ABUNDANCE CORRELATES WITH EXTINCTION RATE: BIOLOGICAL SIGNAL DECOUPLED FROM SAMPLING EFFECTS IN POST-PALEOZOIC BIVALVE GENERA

We find that post-Paleozoic bivalve genera show a significant negative correlation between extinction rates and local abundance. This result is expected based on macroevolutionary theory and has been empirically shown in biological studies in the Recent. The biological basis for this correlation has been difficult to establish in the fossil record due to sampling artifacts that mimic the expected relationship between abundance and durations. Sampling will artifactually shorten the observed durations of rare genera. If the effect of sampling is strong, it will be manifested in the following ways: (1) short-duration genera will have lower abundance; (2) rare taxa will have shorter observed durations. We observe, contrary to our sampling expectations, no difference in the abundance distributions of short- and long-duration bivalve genera. Furthermore, not all rare genera are restricted to a single 10-million-year temporal bin. A positive correlation exists between genus occurrences and durations. To test whether short-duration genera are the product of poor sampling (i.e., few occurrences) or were truly short-lived, we rarefy the number of occurrences necessary to estimate the observed durations for long-duration genera, and compare these values with the number of occurrences observed for short-duration genera. The number of occurrences needed to accurately estimate long durations falls within the range observed for short-duration genera. These analyses show that a significant proportion of short-duration genera were truly short-lived and that sampling is not responsible for the significant correlation we observe between extinction rates and local abundance in the post-Paleozoic bivalve fossil record.