Periodicity of Extinctions in the Geologic Past
In 1984, palaeontologists David Raup and Jack Sepkoski published a foundational statistical study examining whether major extinction events had occurred randomly or at regular intervals during the geologic past.
The researchers analysed fossil-family extinction intensity records for marine vertebrates, invertebrates and protozoans covering approximately the previous 250 million years. They identified 12 major extinction events within the dataset.
Their analysis indicated that these events displayed a statistically significant periodicity, with a mean interval of approximately 26 million years. The reported probability of obtaining this pattern by chance was less than 1 per cent, or P < 0.01.
Two of the extinction events included in the sequence coincided with events that had previously been associated with meteorite or asteroid impacts. However, the study did not establish a single physical mechanism responsible for the proposed cycle.
Why this study matters
The paper was one of the first widely discussed attempts to identify a regular long-term cycle in the fossil extinction record. It stimulated decades of research into extinction periodicity, impact events, astronomical influences, geological processes, fossil-record dating and the statistical methods used to identify cycles in unevenly spaced palaeontological data.
The proposed 26-million-year interval should therefore be treated as a statistical result reported from the available fossil data, rather than as proof that extinctions must occur according to a fixed timetable.
Questions for discussion
-
How reliable were the fossil-family extinction dates available in 1984?
-
Could uncertainties in geological dating create, weaken or shift an apparent periodicity?
-
Is the reported 26-million-year pattern still detectable when newer extinction databases and timescales are used?
-
Could the pattern result from the way extinction events were selected or grouped?
-
What natural mechanism could plausibly operate on a cycle of approximately 26 million years?
-
Do impact events, volcanic episodes, sea-level changes or astronomical motions provide convincing explanations?
-
How should statistical significance be interpreted when researchers test many possible cycle lengths?
Study details
Authors: David M. Raup and J. John Sepkoski Jr.
Publication: Proceedings of the National Academy of Sciences
Year: 1984
Proposed period: Approximately 26 million years
Dataset: Marine fossil-family extinction intensity records
Reported significance: P < 0.01
Source: https://pmc.ncbi.nlm.nih.gov/articles/PMC344925/