INTRODUCTION

Caves were recognized as potential sources of paleontological data as early as 1821 when William Buckland reported fossil remains from cave deposits in Kirkdale, England (Buckland 1821). Since that time, numerous research projects carried out in fossiliferous cave deposits have improved our understanding of vertebrate history (Lundelius 2006). Factors contributing to the utility of caves as paleontological data sources include the preservation of large sample sizes (e.g., Barnosky and Bell 2003), preservation of uniquely fossilized materials (e.g., dung deposits; Martin et al. 1961), and the preservation of independent paleoecological data sets (e.g., packrat middens, palynological data, vertebrate fossils, isotopic data derived from speleothems). This combination of features suggests that there is great potential for researchers working in cave deposits to utilize independent data sets for the development of site chronologies and for the interpretation of past biotic conditions. While congruence in such data sets would produce ideal scientific results, the reality is that independent data sets are sometimes found to be inconsistent with each other.

Cathedral Cave, Nevada, preserved a diverse arvicoline rodent assemblage (i.e., voles, lemmings, and muskrats) that was incongruent with late Pleistocene age estimates based on uranium-series dates (Bell 1995; Bell and Barnosky 2000; Bell et al. 2004a; Jass 2005). These conflicting data sets led to new investigations at Cathedral Cave that incorporated fine-scale (5 cm levels) excavation, uranium-series analysis, and paleomagnetic analysis in relation to the occurrence of arvicolines throughout the excavated sedimentary levels (Jass 2007; Jass and Bell in press). Because of the close geographic proximity of Cathedral Cave to other Quaternary cave deposits, I wanted to determine whether a similar scenario might be recorded in other local sites.

In conjunction with research at Cathedral Cave, Nevada, I re-examined arvicoline rodent fossils collected from previous archaeological excavations conducted at Smith Creek Cave (SCC). My re-examination of the SCC arvicoline specimens yielded unexpected results, including the identification of tooth morphologies that resemble taxa not known to persist into the latest Pleistocene (e.g., Microtus paroperarius and M. meadensis). The presence of these morphologies was perplexing, given the known chronological distribution of Microtus paroperarius and M. meadensis, and the published radiocarbon chronology and sedimentary sequence of SCC. In this paper I provide a comprehensive description of the arvicoline rodent taxa from SCC, discuss the possible implications of the arvicoline fauna with respect to age control at the site, and discuss some complexities involved in resolving chronologic sequences in North American cave deposits.

Physical Setting, Background, and Chronologic Framework for Smith Creek Cave

Smith Creek Cave is a large rock shelter (~ 50 m x 18 m x 30 m) located at the mouth of Smith Creek Canyon in the northern Snake Range of eastern Nevada. Since 1925 several archaeological and paleontological excavations were conducted at SCC with the most recent field research efforts taking place in 1968, 1971, and 1974 (see summaries in Bryan 1979 and Mead et al. 1982). Publications addressing the vertebrate fauna from SCC include type descriptions of Oreamnos harringtoni (Stock 1936), Spizaetus willetti (Howard 1935), and Teratornis incrediblis (Howard 1952). Additional reports provide a discussion of the herpetofauna (Brattstrom 1958, 1976), a summary faunal list (Goodrich 1965), a mammalian faunal list (Miller 1979), a detailed review of the late Pleistocene/Holocene fauna known from multiple sites in Smith Creek Canyon (Mead et al. 1982), and additional records of arvicoline rodents (Mead et al. 1992; Bell and Mead 1998).

Early reports on fossils from SCC provided limited data regarding site stratigraphy or age control (see Mead et al. 1982), and the primary chronologic framework for the site was established in conjunction with the archaeological investigations of 1968, 1971, and 1974 (Bryan 1979). A complicated site stratigraphy was reported as part of these investigations, with friable sediments, human occupation, rodent burrowing, and recent human activity all contributing to the complexity of the sedimentary sequence at SCC (Bryan 1979; Miller 1979). Nineteen radiocarbon dates taken on charcoal (n = 14), wood (n = 2), pine needles (n = 1), vegetation (n = 1), and bone (n = 1), placed much of the SCC deposit within the late Pleistocene-Holocene with dates ranging from 28,650 ± 760 yr BP to 1675 ± 90 yr BP (Bryan 1979).

Arvicoline Rodents as Time Markers

Arvicoline rodents are known from the late Miocene-Recent in North America (Repenning et al. 1990). Rapid rates of reproduction known in extant arvicolines, along with a highly diversified and rapidly evolving dentition, underscore the significance of this group in late Tertiary through Quaternary biochronology. Recognition of the importance of arvicoline rodents as biochronological markers in North American terrestrial deposits began with research in the Meade Basin of Kansas (e.g., Hibbard 1944, 1949). Arvicoline biochronology advanced significantly in the 1970s and 1980s with the work of Martin (1979) and Repenning (1978, 1980, 1984, 1987) who developed conceptually and contextually distinct temporal divisions for the continent (Bell 2000). Ultimately, relative chronologic frameworks are subject to change with the addition of new data and/or independent dating methods, and this fluidity is reflected in recent work summarizing the chronological distribution of arvicoline rodent taxa (e.g., Bell and Barnosky 2000; Bell et al. 2004a, 2004b; Martin et al. 2008). Nonetheless, because they are known to occupy discrete time intervals (see Bell 2000, for review), the presence of unique arvicoline tooth morphologies within a sedimentary sequence can be used to bracket the age of fossiliferous deposits.