ASSESSING BIASES IN THE MAMMALIAN FOSSIL RECORD USING LATE PLEISTOCENE MAMMALS FROM NORTH AMERICA

Recent analyses of marine bivalves indicate that taxa that fail to become incorporated into the fossil record are a non-random sample of the available taxa (Valentine et al. 2006). They tend to have small body sizes and narrow geographic ranges among other characteristics. We performed a similar analysis for mammals using the late Pleistocene record for North America. We analyzed whether present day mammal species missing from the late Pleistocene record were biased with respect to body size, range size, diet or habitat preferences. Of the 293 non-volant species of mammals that presently occur north of Mexico, 222 (68 percent) were recorded in the fossil record over the last 40,000 years. Species that fail to make it into the record tend to be significantly smaller bodied (mean body size = 80.17 g vs. 194.98 g) and have significantly smaller geographic ranges (mean range size = 9.89 vs. 22.214 degrees latitude). Although a large proportion of both missing and represented species were herbivores and granivores, carnivores were more likely to make it into the mammalian fossil record. Mammalian carnivores tend to be larger bodied making them more likely to survive fossilization and to subsequently be discovered. These results suggest that ecological analyses of fossil data should be designed to take these biases into account. However, methods developed for marine systems likely will be useful for terrestrial systems and vice versa. In general, the mammalian record is biased in similar ways to the marine bivalve record despite the differences in depositional environments. This similarity suggests that this common sense prediction, that small-bodied, less robust taxa and narrow-ranging, rare taxa should be less likely to make it into the fossil record, is a general property of the fossil record and not unique to particular groups.