Estimating extinction with the fossil record

Many ecological and palaeontological studies focus on extinction. The fossil record is particularly important for studying long-term patterns in extinction: although analyses of extant phylogenies can estimate extinction rates (e.g. Alfaro et al. 2009) and even suggest mass extinctions (e.g. Crisp & Cook 2009), they cannot imply trilobites ever existed or that sphenodonts (now represented only by the tuatara) were once as diverse as lepidosaurs (lizards and snakes). However, workers also use the fossil record to test ideas about the pace of major extinction events that use methods similar to those that conservation biology might use. Here we will review current palaeobiological methods for inferring extinction patterns, spanning ‘traditional’ methods using stratigraphic ranges to methods using more exact information about distributions of finds within stratigraphic ranges. Palaeontological studies of diversity patterns also focus extensively on standing richness and origination rates. We do not focus on either of these parameters for their own sake. However, standing richness (usually referred to simply as ‘diversity’ in palaeontological studies) is an important parameter when discussing extinction as extinction metrics necessarily rely on changes in standing richness. Origination rates also can be an important parameter as the distribution of originations within time spans (i.e. evenly throughout the interval vs concentrated in the beginning) have small but important effects on the predictions of extinction rate hypotheses. Moreover, palaeontological techniques for estimating origination rates are essentially identical to those for estimating extinction, save that they are done ‘in reverse’. Thus, discussions of how one estimates extinction rates using fossil data can almost double for discussion of how one estimates origination from fossil data. The other parameter that will be important throughout this review is preservation (i.e. sampling) rate. Just as ecologists know that incomplete sampling affects sampled richness (e.g. Hurlbert 1971) and implied extinction (e.g. Solow 1993a), palaeontologists know that incomplete sampling affects the first and last appearances of fossils, both in local sedimentary sections (Signor & Lipps 1982) and globally (Sepkoski 1975). Workers have expressed concern that major extinctions are exaggerated by or possibly even illusions of intervals of poor preservation (Raup 1979; Smith et al. 2001; Peters & Foote 2002). Because sampling intensity is of interest to both fields, we shall also discuss how palaeontological studies address sampling when estimating diversity and diversity dynamics.