EXAMINING THE LATITUDINAL DIVERSITY GRADIENT IN PALEOZOIC TEREBRATULIDE BRACHIOPODS : SHOULD SINGLETON DATA BE REMOVED?

Studies of taxonomic diversity over time count and compare FADs and LADs over a succession of temporal intervals, and interpret them with respect to taxon origination and extinction. Singleton taxa, that first appear and last appear in the same temporal interval, are often removed from analyses because they might result from preservational biases rather than evolutionary processes, or they might represent non-independent FADs and LADs. Should singleton taxa always be excluded? We argue that in the case of Paleozoic terebratulide brachiopods, singletons should be included because 1) they are temporally distributed in proportion to non-singletons, and 2) they occur in proportion to the ratio of singleton to non-singleton taxa in localities known to be very completely preserved. In this case, singletons may reflect rapid evolutionary turnover of taxa, rather than preservational bias, and methods that accommodate singleton taxa should be used to study their diversity.

Singletons can be critical in testing diversity hypotheses in the study of latitudinal diversity gradients. Contrary to the anti-tropical diversity pattern in modern articulated brachiopods, Paleozoic terebratulides show a tropically-biased latitudinal diversity gradient. Testing the hypothesis that the tropics are either a diversity source or sink relies on comparing FAD and LAD latitudes. In singletons, these latitudes are derived from the same data points and should be removed. However, taxon age distributions are more suitable to singleton data as longevity considers origination and extinction simultaneously. We used taxon age distributions to test the hypothesis that the observed Paleozoic diversity gradient is the result of latitudinal bias in generic turnover rate. Temporal (stage-level) and biogeographic data (in 10 degree latitudinal bands) for all Paleozoic terebratulide brachiopod genera were gathered from literature sources, supplemented by examination of NMNH and YPM collections. Paleolatitudes were inferred from the Scotese Paleogeographic Map Archive. Regression analysis indicates a significant correlation between turnover rate and latitude. Proportionally more singleton genera occur in the tropics. Although singletons are randomly distributed in time, they are not randomly distributed over latitude.