ARE CORRELATES OF EXTINCTION RISK BROADLY APPLICABLE OR CLADE-DEPENDENT? A COMPARATIVE ANALYSIS OF EARLY CENOZOIC BIVALVES

Understanding the factors that contribute to extinction risk is essential for predicting the response of species to environmental change. Yet, identifying biological characteristics that are consistent predictors of extinction risk across the diversity of life is a formidable task. Geographic range size, for example, should always be important, but for a given set of taxa at any point in time other biological characteristics may be more critical. Here, I use species in three ecologically-disparate superfamilies of marine bivalves (the Carditoidea, Pectinoidea, and Veneroidea) to assess whether intrinsic biological correlates of extinction risk were generally applicable or clade-dependent during the early Cenozoic in eastern North America. I compare a model in which each clade differed in its association between biological characteristics and extinction risk with a simpler model describing the dynamics of all bivalve species as a whole. The data used for these analyses were generated through field work in the coastal plain, use of museum collections, and literature compilation. I find that the more complex model that accounts for clade-level differences is much better supported than the clade-independent model. However, geographic range has a strong direct effect on extinction risk in each of the clades, with geographically-widespread species persisting for longer periods of time. Body size also contributes significantly to extinction risk but in opposing directions in the Pectinoidea and Veneroidea, such that it has no net effect on extinction risk for bivalves as a whole; larger size is associated with greater extinction risk in pectinoid species whereas the converse is true among veneroid species. Abundance contributes directly to extinction risk only in the Pectinoidea, in which abundant species are at greater risk. A structural equation model of all bivalve species reveals important indirect effects of both abundance and body size on extinction risk via their positive influence on geographic range size. These results provide additional support for the general role of geographic range in structuring extinction risk over geologic time but also the importance of incorporating clade-level biological characteristics into extinction models.