compendium mentions almost 300 geologic intervals, to which Muller and Rohde assigned numerical dates based on the new time scale. The task was not entirely straightforward because of changes and variations in nomenclature. For example, Sepkoski refers to a Wolf-campian epoch, which is not recognized in GTS2004; Muller and Rohde defined it as the union of two stages.

comments powered by Disqus


Subscribe to American Scientist