Printer Friendly

Pleistocene vertebrates from the pre-late Wisconsinan Mill Creek site, St. Clair County, Michigan.


Fossil remains of five fish and eleven mammal taxa are described from sediments underlying tills at the Mill Creek site, northwest of Port Huron, St. Clair County, Michigan. Prior to the discovery of this site, twelve of the sixteen taxa were not known from the fossil record of Michigan. The stratigraphic data, plus the associated fossil flora and fauna, and attempted dating suggest a Pleistocene age of pre-late Wisconsinan, representing at least an interstadial if not an interglacial biota. The dating methods used suggest that at least the molluscs may represent a cluster of two ages, although whether this is also true of the vertebrate fossils is unknown. Although the fish fossils represent wide-ranging taxa, the mammal fossils include northern taxa such as Dicrostonyx sp., Lemmus sp., Synaptomys borealis, and Microtus xanthoganthus. This particular combination of eleven mammal taxa does not co-occur today, although seven of the eleven do co-occur. Although this combination of mammals has not been reported previously as a fossil assemblage, several pre-late Wisconsinan sites have very similar ones. This report of collared lemming fossils, Dicrostonyx sp., is the earliest south of Beringia and, if a larger sample were available, could have an important bearing on the evolution of this genus in North America.



The Mill Creek site consists of fossiliferous sediments (containing pollen, ostracodes, molluscs, and vertebrates) below several tills northwest of Port Huron, St. Clair County, Michigan. A report on this locality, including its location, stratigraphy, fossils, and dating, has been published elsewhere (Karrow et al. 1997). The age of the site is not yet determined. The stratigraphy suggests a Middle Wisconsinan (in other words interstadial) or greater age. Three dating methods have been attempted with only partially concurrent results: a carbon date of 48.3 [+ or -] 0.8 ka, a thermoluminescence date of 57 [+ or -] 9 ka, but perhaps as old as 300 ka, and two clusters of amino acid allo-isoleucine/isoleucine values obtained from molluscs suggest co-occurrence of early Sangamonian and Illinoian specimens (Karrow et al. 1997). The possibility of a mixed age fauna was discussed by Karrow et al. (1997) and has implications for the vertebrate assemblage described here.

As reported in Karrow et al. (1997, Table 4) all but four (Esox sp., Aplodinotus sp., Catostomidae, and Microtus pennsylvanicus) of the sixteen vertebrate taxa from this site had not been reported previously from the Pleistocene of Michigan. Consequently, this fauna featured prominently in a review of the Great Lakes Pleistocene faunas (Holman 2001). However, due to space limitations, the section on fossil vertebrates presented in Karrow et al. (1997) consisted primarily of palaeoenvironmental interpretation. This contribution presents a more complete analysis of this important fauna than could be included in Karrow et al. (1997).


yr. B.P. = years before present

I = upper incisor

M = upper molar

m = lower molar

MNI = minimum number of individuals

P = upper premolar

ROM = Royal Ontario Museum.


Seventy-nine specimens from sixteen taxa described here represent five fish and eleven mammals. No additional material has been identified since the report presented in Karrow et al. (1997). Many taxa are not identified to species level due to the incomplete nature of these fossils. All specimens are catalogued into the collection of the Department of Palaeobiology of the Royal Ontario Museum (ROM) where additional unidentified fragments are also housed.
 Class Osteichthyes
 Order Salmoniformes
 Family Esocidae (Pikes)
 Esox sp.
 ROM 43693, 43694 complete teeth.

Crossman and Harington (1970) summarized the meager fossil record for this genus and noted that single incomplete teeth are impossible to identify with confidence to species. Nevertheless, the overall shape of the teeth is diagnostic to the genus (see Figure 5 in Crossman and Harington, 1970). The only previous fossil records of this genus in Michigan were muskellunge (Esox masquinongy), reported by Wilson (1967) from Fenton Lake (probably a Holocene site), and pike (Esox lucius) from the Pleistocene Shelton mastodon site (Shoshani et al. 1989). The earliest record of Esox sp. in North America is from the lower Pliocene of Oklahoma, while several Pleistocene records include specimens from Kansas, Florida, Ohio, Oklahoma, Texas, Yukon Territory, and Ontario. The single Ontario specimen, an isolated tooth reported by Crossman and Harington (1970) from the Don Valley Formation, Toronto (Figure 5 in Crossman and Harington 1970) has been transferred from the University of Waterloo and is now ROM 29009 in the Department of Palaeobiology, ROM. As noted by Crossman and Harington (1970), the living Northern pike, Esoxlucius, is successful in a wide range of environments and can basically be considered to be a temperate to arctic species, whereas the muskellunge has a more restricted, southerly and easterly, temperate distribution. Both the Northern pike and muskellunge occur in southeast Michigan today (Page and Burr 1991).

 Order Cypriniformes
 Family Cyprinidae (Minnows)
 Cyprinidae, genus undetermined
 ROM 43684, 43685 incomplete pharyngeal arch; 43652, 43686-
 43691 pharyngeal teeth.

The pharyngeal teeth of small cyprinids can usually be separated from the pharyngeal teeth of other North American fish families by having a hook, of variable shape, at the tooth tip. Although the pharyngeal teeth of small catostomids may resemble those of small cyprinids, small catostomids often have a fissure at the tip, whereas adult catostomid pharyngeal teeth have a broad, flattened tip. Generic identification of these specimens was not attempted due to the incomplete nature of the specimens. Although six species of minnows have been reported from the Late Pleistocene of Ohio (Ford et al. 1996), the Mill Creek cyprinid fossils are the first from Michigan. The minnow family is the largest family of fishes; at least 28 species of cyprinid occur in the Mill Creek area today (Page and Burr 1991).
 Family Catostomidae (Suckers)
 Catostomidae, genus undetermined
 ROM 43670-43673 incomplete pharyngeal teeth.

See the comments on identification under Cyprinidae above. As with the cyprinid fossils, these catostomid specimens were not identified to genus due to their fragmentary nature. Four species of catostomid were reported by Wilson (1967), probably from the Holocene of Michigan, and Holman (1979) reported additional specimens of Catostomus commersoni. Eleven species of catostomid are present in the Mill Creek vicinity today (Page and Burr 1991) with wide habitat occurrence.
 Order Gasterosteiformes
 Family Gasterosteidae (Sticklebacks)
 Pungitius sp. cf. P. pungitius
 ROM 26382 (Figure 1A), 43683 incomplete left pelvic spine;
 43674 incomplete right pelvic spine.

 cf. Pungitius sp.
 ROM 43682 incomplete right pelvic spine; 43679, 43680
 incomplete left pelvic spines; 43681 incomplete pelvic spine.

 Gasterosteidae, genus undetermined
 ROM 43678 incomplete left pelvic spine; 43675, 43676 incomplete
 pelvic spines.

Pungitius pelvic spines can be differentiated from those of Gasterosteus, Apeltes, and Culaea by the lack of small tubercles covering the spine (as in Gasterosteus) and by the shape and arrangement of the small spinelets on the dorsal surface of the spine. Unfortunately, all these fossil spines are waterworn to some extent and there is the possibility that a waterworn Gasterosteus spine could be that of Pungitius. The least waterworn spine is ROM 26382 (Figure 1A) and it most closely resembles Pungitius. The only species of stickleback present today in southeast Michigan is Culaea inconstans, the Brook stickleback (Page and Burr 1991), although Pungitius pungitius ranges into northern and western Michigan. The Mill Creek Pungitius fossils are the first stickleback fossils to be reported from Michigan. Pungitius pungitius prefers shallow vegetated areas of lakes, ponds, and pools of streams (Page and Burr 1991).
 Order Perciformes
 Family Sciaenidae (Drumfish)
 Aplodinotus sp. cf. A. grunniens
 ROM 43692 complete pharyngeal tooth.

The pharyngeal teeth of sciaenids are easily identified by their overall rounded or quadrate shape and the shape of the tooth root. Aplodinotus grunniens is the only extensively freshwater member of this family in North America (Lee et al. 1980). Hubbs (1940) previously reported Aplodinotus grunniens from postglacial sediments of Michigan and Harington (1971) summarized the Pleistocene and postglacial records for this species. The Freshwater drum lives in southeast Michigan today, where it prefers the bottom of medium to large rivers and lakes (Page and Burr 1991).
 Class Mammalia
 Order Insectivora
 Family Soricidae (Shrews)
 Sorex hoyi
 ROM 26380 complete left P4 (Figure 1B).

The identification of this single shrew tooth proved difficult; however, the overall size and morphology indicates S. hoyi. Karrow and Seymour (1991) first reported this tooth as Sorex sp. cf. S. arcticus, but the identification of this tooth as S. hoyi had been ascertained by the time Karrow et al. (1997) was published. Mesiodistal length of this tooth is 1.28 mm; comparative measures for Recent Sorex species are as follows (mean, followed by sample size in parentheses; all specimens housed in the Department of Mammalogy, ROM): S. cinereus from the Yukon Territory 1.17 (10), S. cinereus from Ontario 1.19 (12), S. hoyi from Ontario 1.25 (11), S. tundrensis from British Columbia 1.39 (4), S. fumeus from Ontario 1.45 (12), S. arcticus from Ontario 1.49 (12). The posteromedial cusp (= hypocone of van Zyll de Jong 1983) is well developed posteriorly and most closely resembles Recent S. hoyi. This cusp is not well developed in S. cinereus or S. tundrensis and only partially expanded in S. arcticus and S. fumeus. The transverse ridge posterior to the metacone is well developed as in Recent S. hoyi; this feature is not well developed in S. arcticus. The metacone is relatively short (compared to the mesiodistal length) as in Recent S. hoyi and unlike S. arcticus or S. fumeus. Therefore the combination of these three features defines S. hoyi. In addition, at least one of these features did not match with S. bendirii, S. dispar, S. longirostris, S. monticolus, S. palustris, S. trowbridgii, or S. vagrans.

S. hoyi occurs in the Lower Peninsula of Michigan today but apparently not in the Mill Creek area (Forsyth 1985; Kays and Wilson 2002). It prefers grassy glades within the boreal forest as well as sphagnum bogs or the shrubby borders of these bogs (Banfield 1974). The only previous record of fossil Sorex from Michigan was a late holocene record of Sorex cinereus from Mackinac County (Holman et al. 2003). S. hoyi has been reported from at least 10 other Pleistocene fossil localities in North America (Kurten and Anderson 1980).
 Order Carnivora
 Family Mustelidae (Weasels)
 Mustela erminea
 ROM 26381 complete left humerus.

Although this specimen was broken in three places, the break apparently occurred during collection or screenwashing, as the pieces were reassembled easily. The bone matches the size and morphology of Recent female M. erminea. Greatest length of fossil: 24.9 mm; average lengths of Recent species (sample size in brackets; specimens primarily from Ontario and located in the ROM Mammalogy and Palaeobiology Departments): M. nivalis 18.6 (2); M. erminea females 23.8 (14); M. erminea males 29.1 (30); M. frenata females 30.2 (8); M. frenata males 33.2 (33). Ralls and Harvey (1985) showed there was a pronounced sexual dimorphism for these three species, as well as geographic variation. The average size of these species could have fluctuated during the Pleistocene, much as Kurten (1968) demonstrated for several other European carnivores, although this has not been documented for the ermine, in either North America or Europe. Therefore size and overall morphology are used here to identify this fossil. There are at least 14 Rancholabrean fossil localities of M. erminea in North America (Anderson 1984). The only fossil mustelid reported previously from Michigan was Martes americana, the marten, from the Holocene Sleeping Bear Dune locality (Pruitt 1954), so the Mill Creek specimen is the first record of a fossil Mustela weasel from Michigan. The ermine lives in a wide range of habitats from deep forest to open tundra but prefers brushy or rocky cover, often near water; today it is near the southern limit of its range in southeastern Michigan (Forsyth 1985; Kays and Wilson 2002).
 Order Rodentia
 Family Muridae
 Subfamily Crictinae (Mice)
 Peromyscus sp.
 ROM 43651 complete left upper M1 (Figure 1C).

The teeth of P. maniculatus and P. leucopus are very similar. Guilday and Handley (1967) and Guilday et al. (1977) used the morphology of the first lower molar to distinguish between the two species. Foley (1984) also used morphology but had less success using size. On geographic grounds, P. maniculatus might be the most likely identity for this tooth, as today this species ranges much further north than does P. leucopus (Banfield 1974; Kays and Wilson 2002). However, since the literature identifications are based on lower molars only, this single upper molar is not here identified to the specific level. P. maniculatus has been identified from more complete fossil material from more than 25 fossil localities, and P. leucopus is also widely distributed in localities of Rancholabrean age (Kurten and Anderson 1980). The only previous record of fossil Peromyscus from Michigan was a late Holocene report from Mackinac County (Holman et al. 2003). P. maniculatus has been found in an extremely wide variety of habitats; when associated with P. leucopus, it tends to be restricted to more open habitats (Kurten and Anderson 1980).
 Subfamily Arvicolinae (Voles and lemmings)
 Dicrostonyx sp.

ROM 43664 incomplete frontal; 43663 incomplete palate; 26386 incomplete right M1; 26387 complete left M1, Type I (Figure 1D); 43697 incomplete M2; 26368 complete right M2, Type I (Figure 1E); 26385 incomplete left M3, Type II (Figure 1F); 26365 incomplete right M3, Type III (Figure 1G); 26383 incomplete right M3; 26388 complete right m1, Type II (Figure 1H); 26366, 26371 incomplete right m1; 26370 incomplete left m1; 26384 incomplete right m3; 26367 incomplete left m3; 26369 complete left m3. MNI = 3 using right m1

All of this material represents Dicrostonyx due to the deep re-entrant angle present on both sides of the teeth. Mead and Mead (1989) reported 24 North American fossil localities for Dicrostonyx; none were in Michigan. The Mill Creek site falls midway between the most easterly known fossils of D. groenlandicus (= D. torquatus; Moscow Fissure, Wisconsin; Foley, 1984) and the closest site for fossil D. hudsonius (New Paris No. 4, Pennsylvania; Guilday et al., 1964); Figure 2. On geographic grounds, either species might be expected to occur in Michigan. Two issues further complicate the possible species level identification of this fossil material: the recent recognition of D. richardsoni as a valid living species (Engstrom et al. 1993) and the fact that Recent D. hudsonius may be indistinguishable from the earlier fossil species D. simplicior. D. simplicior was widespread in the Holarctic, but it is not clear how long it persisted. The potential confusion of D. simplicior and D. hudsonius may be a problem for a fossil site the age of the Mill Creek site.


As these Mill Creek specimens constitute the earliest known occurrence of Dicrostonyx south of Beringia, they may have an important bearing on the understanding of the divergence of D. hudsonius and D. groenlandicus in North America. D. richardsoni can't yet be identified using dental data, so the following discussion will concern only the other species. ROM 26384 preserves a portion showing one of the extra cusps present in living D. groenlandicus that are not usually present in D. hudsonius. It was on the basis of this tooth that Karrow and Seymour (1991) reported D. groenlandicus from this locality. However, Agadjanian (1986) has shown that these extra cusps in living D. groenlandicus are more likely the end product of a gradual series of changes from the Middle Pleistocene to the Recent. In particular, the three upper molars and the first lower molar show gradual additions of cusps, which he subdivided into four morphotypes (Types I to IV). Large samples of Middle Pleistocene Palearctic Dicrostonyx cf. D. simplicior show a predominance of the simpler Type I morphotype, while Recent Palearctic and Nearctic Dicrostonyx groenlandicus show a predominance of the more complicated Type III morphotype, with the occasional presence of morphotypes Type II or IV. Intermediate-aged specimens have been called D. gulielmi, D. henseli, or D. gulielmi-henseli (Agadjanian and von Koenigswald 1977; Agadjanian 1986). These specimens show primarily Type II morphotypes with some Type I and Type III present, depending on which molar is considered. Only large samples (most of Agadjanian's samples were larger than 50 teeth) demonstrated that with time the proportion of each morphotype present shifted toward the more complex. Consequently, only a histogram based on a large sample can form the basis for a species identification of a group of teeth. Even this approach will not necessarily identify an individual tooth. Using a similar approach, Morlan (1984) showed that Middle Wisconsinan Dicrostonyx from the Yukon may include the final stages of evolution from D. gulielmi to D. groenlandicus.

There are only five teeth from the Mill Creek site that are complete enough to assign to a morphotype of Agadjanian: there are two Type I (one each of M1 and M2), two Type II (one each of M3 and m1), and one Type III (an M3). According to Agadjanian (1986), both Type I and Type II molars occur in most populations of Middle to Late Pleistocene age in Eurasia. Two teeth of each Type is not a sufficient sample upon which to make a taxonomic judgment. However, Type III UM3s do not occur in the Middle or Late Pleistocene precursors of D. groenlandicus nor do they occur in the Recent D. hudsonius. Thus this single tooth is the only evidence of an "advanced" Dicrostonyx at Mill Creek. Therefore, this sample of Pre-Late Wisconsinan Dicrostonyx from Michigan may represent a population of collared lemmings already evolving into D. groenlandicus. This is similar to Morlan's (1984) Unit 2A at localities 12 and 15, Old Crow, Yukon Territory dated at >51,000 yr. B.P., in which he reported teeth representing the transitional D. gulielmi-torquatus (torquatus = groenlandicus).

Although today these species are tundra obligates, North American fossil Dicrostonyx sometimes has been recovered in disharmonious or nonanalogue faunas (as defined by Graham 1979, 1985; Graham and Mead 1987; Lundelius et al. 1983) and may only be indicative of boreal or steppe environments and not necessarily strict tundra (Mead and Mead 1989). Today the collared lemming prefers to eat shrubs, particularly willow leaves and herbs (Forsyth 1985; Rodgers and Lewis 1985).
 Synaptomys (Mictomys) borealis
 ROM 26389 incomplete left m1 (Figure 1I).

Although incomplete, the morphology compares very favorably with Recent S. borealis, with deep lingual reentrants and the lack of labial triangles. Mead et al. (1992) noted that the lower molars of the northern (S. borealis) and southern (S. cooperi) bog lemmings are readily identifiable to species. The evolutionary history of the bog lemmings is complex and several different versions have been suggested (Barnosky and Rasmussen 1988; Hibbard 1956; von Koenigswald and Martin 1984; Repenning 1987; Repenning and Grady 1988; Repenning et al. 1987). Paleontologists tend to use Mictomys as a genus for S. borealis whereas mammalogists generally use it as a subgenus; here I follow the mammalogists. The only previous record of fossil Synaptomys from Michigan was from the late Holocene of Mackinac County (Holman et al. 2003). This species has been recovered from at least nine North American Pleistocene sites, as far south as northern Arkansas (Kurten and Anderson 1980). Today, the northern bog lemming ranges across Canada and into Alaska but does not occur in Michigan. It consumes grasses and sedges and prefers sphagnum bogs, although it may also occur in mossy spruce woods, wet subalpine meadows, and tundra (Banfield 1974).
 Lemmus sp. cf. L. trimucronatus (= sibiricus)
 ROM 43699 incomplete right M1 (Figure 1J).

Although incomplete, the large size and deep buccal bays indicate this tooth probably belongs to the genus Lemmus. Additionally, in the Recent sample available (in both ROM Mammalogy and Palaeobiology), the anteriad curvature of the anterior enamel portion of each buccal bay suggests Lemmus, as this enamel portion is straighter in both Synaptomys species. This constitutes the third record of fossil Lemmus south of Beringia, and the first for Michigan. The previous two records were from January Cave, Alberta (Burns 1980) and Prairieburg, Iowa (Foley and Raue 1987). Except for parts of north central British Columbia and northeastern Manitoba, the brown lemming today does not occur south of the sixtieth parallel. The brown lemming eats grasses, sedges, and moss (Rodgers and Lewis 1985) and prefers to inhabit wet tundra swales but may also occur on lakeshores, stream banks, grassy slopes, and even sphagnum bogs (Banfield 1974). I follow Jarrell and Fedga (1993) in using the name L. trimucronatus instead of L. sibiricus for this species.
 Clethrionomys sp.
 ROM 43661 complete left M2 (Figure 1K).

The rounded enamel ridges of this tooth match the morphology of Recent comparative C. gapperi. Red-backed voles begin life with rootless teeth that develop roots as they mature (Tupikova et al. 1968). This tooth is still rootless and so does not represent an adult individual. Unfortunately, the two North American species of Clethrionomys, C. gapperi and C. rutilus, are dentally indistinguishable (Rausch and Rausch 1972; Kurten and Anderson 1980). Nevertheless, Clethrionomys gapperi has been reported from 42 fossil sites in eastern and central North America alone (Richards 1986). The only previous record of fossil Clethrionomys from Michigan was from the late Holocene of Mackinac County (Holman et al. 2003). Kurten and Anderson (1980) reported at least 20 sites from across North America, including several that are late Irvingtonian in age although most are Rancholabrean. Many fossils have been reported south of its present distribution (Richards 1986), as is the case here. Today, C. gapperi can be found in both coniferous and mixed-hardwood forested areas as well as shrubby habitats. It is omnivorous, eating such food as vegetative parts of grasses, conifer seeds, roots, bark, insects, and fungi (Forsyth 1985).
 Microtus pennsylvanicus
 ROM 26374, 43696 incomplete right M2; 26395 complete right
 M3; 26394 complete left m1; 26396 incomplete left m1.

 Microtus cf. M. pennsylvanicus
 ROM 43662 incomplete palate; 26393 complete left m3.
 MNI = 2 using left m1.

These specimens represent the common and widespread meadow vole. ROM 26395 has an incipient fourth lingual reentrant on its posterior loop. The m1s have well-developed fourth buccal reentrants (cemented in ROM 26396) and fifth lingual reentrants (cemented in ROM 26394). Additionally, these teeth exhibit the typical meadow vole pattern of differentiated enamel thickness, with thicker anterior enamel on the triangles of the lower teeth and thicker posterior enamel on the triangles of the upper teeth. The specimens identified as cf. M. pennsylvanicus are specimens of Microtus sp. that appear to represent M. pennsylvanicus but for which there is still some doubt. Dental variation in this species was studied by Guilday (1982); he noted that occlusal patterns varied so much that different species may show identical occlusal patterns. There are at least three Michigan Pleistocene localities for this species (Holman 1979; Holman et al. 1986; Shoshani et al. 1989) as well as two Holocene reports (Holman et al. 2003; Wilson 1967) and over 40 late Pleistocene localities across North America (Kurten and Anderson 1980). Today this species eats primarily herbaceous vegetation, seeds, bark, insects, snails, and other small invertebrates and vertebrates (Forsyth 1985). Its preferred habitat is wet meadows but it may inhabit almost any grassland habitat (Banfield 1974).
 Microtus xanthognathus
 ROM 43647 incomplete right M1; 26390 complete right m1
 (Figure 1L); 26392, 43648, 43666 incomplete right m1; 26372
 incomplete left m1.

 Microtus cf. M. xanthognathus
 ROM 26373, 26391 incomplete right M1; 43667 complete right
 M2; 26397 complete left M3; 43695 incomplete right m1; 43665
 incomplete right m2.
 MNI = 3 using right m1.

The large size of these specimens was the first clue to the presence of this species. Using Recent comparative material, the morphology, such as the intermediate enamel thickness differentiation, also matched quite well. The specimens identified as cf. M. xanthognathus are specimens of Microtus sp. that appear to represent M. xanthognathus but with some doubt. Large size was the main reason why they were not defined as Microtus sp. Dental variation in this species was studied by Guilday (1982), who noted that occlusal patterns varied so much that sometimes different species show identical occlusal patterns. The exceptions here are ROM 26397 and 43695 which are a little smaller than the Recent specimens examined but which appear to match the morphology of this species. This species has been found before at over a dozen sites in various states south of its present range (Kurten and Anderson 1980), but the Mill Creek specimens are the first from Michigan. Today the species inhabits riparian boreal forest edge and prefers to consume horsetails, lichens, grasses, and berries (Banfield 1974; Forsyth 1985).
 Microtus sp.
 ROM 43698 complete right M2; 26398 incomplete left M2;
 26375 incomplete left M2?; 26376, 26379, 26399, 43649, 43650
 incomplete left m1; 26377, 26378 incomplete right m1.

These specimens were either too incomplete for specific identification or of uncertain specific status; however, they appear to represent the genus Microtus. Martin and Prince (1990) demonstrated in one locality that statistically significant intraspecific differences in M. pennsylvanicus teeth could be found over a period of 15,000 years. It is not surprising, therefore, that some teeth from Mill Creek, which may represent a mixed-aged site (Karrow et al. 1997), could not be confidently identified to the specific level using Recent comparative material.
 Family Zapodidae (jumping mice)
 Zapus sp. cf. Z. hudsonius
 ROM 43668 incomplete right I.

Although other small rodents have grooved upper incisors (for example Lemmus and Synaptomys), the shape and curvature of this specimen best compares with Recent zapodid incisors, in particular Zapus. There are at least eight other fossil localities reported from a variety of central and eastern states (Kurten and Anderson 1980), but the Mill Creek specimens are the first from Michigan. Zapus hudsonius lives throughout Michigan today, where it prefers grasslands, meadows, and forest edges. An omnivore, it eats primarily seeds and insects and their larvae (Forsyth 1985).
 Order Lagomorpha
 Family Leporidae (rabbits and hares)
 Leporidae, genus undetermined
 43669 incomplete upper tooth.

Although this tooth is very incomplete, its large size and compressed shape indicate a lagomorph, most likely a leporid. An ochotonid is tentatively ruled out because of the large size of this tooth, rather than for zoogeographic reasons (Mead 1987; Mead and Grady 1996), as there are two records of fossil pika from Ontario (Churcher and Dods 1979; Savage 1994). This Mill Creek specimen represents the first fossil leporid from the Pleistocene of Michigan.


Although fragmentary, this fauna is important because only two previous records exist of interstadial fossil vertebrates from Michigan: a Lesser scaup (Aythya affinis) bone from Muskegon County (Kapp 1978; Holman 1976), and mandible fragments and teeth of a Jefferson mammoth (Mammuthus jeffersonii) from Midland County (Kapp 1970). Only four of the 16 taxa reported here have been recovered from a fossil site in Michigan before (Esox sp., Catostomidae, Aplodinotus sp., and Microtus pennsylvanicus).

This is a small fauna, with eight of the 11 mammals being represented by a single specimen. For the other three mammals the minimum number of individuals represented is three for Dicrostonyx sp. and Microtus xanthognathus and two for M. pennsylvanicus. No tooth was found in association with any other and all skull parts identified were lacking teeth. Most teeth were incomplete and some showed signs of abrasion or etching, suggesting specimen transport prior to deposition or possible post-depositional decay. Without tooth rows or very many complete teeth for the mammals or complete pharyngeal arches for the fish, the identifications must remain incompletely determined for many taxa. Most of the remaining unidentified and uncatalogued specimens are extremely fragmentary or represent difficult to identify bones, such as incomplete fish vertebrae, rodent vertebral epiphyses, or rodent podial elements.

At the present time the sedimentology of the bone-bearing horizon remains unstudied. Very few bones were recovered per bag of matrix, however, suggesting a sparse distribution of bones or perhaps small pockets within this layer.

With fish and mammals represented, this fossil deposit obviously samples both aquatic and terrestrial ecosystems. The fish fossils do not indicate any particular aquatic environment and are all fairly wide-ranging today. The habitat preferences for most of the mammals recovered includes riparian forest, bogs, or wet meadows or swales which could have been located next to the stream, river, or lake in which the fish lived. As far as can be determined all the fish except Pungitius occur in the Mill Creek area today.


The story for the mammals is quite different. Six of the eleven mammals (Mustela erminea, Dicrostonyx sp. Synaptomys borealis, Lemmus sp. cf. L. trimucronatus, Clethrionomys sp., and Microtus xanthognathus) do not occur in the Mill Creek area today. As discussed in Karrow et al. (1997), of these six, the two lemming genera (Dicrostonyx and Lemmus) today are tundra obligates, whereas the other four species have boreal affinities. The remaining five mammal species are more cosmopolitan in distribution. Seven of eleven of these mammal species today co-occur in an area in northern Canada and Alaska (Figure 3). However, the particular combination of eleven mammal species found at Mill Creek does not co-exist today anywhere in North America. This mammal fauna then could be one more nonanalogue Pleistocene fauna (Graham and Mead 1987), many of which already have been reported (Lundelius et al. 1983; Graham and Mead 1987).

However, two clusters of amino acid allo-isoleucine/isoleucine values obtained from molluscs suggest co-occurrence of early Sangamonian and Illinoian specimens, or at least interglacial and interstadial specimens (Karrow et al. 1997). This suggests that this fossil vertebrate deposit may not have formed at one time, thus these species may not have co-occurred at this location. An alternative explanation is that, because of the paucity of similarly aged faunas in this region of the continent, this combination of animals may reflect geographic location and taphonomic sampling rather than being a mixed-aged site (Karrow et al. 1997). At the moment I cannot choose between these two alternatives. As evidence of the possibility that this may be a synchronously deposited fauna, there are several other faunas, discussed below, which have preserved very similar faunas.

Beside Pleistocene fossil sites in Alaska (e.g., Guthrie 1968; Repenning et al. 1964) and the Yukon (e.g., Morlan 1989) that record arctic or boreal species of small mammals similar to those found at Mill Creek, there are two sites from south of Beringia that preserve similar faunas: Conklin Quarry, Iowa (Baker et al. 1986) and Moscow Fissure, Wisconsin (Foley 1984). Both of these sites are older than latest Pleistocene: Conklin Quarry dates between 18,090 [+ or -] 190 and 16,710 [+ or -] 270 yr. B.P. and Moscow Fissure dates from 17,050 [+ or -] 1,500 yr. B.P. Both preserve what are called "full glacial biotas" by Baker et al. (1986). Mill Creek certainly appears to be another such fauna, albeit slightly older. The mammals present at Mill Creek that are also present at Conklin Quarry are Clethrionomys sp., Dicrostonyx sp., Microtus pennsylvanicus, and M. xanthognathus. Only two boreal mammals found at Conklin Quarry are not found at Mill Creek: Phenacomys intermedius and Microtus sp. cf. M. miurus. All mammals found at Mill Creek are found at Moscow Fissure, Wisconsin except Lemmus sp. and Mustela erminea. Therefore, the particular combination of fossil mammals found so far at Mill Creek has not yet been found elsewhere in North America.

There are also several latest glacial sites that preserve similar faunas. Two are in Appalachia: Clark's Cave, Virginia dating at >10,000 yr. B.P. (preserved all mammals found at Mill Creek except the two lemmings Lemmus and Dicrostonyx; Guilday et al. 1977), and New Paris No. 4, Pennsylvania dating at 11,300 BP (preserved all Mill Creek mammals except Lemmus and Mustela erminea; Guilday et al. 1964). An additional similar site is the Sheridan Pit at Indian Trail Caverns, Ohio (Hansen 1992; McDonald 1994; Ford et al. 1996). This site preserves more than just a microvertebrate fauna but seven of the eleven Mill Creek mammals have been recovered so far (lacking so far are Dicrostonyx sp., Lemmus sp. cf. L. trimucronatus, and Zapus sp. cf. Z. hudsonius).

So far no fossil amphibians, reptiles, or birds have been found at Mill Creek. The Mill Creek fauna represents a boreal environment and has no living analogue only because of the presence of the two tundra obligate lemming genera. As Mead and Mead (1989) pointed out, however, Pleistocene Dicrostonyx may only indicate boreal conditions and not necessarily tundra, just as Pleistocene Ochotona may indicate boreal conditions and not necessarily talus slopes (Mead 1987; although see also Hafner 1993).


Thanks are due to P. F. Karrow for inviting me to work on this fauna, and to J. A. Burns, C. A. Repenning, and especially R. E. Morlan for examining parts of the Mill Creek collection of microvertebrates and for their generosity with their time and advice. J. A. Holman reviewed this paper and provided helpful advice that improved the manuscript. P. Diebolt, V. Gaitan, and A. L. M. Davis screenwashed the sediments. J. Mulock drew Figure 1, N. Bahar prepared Figures 2 and 3, and P. Fenton redrafted Figure 2.


AGADJANIAN, A. K. 1986. The history of collared lemmings in the Pleistocene. In Beringia in the Cenozoic era (translation of the 1976 Beringiya v Kainozoe), ed. V. L. Kontrimavichus. Rotterdam: A. A. Balkema.

AGADJANIAN, A. K., AND W. VON KOENIGSWALD. 1977. Merkmalsverschiebung an den oberen Molaren von Dicrostonyx (Rodentia, Mammalia) im Jungquartar. Neues Jahrbuch fur Geologie und Palaontologie Abhandlungen 153:33-49.

ANDERSON, E. 1984. Review of the small carnivores of North America during the last 3.5 million years. Carnegie Museum of Natural History Special Publication 8:257-66.

BAKER, R. G., R. S. RHODES II, D. P. SCHWERT, A. C. ASHWORTH, T. J. FREST, G. R. HALLBERG, AND J. A. JANSSENS. 1986. A full-glacial biota from Southeastern Iowa, U.S.A. Journal of Quaternary Science 1:91-107.

BANFIELD, A. W. F. 1974. The mammals of Canada. Toronto: University of Toronto Press.

BARNOSKY, A. D., AND D. L. RASMUSSEN. 1988. Middle Pleistocene arvicoline rodents and environmental change at 2900 meters elevation, Porcupine Cave, South Park, Colorado. Annals of the Carnegie Museum 57:267-92.

BURNS, J. A. 1980. The brown lemming, Lemmus sibiricus (Rodentia, Arvicolidae), in the Late Pleistocene of Alberta and its postglacial dispersal. Canadian Journal of Zoology 58:1507-11.

CHURCHER, C. S., AND R. R. DODS. 1979. Ochotona and other vertebrates of possible Illinoian age, from Kelso Cave, Halton County, Ontario. Canadian Journal of Earth Sciences 16:1613-20.

CROSSMAN, E. J., AND C. R. HARINGTON. 1970. Pleistocene pike, Esox lucius, and Esox sp., from the Yukon Territory and Ontario. Canadian Journal of Earth Sciences 7:1130-8.

ENGSTROM, M. D., A. J. BAKER, J. L. EGER, R. BOONSTRA, AND R. J. BROOKS. 1993. Chromosomal and mitochondrial DNA variation in four laboratory populations of collared lemmings (Dicrostonyx). Canadian Journal of Zoology 71:42-48.

FOLEY, R. L. 1984. Late Pleistocene (Woodfordian) vertebrates from the Driftless area of Southwestern Wisconsin, the Moscow Fissure local fauna. Illinois State Museum Report of Investigations 39:1-50.

FOLEY, R. L., AND L. E. RAUE. 1987. Lemmus sibiricus from the Late Quaternary of the midwestern United States. Current Research in the Pleistocene 4:105-7.

FORD, III, K. M., A. R. BAIR, AND J. A. HOLMAN. 1996. Late Pleistocene fishes from Sheriden Pit, northwestern Ohio. Michigan Academician 28:135-45.

FORSYTH, A. 1985. Mammals of the Canadian wild. Camden East, Ontario: Camden House.

GRAHAM, R. W. 1979. Paleoclimates and late Pleistocene faunal provinces in North America. In Pre-Llano cultures of the Americas: Paradoxes and possibilities, ed. R. L. Humphrey and D. Stanford. Washington D. C.: The Anthropological Society of Washington.

_____. 1985. Diversity and community structure of the late Pleistocene mammal fauna of North America. Acta Zoologica Fennica 170:181-92.

GRAHAM, R. W., AND J. I. MEAD. 1987. Environmental fluctuations and evolution of mammalian faunas during the last deglaciation in North America. In North America and adjacent oceans during the last deglaciation: The geology of North America, Volume K-3, ed. W. F. Ruddiman and H. E. Wright Jr. Washington, D.C.: Geological Society of America.

GUILDAY, J. E. 1982. Dental variation in Microtus xanthognathus, M. chrotorrhinus, and M. pennsylvanicus (Rodentia: Mammalia). Annals of the Carnegie Museum of Natural History 51:211-30.

GUILDAY, J. E., AND C. O. HANDLEY, JR. 1967. A new Peromyscus (Rodentia: Cricetidae) from the Pleistocene of Maryland. Annals of the Carnegie Museum 39:91-103.

GUILDAY, J. E., P. S. MARTIN, AND A. D. MCCRADY. 1964. New Paris Number 4: A Pleistocene cave deposit in Bedford County, Pennsylvania. National Speleological Society Bulletin 26:121-94.

GUILDAY, J. E., P. W. PARMALEE, AND H. W. HAMILTON. 1977. The Clark's Cave bone deposit and the Late Pleistocene paleoecology of the central Appalachian mountains of Virginia. Carnegie Museum of Natural History Bulletin 2:1-88.

GUTHRIE, R. D. 1968. Paleoecology of a Late Pleistocene small mammal community from interior Alaska. Arctic 21:223-44.

HAFNER, D. J. 1993. North American pika (Ochotona princeps) as a Late Quaternary biogeographic indicator species. Quaternary Research 39:373-80.

HANSEN, M. C. 1992. Indian Trail Caverns: A window on Ohio's Pleistocene bestiary. Rocks and Minerals 67:405-9.

HARINGTON, C. R. 1971. A postglacial freshwater drum (Aplodinotus grunniens) from Ontario, and comments on the zoogeography of the species. Canadian Journal of Earth Sciences 8:1137-44.

HIBBARD, C. W. 1956. Vertebrate fossils from the Meade Formation of southwestern Kansas. Papers of the Michigan Academy of Sciences, Arts, and Letters 49:145-203.

HOLMAN, J. A. 1976. A 25,000-year-old duck, more evidence for a Michigan Wisconsin Interstadial. American Midland Naturalist 96:501-3.

_____. 1979. New fossil vertebrate remains from Michigan. Michigan Academician 11:391-7.

_____. 2001. In Quest of Great Lakes Ice Age Vertebrates. East Lansing, Michigan: Michigan State University Press.

HOLMAN, J. A., D. C. FISHER, AND R. O. KAPP. 1986. Recent discoveries of fossil vertebrates in the lower peninsula of Michigan. Michigan Academician 18:431-63.

HOLMAN, J. A., B. L. LUNDRIGAN, AND P. MYERS. 2003. Late Holocene (Little Ice Age interval) microvertebrates from Mackinac County, Michigan. Michigan Academician 35:159-69.

HUBBS, C. L. 1940. The cranium of a fresh-water sheepshead from post-glacial marl in Cheboygan County, Michigan. Papers of the Michigan Academy of Science, Arts, and Letters 25:293-6.

JARRELL, G. H., AND K. FREDGA. 1993. How many kinds of lemmings? A taxonomic overview. In The biology of lemmings, ed. N. C. Stenseth and R. A. Ims. London: Linnean Society Symposium Series, no. 15, Academic Press.

KAPP, R. O. 1970. A 24,000-year-old Jefferson mammoth from Midland County, Michigan. Michigan Academician 3:95-99.

_____. 1978. Plant remains from a Wisconsinan interstadial dated 25,000 B.P., Muskegon County, Michigan. American Midland Naturalist 100:506-9.

KARROW, P. F., AND K. SEYMOUR. 1991. Interstadial microvertebrates from Mill Creek, St. Clair County, Michigan. Geological Society of America, Abstracts with Programs 23(3): 20.

KARROW, P. F., K. L. SEYMOUR, B. B. MILLER, AND J. E. MIRECKI. 1997. Pre-Late Wisconsinan Pleistocene biota from southeastern Michigan, U.S.A. Palaeogeography, Palaeoclimatology, Palaeoecology 133:81-101.

KAYS, R. W., AND D. E. WILSON. 2002. Mammals of North America. Princeton: Princeton University Press.

KURTEN, B. 1968. The Pleistocene mammals of Europe. London: Weidenfeld and Nicolson.

KURTEN, B., AND E. ANDERSON. 1980. Pleistocene mammals of North America. New York: Columbia University Press.

LEE, D. S., C. R. GILBERT, C. H. HOCUTT, R. E. JENKINS, D. E. MCALLISTER, AND J. R. STAUFFER, JR. 1980. Atlas of North American freshwater fishes. North Carolina Biological Survey Publication 1980-12.

LUNDELIUS JR., E. L., R. W. GRAHAM, E. ANDERSON, J. E. GUILDAY, J. A. HOLMAN, D. W. STEADMAN, AND S. D. WEBB. 1983. Terrestrial vertebrate faunas. In The Late Pleistocene, ed. S. C. Porter. Minneapolis: University of Minnesota Press.

MARTIN, R. A., AND R. H. PRINCE. 1990. Variation and evolutionary trends in the dentition of Late Pleistocene Microtus pennsylvanicus from three levels in Bell Cave, Alabama. Historical Biology 4:117-29.

MCDONALD, H. G. 1994. The Late Pleistocene vertebrate fauna in Ohio: Coinhabitants with Ohio's Paleoindians. In The first discovery of America: Archaeological evidence of the early inhabitants of the Ohio area, ed. W. S. Dancey. Columbus, Ohio: Ohio Archaeological Council.

MEAD, J. I. 1987. Quaternary records of pika, Ochotona, in North America. Boreas 16:165-71.

MEAD, J. I., C. J. BELL, AND L. K. MURRAY. 1992. Mictomys borealis (Northern Bog Lemming) and the Wisconsin paleoecology of the east-central Great Basin. Quaternary Research 37:229-38.

MEAD, J. I., AND F. GRADY. 1996. Ochotona (Lagomorpha) from Late Quaternary cave deposits in eastern North America. Quaternary Research 45:93-101.

MEAD, E. M., AND J. I. MEAD. 1989. Quaternary zoogeography of the Nearctic Dicrostonyx lemmings. Boreas 18:323-32.

MORLAN, R. E. 1984. Biostratigraphy and biogeography of Quaternary microtine rodents from northern Yukon Territory, Eastern Beringia. Carnegie Museum of Natural History Special Publication 8:184-99.

_____. 1989. Paleoecological implications of Late Pleistocene and Holocene microtine rodents from the Bluefish Caves, northern Yukon Territory. Canadian Journal of Earth Sciences 26:149-56.

PAGE, L. M., AND B. M. BURR. 1991. A field guide to freshwater fishes. Boston: Houghton Mifflin Co.

PRUITT JR., W. O. 1954. Additional remains from under Sleeping Bear Dune, Leelanau County, Michigan. Papers of the Michigan Academy of Sciences, Arts, and Letters 39:253-6.

RALLS, K., AND P. H. HARVEY. 1985. Geographic variation in size and sexual dimorphism of North American weasels. Biological Journal of the Linnaean Society 25:119-67.

RAUSCH, R. L., AND V. R. RAUSCH. 1975. Relationships of the redback vole Clethrionomys rutilus (Pallas) in North America: Karyotypes of the subspecies dawsoni and albiventer. Systematic Zoology 24:163-70.

REPENNING, C. A. 1987. Biochronology of the microtine rodents of the United States. In Cenozoic mammals of North America, ed. M. O. Woodburne. Berkeley: University of California Press.

REPENNING, C. A., E. M. BROUWERS, L. D. CARTER, L. MARINCOVICH, AND T. A. AGER. 1987. The Beringian ancestry of Phenacomys (Rodentia: Cricetidae) and the beginning of the modern Arctic Ocean Borderland biota. U.S. Geological Survey Bulletin 1687:1-31.

REPENNING, C. A., AND F. GRADY. 1988. The microtine rodents of the Cheetah room fauna, Hamilton Cave, West Virginia, and the spontaneous origin of Synaptomys. U.S. Geological Survey Bulletin 1853:1-32.

REPENNING, C. A., D. M. HOPKINS, AND M. RUBIN. 1964. Tundra rodents in a Late Pleistocene fauna from the Tofty Placer district, central Alaska. Arctic 17:177-97.

RICHARDS, R. 1986. Late Pleistocene remains of boreal voles (genera Phenacomys and Clethrionomys) from southern Indiana caves. Proceedings of the Indiana Academy of Science 95:537-46.

RODGERS, A. R., AND M. C. LEWIS. 1985. Diet selection in Arctic lemmings (Lemmus sibiricus and Dicrostonyx groenlandicus): Food preferences. Canadian Journal of Zoology 63:1161-73.

SAVAGE, H. 1994. Prehistoric fauna in a vertical fissure cave in the Niagara Escarpment, Dufferin County, Ontario. In Great Lakes Archaeology and Paleoecology, ed. R. I. McDonald. Waterloo, Ontario: University of Waterloo.

SHOSHANI, J., D. C. FISHER, J. M. ZAWISKIE, S. J. THURLOW, S. L. SHOSHANI, W. S. BENNINGHOFF, AND F. H. ZOCH. 1989. The Shelton mastodon site: Multidisciplinary study of a Late Pleistocene (Twocreekan) locality in southeastern Michigan. University of Michigan Contributions from the Museum of Paleontology 27:393-436.

TUPIKOVA, N. V., G. A. SIDOROVA, AND E. A. KONOVALOVA. 1968. A method of age determination in Clethrionomys. Acta Theriologica 13(8):89-115.

VAN ZYLL DE JONG, C. G. 1983. Handbook of Canadian mammals, Part 1: Marsupials and insectivores. Ottawa: National Museums of Canada.

VON KOENIGSWALD, W., AND L. D. MARTIN. 1984. Revision of the fossil and Recent Lemminae (Rodentia: Mammalia). Carnegie Museum of Natural History Special Publication 9:122-37.

WILSON, R. L. 1967. The Pleistocene vertebrates of Michigan. Papers of the Michigan Academy of Sciences, Arts, and Letters 52:197-257.


Royal Ontario Museum
COPYRIGHT 2004 Michigan Academy of Science Arts & Letters
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2004 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Seymour, Kevin L.
Publication:Michigan Academician
Geographic Code:100NA
Date:Jun 22, 2004
Previous Article:Herpetological assemblages of the Michigan Regional Landscape Ecosystems.
Next Article:Distribution of reptiles and amphibians on the islands of eastern Lake Michigan: summary and analysis.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters