Since the 19th century the remains of Quaternary mammals were an important source of data for reconstructing past environmental conditions. I tested two basic assumptions that underlie Quaternary vertebrate paleoecology. The first assumption is that fossils mammals can be identified reliably to species. The second assumption is that correlations established between extant mammals and environmental parameters can be used to interpret reliably the paleoenvironment from the latest Pleistocene. Incorrect specimen identifications could lead to errors in paleoecologic interpretations. I explicitly tested an alternative to the traditional approach to identification by identifying fossil shrews based on apomorphies. My results indicated that some traditional characters are useful for identification, but only complete specimens with a combination of characters can be identified to species. This indicates that previous authors who identified shrews to species did not compare them to the full diversity of species. I tested the reliability of cenograms and species-richness models as approaches for the reconstruction of environmental conditions in the past. I used faunal data from Hall’s Cave, Kerr County, Texas to construct cenograms and species-richness models and compared the results to independent paleoclimate proxies. Neither species-richness models nor cenograms agree with paleoenvironmental reconstructions based on proxy data from the Late Pleistocene and Holocene. Cenograms and species-richness models are unreliable and fraught with problems, and both approaches should be abandoned as tools for paleoecological reconstruction. To test for potential geographic bias in the identification of Quaternary fossils I developed a GIS (geographic information systems) database of Quaternary paleontological sites within Texas. I was able to show that the identification of species of fossil soricids, heteromyids, Odocoileus, and Spilogale was influenced by geography. Those fossils should be treated as generic identifications until they are re-evaluated against the full diversity of species. Utilizing GIS I also developed a method of paleoecological analysis. My analysis showed that the environmental conditions found today in Texas might not be limiting the current range of shrews. Based on the known geographic range of shrew fossils, other ecological factors besides environmental conditions are shaping the current distribution of shrews. / text

Records of change of [delta]13C values in vertebrate teeth offer an opportunity to gain insight into changes in past vegetation. Increasingly, teeth from small mammals are used for such purposes, but because their teeth grow very rapidly, seasonal changes in vegetation potentially provide a large source of variability in carbon isotope composition, complicating interpretations of small mammal tooth isotope data. To investigate the controls of seasonality on the stable isotope composition of fossil teeth, we constructed a Monte-Carlo-based model to simulate the effects of changes in the seasonal pattern of diet in leporid lagomorphs (rabbits and hares) on the distribution of [delta]¹³C values in random populations of leporid teeth from the Edwards Plateau in central Texas. Changes in mean-state, seasonal vegetation range, and relative season length manifest themselves in predictable ways in the median, standard deviation, and skewness of simulated tooth [delta]¹³C populations, provided sufficient numbers of teeth are analyzed. This Monte Carlo model was applied to the interpretation of a 20,000 year record of leporid tooth [delta]¹³C values from Hall's Cave on the Edwards Plateau in central Texas. Variations in the [delta]¹³C values of teeth deposited at the same time (standard deviation = 1.69%) are larger than changes in the mean vegetation composition reconstructed from bulk organic carbon [delta]¹³C, indicating the influence of short-term variability, making it difficult to assess changes in mean C3/C4 vegetation from the tooth [delta]¹³C data. However, populations of teeth from different climate intervals (e.g., the late Glacial, Younger Dryas, and the Holocene) display changes in the shape of the tooth [delta]¹³C distributions. Interpretation of these changes as shifts in seasonal vegetation patterns that are based upon results from our model are consistent with hypothesized climatic changes. An increase in the standard deviation of the tooth population between the late Glacial and the Younger Dryas -- Holocene is consistent with an increase in seasonality. Furthermore, a shift to more C3-dominated vegetation in the tooth [delta]¹³C distribution during the Younger Dryas is accompanied by a more skewed population -- indicative of not only wetter conditions but an increase in the duration in the C3 growing season. However, late Holocene changes in vegetation are not clear in the tooth data, despite the evidence from bulk organic carbon [delta]¹³C values for an increase in % C3 vegetation of 57%. Small mammal teeth can potentially provide unique insights into climate and vegetation on seasonal and longer timescales that complement other data, but should be interpreted with a careful consideration of local conditions, taxon ecology and physiology, and the dominant timescales of isotope variability. / text