Geological Society of AmericaSpecial Paper 3612002191Mammals from the end of the age of dinosaurs in North Dakotaand southeastern Montana, with a reappraisal of geographicdifferentiation among Lancian mammalsJohn P. Hunter*Department of Anatomy, New York College of Osteopathic Medicine of New York Institute of Technology,Old Westbury, New York 11568, USAJ. David Archibald*Department of Biology, San Diego State University, San Diego, California 92182, USAABSTRACTAn end-Cretaceous nonavian dinosaur extinction and an early Paleocene mam-malian radiation is documented primarily in stratigraphic sequences in eastern Mon-tana. To determine how representative these sequences are, we extended investigationof this Cretaceous-Tertiary (K-T) transition to new areas. Studies in southwesternNorth Dakota and southeastern Montana provide new records of mammals throughthe last 1.32–1.68 million years of the Cretaceous and extending into the Paleocene,allowing us to evaluate mammalian faunal differentiation across the geographic land-scape of the North American Western Interior in the latest Cretaceous. In NorthDakota, mammals occur at 15 horizons in the Hell Creek Formation from ⬃80 to 3 mbelow the Hell Creek–Fort Union formational contact. In southeastern Montana,where the Hell Creek Formation is ⬃150 m thick, mammals occur from 85 to 15 mbelow the formational contact, with well-sampled local faunas at 65 and 61 m. Infaunal comparisons, the new study areas are closely similar to contemporaneous localfaunas, but differ from those at higher latitudes in several ways. Although mammalianfaunas of the northwestern coastal plain of the Western Interior Seaway (i.e., the HellCreek Formation and its lateral equivalents) show little differentiation, differencesthat exist are strongly associated with geographic distance rather than latitude perse. The overlying Fort Union Formation in North Dakota and southeastern Montanahas produced a few early Paleocene mammalian specimens, promising that futurework in the area will contribute to knowledge of mammalian radiation after the K-Tboundary.Hunter, J.P., and Archibald, J.D., 2002, Mammals from the end of the age of dinosaurs in North Dakota and southeastern Montana, with a reappraisal ofgeographic differentiation among Lancian mammals, in Hartman, J.H., Johnson, K.R., and Nichols, D.J., eds., The Hell Creek Formation and the Cretaceous-Tertiary boundary in the northern Great Plains: An integrated continental record of the end of the Cretaceous: Boulder, Colorado, Geological Society of AmericaSpecial Paper 361, p. 191–216.*E-mail: Hunter, [email protected]; Archibald, [email protected] end of the Cretaceous through the beginning of theTertiary (Paleogene) was a critical time in the evolutionary his-tory of mammals. During this interval, mammalian communi-ties underwent profound changes in composition. In the West-ern Interior of North America, a Tertiary community withdiverse eutherian mammals replaced a Cretaceous communitydominated by primitive marsupials and the now extinct multi-tuberculates (Archibald, 1996; Hunter, 1999). The highest ratesJ.P. Hunter and J.D. Archibald192of extinction and origination, as well as the largest single in-crease in average body mass observed in the North Americanmammalian fossil record (Alroy, 1999), are associated with theCretaceous-Tertiary (K-T) boundary Most extant orders of pla-cental mammals first appeared in the fossil record in a dramaticPaleogene evolutionary radiation following the extinction ofdinosaurs at the K-T boundary (McKenna and Bell, 1997). Thetemporal clustering of these ordinal appearance events (Archi-bald and Deutschman, 2001) and estimated preservation poten-tial of Late Cretaceous mammals (Foote et al., 1999) demon-strate that the origin and diversification of modern placentalorders deep in the Cretaceous, that is, long before the K-Tboundary, is unlikely.Unfortunately, our knowledge of the history of mammalsleading up to the K-T boundary, arguably the single greatestchange in mammalian communities, is restricted geographi-cally. Although there are isolated occurrences in eastern NorthAmerica (Grandstaff et al., 1992), Europe (Grigorescu andHahn, 1987; Le Loeuff and Buffetaut, 1995), and India (Prasadet al., 1995; Krause et al., 1997), mammals from the end of theCretaceous are best known from the North American WesternInterior (Lancian North American Land Mammal Age[NALMA]) from Alberta (Lillegraven, 1969) to New Mexico(Flynn, 1986). Well-sampled stratigraphic sequences that in-clude fossil mammals on both sides of the K-T boundary areeven more restricted latitudinally to eastern Montana (Sloan andVan Valen, 1965; Archibald, 1982; Clemens, this volume) andsouthern Saskatchewan (Johnston, 1980; Johnston and Fox,1984; Fox, 1989, 1997). Although similar sequences of local-ities occur farther south near Glendive, Montana (Hunter et al.,1997) and in the Hanna Basin of southern Wyoming (Eberleand Lillegraven, 1998a, 1998b; Lillegraven and Eberle, 1999),very earliest Paleogene mammals (Pu0 or Pu1; Archibald andLofgren, 1990) have not been recovered near Glendive, andonly a few Cretaceous mammals from the Hanna Basin.In spite of the geographically small window of observationthrough which we view mammalian evolution leading up to theK-T boundary, scenarios that seek to account for mammalianturnover at the boundary have incorporated a geographic com-ponent. Differences in faunal composition among Lancian sites,e.g., an initially greater observed diversity of eutherians athigher latitudes in Lancian time (Lillegraven, 1969), have beeninterpreted to indicate a north-to-south sequence in the transi-tion from Cretaceous-aspect to Paleocene-aspect vertebrate fau-nas, as expected with an invasion of North America by Asianimmigrants (Sloan, 1969; Russell, 1975). Although some sub-sequent discoveries reaffirmed this interpretation (Archibald,1982), others have rendered it less straightforward (Fox, 1989).Separating the effects of geographic, ecological, and temporaldifferences on the composition of Lancian mammalian localfaunas has remained a problem. This difficulty has stemmedlargely from the fact that no field area could boast a sequenceof well-documented localities through Lancian time, by meansof which at least temporal differences could be factored out(Archibald, 1982). The overall goals of this study are to movetoward providing such stratigraphic sequences and to