A WEB OF INTERCONTINENTAL CORRELATION

Arvicolines as Tools for Intercontinental Correlation

Rep dedicated extraordinary effort to establish intercontinental biostratigraphic and biochronologic correlations. During a time when many mammalian paleontologists focused on local, regional, or continental-scale correlation issues (the Great American Biotic Interchange being a notable exception), Rep centered his attention on intercontinental dispersal of mammals and the potential of such dispersals for forging correlations across vast distances. After initial explorations of the mammalian fauna in toto (Repenning 1967), he focused on using arvicolines to establish correlations across North America, Asia, and Europe. Rep's emphasis on the shared history of arvicolines from Eurasia and North America was pioneering in its attempt to establish intercontinental correlations for higher latitudes. His approach represented a clear departure from much of the earlier work by Hibbard and others that focused on regional patterns of change in the North American arvicoline fauna. Ironically, Rep was inspired in part (Repenning, personal commun. to C.J. Bell) by early comments published by Hibbard that arvicolines from the Cudahy fauna bore some resemblance to those from Norfolk, England (Hibbard 1944, 1950).

Immigrants, Boundary Definitions, and Dispersal Events

Rep's correlations of North American and Eurasian faunas were based on the interpretation that many North American arvicoline lineages were immigrant taxa (Repenning and Fejfar 1977; Repenning 1978, 1980, 1983a, 1983b, 1984, 1987, 1998, 2001, 2003; Repenning et al. 1990). The biogeographic context in which he conceptualized immigration to have occurred was complex and centered on the Arctic Ocean borderland. Rep framed, then set out to test, the hypothesis that similarities between North American and western European arvicoline faunas were a result of bi-directional and nearly simultaneous dispersal events along the Arctic Ocean borderland from source areas in Asia (e.g., Repenning 1983b; Repenning 1987; Repenning et al. 1990). Philosophically, he viewed immigrants as the best tools for the establishment of unambiguous definitions of biostratigraphic and biochronologic boundaries. He considered immigration (on a continental scale) to be a more readily recognizable phenomenon than intracontinental dispersals (Repenning 1998). He viewed correlations based on immigration events as a more reliable foundation for establishment of faunally based temporal divisions than stage-of-evolution correlations established based upon anagenetically evolving endemic lineages.

Nevertheless, stage of evolution of certain arvicoline lineages did play a role in the boundary definitions of two divisions he proposed for the Blancan North American Land Mammal Age (NALMA), both of which lacked evidence of North American immigrants (Blancan II, Blancan IV; Repenning 1987). Ultimately, the proposition and recognition of those boundaries was a result of his effort to recognize equivalent numbers of European and North American dispersal events (Bell et al. 2004b). That suggests to us that biogeographically balanced dispersal was a more important concept for him than was the exclusive use of immigrant taxa for the establishment of boundaries.

Purported North American immigrants in his final published formulations (Repenning 1998, 2001) included Mimomys (in two separate dispersals), Pliopotamys, Mictomys, Plioctomys, Allophaiomys, Phenacomys (Paraphenacomys), Microtus (in two separate dispersals), Phenacomys (Phenacomys), Lasiopodomys, Clethrionomys, Terricola, Lemmus, and Dicrostonyx. A persistent problem for Rep was the unambiguous establishment of immigrant status for many of those taxa. With the notable exception of one of the Mimomys dispersals (Repenning 2003), the immigrant status of arvicolines in North America was inferred rather than demonstrated (we note that he would disagree with this claim). Rep did acknowledge that in several cases earliest known occurrences of taxa in North America pre-dated known Eurasian occurrences. He attributed those, in part, to a lack of radioisotopic data from relevant Eurasian localities, or to misidentification of magnetic polarity events in the Eurasian sequence (Repenning 2001). In such instances, he invoked the occurrence of perceived primitive versus advanced forms (morphotypes) of lineages as a basis for his interpretation of dispersal from Eurasia to North America (Repenning 2001).

Many of the details of Rep's proposed biochronology would stand regardless of whether or not the taxa involved were actually immigrants. His purported immigration events were radiometrically dated or calibrated (e.g., Repenning 1967, 1998) and those data serve as useful anchors for any biochronology. As geologically 'instantaneous' events with global expression, magnetic polarity transitions were viewed by Rep as fundamentally important for intercontinental correlation. He utilized radiometrically dated magnetopolarity chronology (e.g., Mankinen and Dalrymple 1979) to calibrate the history of arvicoline rodents in North America and Eurasia throughout much his work (e.g., Repenning 1987, 1992; Repenning et al. 1990, 1995), but arvicolines themselves were often the ultimate arbiters for determination of which polarity event was represented in a stratigraphic section. Many of the North American arvicoline faunas he considered, particularly those in the western United States, were constrained by age control external to the fauna and independent of magnetic polarity assessment (Repenning 1987). Those faunas formed the backbone of his web of biochronologic correlations. Other faunas that were unconstrained by external age control were positioned on the basis of phylogenetic interpretation. In those instances, he seems to have evaluated population-level variation of component taxa, and by assuming a constant rate of evolutionary change, assigned an age range to those unconstrained faunas (see data and locality summaries provided by Repenning 1987, figure 8.1 [in pocket of book], and individual taxon discussions provided by Repenning 1992).

The hypothesized dispersal events that introduced purported immigrants were outlined early in Rep's work on arvicolines (e.g., Repenning 1980, 1983b, 1984), and were tightly integrated with his arvicoline divisions of the NALMAs. Those events were formally recognized and numbered by Rep in 1987 (dispersal events 1-10; Repenning 1987), but new data forced awkward modifications and revisions to the numbered system (e.g., Repenning et al. 1995). Formal, numbered, events were eventually abandoned in favor of nomenclatural practice tied to particular taxa or time intervals (e.g., Repenning 1998). Eventually, Rep came to recognize 16 arvicoline dispersal events, with all but one bringing immigrants from Asia into North America (Repenning 2001). The exception was the emigration of Microtoscoptes from North America into Asia (Repenning 1998, 2001; current phylogenetic hypotheses [not accepted by Rep] suggest that Microtoscoptes was not an arvicoline, but a cricetid exhibiting dental characteristics convergent with arvicolines; see comments by Musser and Carleton 2005, p. 957). The west-to-east predominance was obvious, unexpected, and eventually suggested, with some comedic flavor, as a biological phenomenon of unknown origin (Repenning 1998, p. 55, footnote).

For Rep, dispersal events and the immigrants they introduced were constrained by both geologic and paleoclimatic events, and represented opportunities to explain patterns of biological and evolutionary change in the broader context of geologic and climatic processes that led to periods of population-level isolation or, conversely, dispersals. He discussed extensively the relationship between climate, geology, dispersal events (or lack thereof), and dispersal routes (e.g., Repenning 1990, 1998, 2001). Elucidating those complex relationships was a major goal for Rep in much of his work.

Temporal Range Extensions and Youngest Known Occurrences

An interesting contrast to Rep's emphasis on first-appearance data (especially those associated with immigration) was a disinclination to explore the youngest known occurrences of many arvicolines. In addition, he found it difficult to accept radiometrically dated occurrences of arvicolines that conflicted with his conceptualization of arvicoline chronology. For example, conservative radioisotopic age estimates supported an age of 252,000±30 ka for the Salamander Cave locality in South Dakota (Mead et al. 1996), but Repenning, as a co-author of the paper, argued for an older, Cudahy-equivalent age (600,000±100,000 years old) for the site, based on the presence of Microtus meadensis, Microtus paroperarius, and Mictomys meltoni/kansasensis, taxa known to co-occur in the Cudahy faunas of west Kansas (Hibbard 1944). Rep was intransigent is his arguments, and the final publication proposed two alternative age interpretations, framed as hypotheses in need of additional testing (Mead et al. 1996). Although Repenning clearly recognized the importance of geologically young taxonomic occurrences, particularly in the context of faunal provinciality (Repenning letter to Jim Mead, January 11, 1993), the independent age data from Salamander Cave conflicted with his conception of the temporal distribution of M. paroperarius. His view on the age of Salamander Cave might be interpreted ungenerously only as an unwillingness to accept the possibility that his chronology was flawed, but other factors played a role in his thinking (Repenning letter to Jim Mead, January 11, 1993). Most important among those was the low sample size involved, when considered in light of the total range of variation in the relevant taxa as they were then understood.

Other reports of potentially young occurrences of M. paroperarius had reached Rep's attention as early as 1990. He argued that M. paroperarius was not likely to be present at Cathedral Cave, Nevada, based primarily on the initial late Pleistocene age assignment for the site: "I did not worry about your Microtus being paroperarius because I assumed that your fauna was less than 450,000 years old" (Repenning letter to Bell, December 17, 1990, p. 4). Although subsequent research pushed back the age assignment for the Cathedral Cave fauna, it is still significantly younger than would be expected given Rep's understanding of the temporal distribution of M. paroperarius. New age data on the Cathedral Cave fauna, which includes M. paroperarius, are consistent with the younger age interpretation for the Salamander Cave material (Jass 2007; Jass and Bell 2011).

Rep's initial reticence to embrace the younger ages is somewhat surprising given his acknowledgment of the paucity of data for faunas younger than Cudahy. In 1994 he wrote an informal report on the identification of some arvicoline specimens from the Denver Museum of Natural History excavation in the Velvet Room in Porcupine Cave. In that report he wrote

"It is a Cudahy-type fauna and has a modern Lemmiscus curtatus in it. Not much Mictomys, but some. Cudahy is 0.61 Ma, as you know, but we really do not know the temporal range of this sort of fauna... Consideration of evolutionary rates would suggest that the fauna would not stay so much like Cudahy for more than about 100 kya, and then sometime about 450 kya we get an immigrant of Old World (modern species) of Microtus." (Repenning, letter to Elaine Anderson, 17 June, 1994, pp. 2-3; emphasis added).

Near the end of his life, Rep came into greater acceptance of young occurrences of some arvicoline taxa and to an acceptance of unexpected taxonomic associations of arvicolines that were discovered in several faunas from the western United States. Most notable were the repeated instances of Allophaiomys being found in direct stratigraphic association with some hypothesized evolutionary descendants (Microtus paroperarius, Lemmiscus curtatus, Terricola meadensis). Such occurrences were anathema to Rep when he first encountered them, but he began to incorporate them into his thinking and to modify his biochronologic schemes accordingly. He once wrote on a manuscript review "Things that we have discussed, such as Cathedral Cave and Porcupine Cave have cast considerable doubt in my assumptions about faunal change (which are easier to make when you know less)" (quoted by Christopher Bell, letter to Charles Repenning, 19 May, 1995, p. 4). Rep was intrigued by the possibility that the anomalous taxonomic assemblages found in high-elevation cave sites in the west might reflect the need for independent biochronologies at high elevations, comparable to the latitudinal and longitudinal differences he already recognized. Occurrences that were at first seen as anomalous (e.g., Salamander Cave) became the norm for western cave sites that preserved biotas from relatively high elevations (Porcupine Cave and Cathedral Cave). The interesting thing about these sites was no longer the 'anomalous' taxonomic associations, but rather that those associations were consistently recovered in western sites of the appropriate age and elevational range.

His accommodations in the face of these new data were based on his increasingly acute sense that North American arvicoline biochronology was sufficiently mature to permit recognition of fine-scale patterns of provincialism and, consequently, of independent biochronologies for relatively fine-scale geographic areas.