1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 This Paper Explores

Home , Arctic lemming, Singing vole, Taiga vole, Tundra vole

Small mammals and paleovironmental context of the terminal pleistocene and early holocene human occupation of central Alaska

Item Type Article

Authors Lanoë, François B.; Reuther, Joshua D.; Holmes, Charles E.; Potter, Ben A.

Citation Lanoë, FB, Reuther, JD, Holmes, CE, Potter, BA. Small mammals and paleovironmental context of the terminal pleistocene and early holocene human occupation of central Alaska. Geoarchaeology. 2019; 1– 13. https://doi.org/10.1002/gea.21768

DOI 10.1002/gea.21768

Publisher WILEY

Journal GEOARCHAEOLOGY-AN INTERNATIONAL JOURNAL

Download date 07/10/2021 13:55:14

Item License http://rightsstatements.org/vocab/InC/1.0/

Version Final accepted manuscript

Link to Item http://hdl.handle.net/10150/634952 Page 2 of 42

1 2 3 1 SMALL MAMMALS AND PALEOVIRONMENTAL CONTEXT OF THE TERMINAL 4 5 2 PLEISTOCENE AND EARLY HOLOCENE HUMAN OCCUPATION OF CENTRAL 6 3 ALASKA 7 8 4 François B. Lanoëab, Joshua D. Reutherbc, Charles E. Holmesc, and Ben A. Potterc 9 10 5 11 a 12 6 Bureau of Applied Research in Anthropology, University of Arizona, 1009 E S Campus Dr, 13 7 Tucson, AZ 85721 14 bArchaeology Department, University of Alaska Museum of the North, 1962 Yukon Dr, 15 8 16 9 Fairbanks, AK 99775 17 10 cDepartment of Anthropology, University of Alaska, 303 Tanana Loop, Fairbanks, AK 99775 18 19 11 20 21 12 Corresponding author: François Lanoë, [email protected] 22 23 13 24 25 14 Abstract 26 27 15 This paper explores paleoenvironmental and paleoecological information that may be obtained 28 16 from small-mammal assemblages recovered at central Alaska archaeological sites dated to the 29 30 17 Terminal Pleistocene and Early Holocene (14,500-8000 cal B.P.). Small-mammal remains in 31 18 these open-air sites are primarily related to deposition by natural death causes and as such 32 19 provide information on site paleoenvironments and landscape heterogeneity. Their presence 33 20 within archaeological occupations likely relates to anthropogenic disturbance and features that 34 35 21 would have favored burrow construction. The co-occurrence of small-mammal remains and 36 22 archaeological occupations provides a chronological framework of presence in the locality for 37 23 most recorded small mammal species. Small-mammal remains document faunal turnover 38 24 between Pleistocene and Holocene communities. The near-contemporaneity of species that 39 25 strongly differ in their ecological requirements suggest that the Terminal Pleistocene and Early 40 41 26 Holocene in central Alaska was a period of dynamic change that may have been characterized by 42 27 patchy vegetation distribution, rather than the climax communities seen today. In addition to the 43 28 biogeographical value of small-mammal remains, the paleoenvironmental information that they 44 29 provide helps to characterize the ecology of early human settlers in the region and the processes 45 46 30 behind human dispersal in Beringia and the Americas at the end of the Ice Age. 47 48 31 49 32 Keywords: Beringia; microtines; paleoenvironments; Pleistocene; Holocene 50 51 33 52 53 34 54 55 35 Abstract 56 57 58 1 59 60 Page 3 of 42

1 2 3 36 This paper explores paleoenvironmental and paleoecological information that may be obtained 4 5 6 37 from small-mammal assemblages recovered at central Alaska archaeological sites dated to the 7 8 38 Terminal Pleistocene and Early Holocene (14,500-8000 cal B.P.). Small-mammal remains in 9 10 39 these open-air sites are primarily related to deposition by natural death causes and as such 11 12 40 provide information on site paleoenvironments and landscape heterogeneity. Their presence 13 14 15 41 within archaeological occupations likely relates to anthropogenic disturbance and features that 16 17 42 would have favored burrow construction. The co-occurrence of small-mammal remains and 18 19 43 archaeological occupations provides a chronological framework of presence in the locality for 20 21 22 44 most recorded small mammal species. Small-mammal remains document faunal turnover 23 24 45 between Pleistocene and Holocene communities. The near-contemporaneity of species that 25 26 46 strongly differ in their ecological requirements suggest that the Terminal Pleistocene and Early 27 28 29 47 Holocene in central Alaska was a period of dynamic change that may have been characterized by 30 31 48 patchy vegetation distribution, rather than the climax communities seen today. In addition to the 32 33 49 biogeographical value of small-mammal remains, the paleoenvironmental information that they 34 35 50 provide helps to characterize the ecology of early human settlers in the region and the processes 36 37 38 51 behind human dispersal in Beringia and the Americas at the end of the Ice Age. 39 40 41 52 42 43 44 53 1. Introduction 45 46 47 54 Paleoenvironmental studies have proved fundamental in characterizing and understanding the 48 49 55 major palaeoecological and biogeographical events associated with the Terminal Pleistocene and 50 51 56 early Holocene (14,500-8000 cal B.P.) of Beringia, when Beringia served as a gateway for 52 53 54 57 dispersal of animal and plant species between northeastern Eurasia and northwestern North 55 56 57 58 2 59 60 Page 4 of 42

1 2 3 58 America. These migrants, including human populations, contributed to the emergence of 4 5 6 59 radically new biotic communities. 7 8 9 60 In particular, human colonization of Beringia during the Terminal Pleistocene corresponds to one 10 11 61 of the major steps of the story of human dispersal around the globe (Hoffecker et al., 2016). 12 13 62 Archaeological and paleogenomic evidence together indicate that Beringian settlers originated 14 15 63 from northeastern Eurasian populations (Llamas et al., 2016; Moreno-Mayar et al., 2018; Sikora 16 17 18 64 et al., 2019) and became archaeologically visible during the Bølling-Allerød chronoperiod 19 20 65 (14,500-12,900 cal B.P.) (Goebel and Buvit, 2011; Meltzer, 2009; Potter et al., 2018). The 21 22 66 colonization of Beringia corresponds to the first evidence for recolonization of latitudes above 23 24 25 67 60°N after the Last Glacial Maximum and set the stage for the colonization of the Americas in 26 27 68 the subsequent millennia (Goebel et al., 2008; Potter et al., 2018; Raghavan et al., 2015). 28 29 30 69 Dispersal of people in and through Beringia was part of and took place during a larger biotic 31 32 70 turnover. Paleobiological studies have shown a biome shift from drier, steppe-like environments 33 34 35 71 over most of Beringia, to wetter biomes characterized by tundra, along the coast, and forested 36 37 72 environments, in the interior (Anderson et al., 2004; Blinnikov et al., 2011; Jones and Yu, 2010; 38 39 73 Oswald et al., 2014; Wooller et al., 2018). Megafauna communities were accordingly affected 40 41 74 with a decrease in number of species, the extinction of many grazing herbivores and 42 43 44 75 hypercarnivorous predators (Guthrie, 2006; Mann et al., 2015), and their eventual replacement 45 46 76 by dominantly mixed feeding and browsing herbivores and omnivorous and hypocarnivorous 47 48 77 predators (Guthrie, 1982; Lanoë et al., 2017). 49 50 51 78 Causes of ecological change in Beringia at the end of the Pleistocene ultimately relate to global 52 53 54 79 climate change associated with the end of the Ice Age and the beginning of the current 55 56 80 interglacial period (Mann et al., 2018). More proximate drivers of environmental change may 57 58 3 59 60 Page 5 of 42

1 2 3 81 have included regional climatic factors such as changes in atmospheric circulation and 4 5 6 82 precipitation (Guthrie, 2001). In parallel, invasion and extinction of keystone ecological species 7 8 83 such as ecological engineers and apex predators, including humans, may have triggered sudden 9 10 84 biome shifts (Guthrie, 1984; Lanoë et al., 2017; Surovell et al., 2005; Zimov et al., 1995). 11 12 13 85 Research in Beringia has long focused on characterizing paleoenvironments and their evolution 14 15 86 at a regional level, relying on proxies that provide information on past environments at a low 16 17 18 87 spatial resolution, such as pollen (Edwards et al., 2000) and megafauna fossils (Guthrie, 1990, 19 20 88 1968a). A different set of proxies that provide paleoenvironmental information at a higher spatial 21 22 89 resolution, such as macrobotanical remains (Zazula et al., 2007), insects (Elias, 2001), and 23 24 25 90 sedimentological, pedological, biogeochemical, and genetic analyses of lacustrine and terrestrial 26 27 91 sediments (Graham et al., 2016; Kaufman et al., 2016; Kokorowski et al., 2008; Reuther, 2013; 28 29 92 Vachula et al., 2019; Willerslev et al., 2014), has shown the high degree of environmental 30 31 32 93 heterogeneity in Beringia during the Terminal Pleistocene and Early Holocene. 33 34 35 94 Small-mammal remains constitute environmental proxies that, in contrast to large mammals, 36 37 95 have seen considerably less attention in Alaska since pioneering studies of Repenning et al. 38 39 96 (1964) and Guthrie (1968b). Because of their diminutive size, small-mammal remains don’t 40 41 97 preserve well, are often overlooked in geological or archaeological contexts, and/or require a 42 43 44 98 time-intensive sampling methodology. Even when recovered, such remains are seldom directly 45 46 99 dated, particularly until recent methodological improvements in the dating of small bone 47 48 100 samples. Instead, in Alaska their age has usually been inferred from stratigraphic association 49 50 51 101 with geological layers, particularly in karstic contexts (Endacott, 2008; Georgina, 2001; 52 53 102 Harington and Cinq-Mars, 2008; Morlan, 1989), and as a result they offer a very low 54 55 103 chronological resolution. 56 57 58 4 59 60 Page 6 of 42

1 2 3 104 Small-mammal remains have a high potential for providing paleoenvironmental information at 4 5 6 105 both local and regional scales. In contrast to most large mammals, small mammals have small 7 8 106 home ranges (typically below 0.5 Ha for herbivore species under 1.0 Kg [Haskell et al. 2002]) 9 10 107 and are tied to micro-environments with specific hydric regimes and/or vegetation communities 11 12 108 (Merritt, 2010). This high-resolution spatial paleoenvironmental information allows for a better 13 14 15 109 understanding of the local environmental setting of archaeological sites. Changes in the small- 16 17 110 mammal assemblage at a locality over time also indicate how regional climatic and 18 19 111 environmental change impacts the local setting of a geological or archaeological site. At a larger 20 21 22 112 scale, distribution of small-mammal communities across different localities provides information 23 24 113 on the different components of a landscape and its homogeneity. This type of 25 26 114 paleoenvironmental information is crucial for developing models of settlement and land use, and 27 28 29 115 ultimately to address issues related to human ecology, economy, and geography. 30 31 32 116 In this paper we present and discuss paleoenvironmental information obtained from small 33 34 117 mammals remains recovered from archaeological sites in the middle Tanana valley of Central 35 36 118 Alaska dated to the Terminal Pleistocene and Early Holocene. The archaeological landscape of 37 38 119 the middle Tanana valley is the best documented within Beringia and provides strong evidence 39 40 41 120 for early dispersal in and through Beringia (Lanoë et al., 2018b; Potter, 2011; Potter et al., 2013). 42 43 121 Paleoenvironmental information from small mammals contributes to establishing environmental 44 45 122 contexts for these early settlers and subsequent populations, and its implications in understanding 46 47 48 123 their economy and lifeways. 49 50 51 124 52 53 54 125 2. Materials & Methods 55 56 57 58 5 59 60 Page 7 of 42

1 2 3 126 Small-mammal remains were recovered as part of archaeological excavations at sites in the 4 5 6 127 middle Tanana valley (Table 1, Figure 1). Studied materials were restricted to layers dating to 7 8 128 the Terminal Pleistocene (14,500-11,700 cal B.P.) and Early Holocene (11,700-8000 cal B.P.). 9 10 129 Materials were collected from sediment screened at a mesh of 3.2 mm, in 5 or 10 cm levels. Due 11 12 130 to their general rarity in Alaska aeolian sediments (Guthrie, 1968b), and because the main 13 14 15 131 purpose of fieldwork was to retrieve archaeological materials, no specific sampling method was 16 17 132 designed to systematically retrieve small mammal skeletal materials. Exceptions included small- 18 19 133 mammal nests encountered whole during excavation, which were collected and screened in the 20 21 22 134 laboratory in nested screens to a size of 63 μm. 23 24 25 135 Small mammals (<1.0 Kg) in central Alaska include vole and lemmings (Cricetidae), jumping 26 27 136 mice (Dipodidae), squirrels (Sciuridae), pikas (Ochotonidae), shrews (Soricidae), bats 28 29 137 (Verspertilionidae), and carnivorans (Mustelidae) (MacDonald and Cook, 2009). Species 30 31 32 138 identification was conducted with the help of reference collections in the Mammal Collection at 33 34 139 the University of Alaska Museum of the North in Fairbanks, and the Stanley J. Olsen Laboratory 35 36 140 of Zooarchaeology at the Arizona State Museum of the University of Arizona in Tucson. Species 37 38 141 identification of skeletal remains of Arvicolinae (Cricetidae) was limited to the lower first molar 39 40 41 142 (m1), which is diagnostic to species for most of the Arvicolinae species expected in the higher 42 43 143 latitudes of North America (Semken and Wallace, 2002). Exceptions include the m1 of Microtus 44 45 144 miurus and Microtus pennsylvanicus that are of similar shape and can overlap in size (Morlan, 46 47 48 145 1989). 49 50 51 146 Remains were quantified using Number of Identified Specimens or NISP (Lyman, 2008), with 52 53 147 quantification of skeletal remains of dipodids limited to the m1 for the sake of consistency with 54 55 148 the similar-sized cricetids. No shrews, bats, or canivoran remains identifiable to species were 56 57 58 6 59 60 Page 8 of 42

1 2 3 149 recorded. The presence of arctic ground squirrel (Urocitellus parryi) was assessed by the 4 5 6 150 presence of their skeletal remains as well as their characteristic, ~10 cm-wide, burrow or 7 8 151 krotovina (Zazula et al., 2007). Observations of bone surface modifications were conducted with 9 10 152 a 20X hand lens on a sample of 305 cranial and postcranial elements, including 26 cricetid and 11 12 153 dipodid m1, accounting for 37% of the total NISP as defined above. 13 14 15 154 16 17 18 155 3. Results 19 20 21 156 Skeletal remains of small mammals were recovered at five of the six archaeological sites studied 22 23 24 157 for the time interval between 14,500 and 8000 cal B.P. (Table 2). Numbers of recovered 25 26 158 specimens vary greatly between sites, ranging from a total absence of vole, lemming, and 27 28 159 jumping mice at the Keystone Dune site, to as many as 6 per square meter at the Bachner site. 29 30 31 160 On average, identified vole, lemming, and jumping mice specimens for the entire set of sites 32 33 161 numbered 0.26 per excavated m2, or 0.04 per m2 per millennium of deposition. 34 35 36 162 Density of identified specimens is significantly lower than typically observed in karstic contexts. 37 38 163 As a comparison, at Lime Hills Cave in southwestern Alaska 395 vole, lemming, and jumping 39 40 164 mice remains were recorded with similar methods (3.2 mm mesh size, quantification of m1 only) 41 42 2 43 165 over 17 m of excavation and in stratigraphic layers spanning 10,000 to 20,000 years (Georgina, 44 45 166 2001), or a density of 1.16 to 2.32 identified specimen per m2 per millennium of deposition. In 46 47 167 contrast, studied loess sections near Fairbanks in central Alaska have yielded 31 vole, lemming, 48 49 3 50 168 and jumping mice lower first molars in about 55 tons of Late Pleistocene loess, or about 60 m , 51 52 169 using a mesh size of 0.6 mm (Guthrie, 1968b). Assuming loess deposition rates of 1 to 5 cm per 53 54 170 century (Lanoë et al., 2018b), the Fairbanks loess sections provide densities of 0.05 to 0.26 55 56 57 58 7 59 60 Page 9 of 42

1 2 3 171 identified specimens per m2 per millennium of deposition. These densities are of the same order 4 5 6 172 of magnitude as sites in the middle Tanana valley. 7 8 9 173 Species identified in this study include the brown lemming (Lemmus trimucronatus), the 10 11 174 meadow vole (Microtus pennsylvanicus), the red-backed vole (Myodes rutilus), the singing vole 12 13 175 (Microtus miurus), the taiga vole (Microtus xanthognathus), the tundra vole (Microtus 14 15 176 oeconomus), the jumping mouse (Zapus hudsonius), and the arctic ground squirrel (Urocitellus 16 17 18 177 parryii) (Figure 2). Several specimens could not be distinguished between meadow and singing 19 20 178 vole. Other taxa were not identified. 21 22 23 179 Three of the species represented (arctic ground squirrel, brown lemming, singing vole) do not 24 25 180 occur in the lowlands of the middle Tanana valley today (MacDonald and Cook, 2009, 2001). 26 27 28 181 Five similar-sized species that do occur in the middle Tanana valley lowlands today, including 29 30 182 the bog lemming (Synaptomys borealis), the flying squirrel (Glaucomys sabrinus), the long- 31 32 183 tailed vole (Microtus longicaudus), the muskrat (Ondatra zibethicus), and the red squirrel 33 34 35 184 (Tamiasciurus hudsonicus), are not represented in the assemblage. The arctic lemming 36 37 185 (Dicrostonyx torquatus), recorded in undated Late Pleistocene loess of the lower Tanana valley 38 39 186 (Guthrie, 1968b; Repenning et al., 1964), is not represented either. 40 41 42 187 A majority of vole and lemming m1 were recovered in (63.5%) or at the same level (24.3%) than 43 44 188 anthropogenic features (pits or hearths), with only a few (12.2%) specimens “floating” in 45 46 47 189 between archaeological occupations (Figure 2). At the Upward Sun River site, 54.5% of the 48 49 190 recorded vole and lemming specimens were recovered in a pit associated with a cremation hearth 50 51 191 and a burial that represents < 1% of the total area excavated at the site (Potter et al., 2014). At the 52 53 54 192 Hollembaek site, 75.0% of the recorded vole and lemming specimens were recovered in an 55 56 193 anthropogenic pit feature probably associated with trash management and that represents ~50% 57 58 8 59 60 Page 10 of 42

1 2 3 194 of the total area excavated at the site (Reuther and Lanoë, 2018). At the Swan Point site, all 4 5 6 195 recorded vole and lemming specimens were recovered in association, within Cultural Zone 4b, 7 8 196 with two large hearths filled with mammoth (Mammuthus primigenius) ivory fragments and 9 10 197 mammoth and other animal bones (Lanoë and Holmes, 2016). All recorded vole and lemming 11 12 198 specimens at the Bachner and Cook sites, as well as 40.0% and 31.8% of those recorded at the 13 14 15 199 sites of Hollembaek and Upward Sun River (excluding those recorded in the pit features), 16 17 200 respectively, were recovered at the same stratigraphic level as archaeological occupations 18 19 201 (Figure 2). Lastly, the Keystone Dune site, characterized by an ephemeral archaeological 20 21 22 202 occupation with no anthropogenic feature and by comparatively high sedimentation rates (Lanoë 23 24 203 et al., 2018a), yielded no small mammal remains (m1 or other). 25 26 27 204 Jumping mice specimens were generally recovered deep in the stratigraphy (Figure 2). At the 28 29 205 Swan Point site where they are most prevalent, jumping mice remains were recovered as 30 31 32 206 complete skeletons including both cranial and post-cranial remains, often in nests consisting of 33 34 207 dry vegetation (Lanoë and Holmes, 2016). Like vole and lemming remains, jumping mice m1 35 36 208 and associated skeletons were primarily (90.5%) found in association with anthropogenic 37 38 209 features. At the Swan Point site, 87.5% of jumping mice specimens were recovered within 39 40 41 210 Cultural Zone 4b, including 38.1% in association with the same hearths where vole and lemming 42 43 211 remains were found. At the Hollembaek site, all jumping mice specimens were recovered 44 45 212 immediately underneath the same pit where vole and lemming remains were dominantly found. 46 47 48 213 Evidence of ground squirrel activity in the form of skeletal remains or burrows was recorded at 49 50 51 214 all sites except Keystone Dune (Figure 2). Their absence at Keystone Dune may result from the 52 53 215 homogeneous sand stratigraphy that hinders recognition of possible burrow outline and filling 54 55 216 (Reuther et al., 2016). Where it was recorded, evidence of ground squirrel activity occurred in 56 57 58 9 59 60 Page 11 of 42

1 2 3 217 the lower portions of the stratigraphy in loess and/or sand, down to bedrock, in layers generally 4 5 6 218 dated to 12,000 cal B.P. or older (Figure 2). In addition to burrows, ground squirrel skeletal 7 8 219 remains were recovered at all sites except Keystone Dune (Lanoë and Holmes, 2016; Potter et 9 10 220 al., 2011). 11 12 13 221 Preservation of small-mammal remains (cranial and postcranial) was on average better than for 14 15 222 large mammals. Taphonomic observations at the Swan Point site show a scarcity of bone 16 17 18 223 breakage and a low impact of weathering and surface dissolution on the vole, lemming, and 19 20 224 jumping mice assemblage, compared to the large mammal assemblage, while preservation of 21 22 225 ground squirrel remains generally shows intermediate stages between those recorded for vole, 23 24 25 226 lemming, and jumping mice remains, and those recorded for large mammal remains (reported in 26 27 227 Lanoë and Holmes, 2016). At the Upward Sun River site, a high proportion of the vole, 28 29 228 lemming, jumping mice, as well as ground squirrel remains recovered in the hearth/burial pit 30 31 32 229 were combusted (reported in Potter et al., 2014, 2011). 33 34 35 230 Five radiocarbon dates were obtained on small-mammal remains or on materials associated with 36 37 231 small-mammal features from the Swan Point site (Holmes et al., 1996; Lanoë and Holmes, 38 39 232 2016). A jumping mouse element recovered in a nest 80-85 cm below surface was dated to 5590- 40 41 233 5330 cal B.P. (Beta-186682), in a layer otherwise dated to ~14,000 cal B.P. (Figure 2). Within 42 43 44 234 the same layer, charcoal fragments associated with 10-15 cm diameter (ground squirrel size) 45 46 235 burrows 75-80 cm below surface were dated to, 13,600-13,340 (Beta-71372), and 13,610-13,300 47 48 236 cal B.P. (Beta-56667), respectively; and a charcoal fragment associated with a ~4 cm diameter 49 50 51 237 (vole, lemming, or jumping mouse size) burrow 75-80 cm below surface was dated to 9540-9310 52 53 238 (Beta-215328) cal B.P. (Figure 3). A charcoal fragment recovered in a burrow ~4 cm diameter 54 55 56 57 58 10 59 60 Page 12 of 42

1 2 3 239 50-55 cm below surface was dated to 5580-5310 cal B.P. (Beta-190577), in a layer otherwise 4 5 6 240 dated to ~12,000 cal B.P. (Figure 2). 7 8 9 241 10 11 12 242 4. Discussion 13 14 243 4.1. Modes of Accumulation 15 16 17 244 Small-mammal paleontological assemblages in the more commonly studied karstic contexts are 18 19 20 245 primarily created by the action of avian and mammalian predators who sample small mammals 21 22 246 in the vicinity of their denning/nesting site (Andrews, 1990; Morlan, 1989). In contrast, large 23 24 247 mammal assemblages in open-air archaeological sites are primarily created by human predators 25 26 27 248 who discard skeletal elements after consumption of the edible portions of a carcass. 28 29 30 249 The settings of the sites studied are unconducive to predator denning or nesting with a lack of 31 32 250 rock outcrops and a paucity of large trees for the time period considered (Bigelow and Powers, 33 34 251 2001). The low density of small-mammal remains and their general lack of breakage makes them 35 36 37 252 unlikely to be related to animal predation. In parallel, the low amount of weathering or surface 38 39 253 dissolution suggests that small-mammal remains were not left on the surface or near surface 40 41 254 contexts for extended periods for time like those of large mammals, but instead were deposited 42 43 255 under the surface. 44 45 46 256 All the small mammal species recorded are at least seasonally fossorial and use burrows for 47 48 49 257 hibernation, shelter, nest, and/or storage; non-fossorial species such as the red squirrel are not 50 51 258 represented (MacDonald and Cook, 2009). Taphonomic observations, along with the recovery of 52 53 259 nests much younger than the surrounding sediments, and the prevalence of burrowing in the 54 55 56 260 sediment, suggest a large proportion of the small-mammal remains studied were deposited in 57 58 11 59 60 Page 13 of 42

1 2 3 261 burrows, in which case the cause of death of the individuals would have been primarily related to 4 5 6 262 thermal or nutritional stress (Morrison and Galster, 1975). The ground squirrel assemblage 7 8 263 displays a taphonomic signature closer to the one observed for large mammal remains. Ground 9 10 264 squirrels are much larger-bodied than vole, lemming, and jumping mice, and in some cases may 11 12 265 have been acquired and consumed by the people that produced the archaeological assemblages 13 14 15 266 (see Potter et al., 2011; Yesner, 1996). 16 17 18 267 The location of recovery of small-mammal remains at the middle Tanana sites suggests that the 19 20 268 location of small-mammal burrows was strongly influenced by pre-existing anthropogenic 21 22 269 features and accumulation of artifacts. Small mammals may have been attracted by materials left 23 24 25 270 by people, such as ashes, animal bones and tissues, or feces, that provided food and nutrients. 26 27 271 Sediment disturbance by people may also have facilitated burrow excavation. Large branches 28 29 272 and roots contributed to the structure of modern Alaska burrows (Wolff and Lidicker, 1980); 30 31 32 273 discarded large archaeological artifacts and anthropogenic features may have provided the same 33 34 274 function in loess settings otherwise void of inclusions and with few or no large trees. The large 35 36 275 quantity of taiga vole remains in the anthropogenic feature at the Hollembaek site probably 37 38 276 reflects the local establishment of a taiga vole colony in these disturbed sediments shortly after 39 40 41 277 human abandonment of the site. Likewise, at the Upward Sun River site vole and lemming 42 43 278 skeletal remains, along with numerous ground squirrel remains, are located in and within 10 cm 44 45 279 underneath a human cremation hearth, above an underlying human burial (Potter et al., 2014), 46 47 48 280 and may reflect burrows excavated in the burial pit after initial abandonment of the site by 49 50 281 people. 51 52 53 282 In some cases, anthropogenic disturbance must also have impacted the preservation of small 54 55 283 mammal remains. Hollembaek site anthropogenic feature includes, in addition to taiga vole 56 57 58 12 59 60 Page 14 of 42

1 2 3 284 remains, those of vole species (red-backed vole, singing vole) more typical of Late Pleistocene 4 5 6 285 layers at this and other sites (Figure 2). Those specimens likely relate to the reworking of Late 7 8 286 Pleistocene sediments during the construction of the feature. In parallel, combustion traces 9 10 287 observed on the vole and ground squirrel remains in the Upward Sun River burial pit may 11 12 288 correspond to indirect thermal alteration (Bennett, 1999; Stiner et al., 1995) from the subsequent 13 14 15 289 superficial excavation of the pit and construction of the cremation hearth during reoccupation of 16 17 290 the site by people (Potter et al., 2011), rather than to human consumption. 18 19 20 291 The creation of the vole, lemming, and jumping mice assemblage by causes mostly unrelated to 21 22 292 predation makes it less likely to provide biased paleoenvironmental information than an 23 24 25 293 assemblage created by mammalian or avian predators. Assemblages created by predators may 26 27 294 reflect as much the predators’ dietary and habitat preferences than the range of micro- 28 29 295 environments present at and around the sampling site (Andrews, 1990; Grayson, 1981). In 30 31 32 296 contrast, assemblages created mainly by burrowing are composed of species and individuals that 33 34 297 actually occupied the local site setting. Anthropogenic features may present avenues for bias by 35 36 298 favoring species of small mammals that rely on ground disturbance, with for instance taiga vole 37 38 299 in the Hollembaek site despite their being relatively uncommon in the region today (MacDonald 39 40 41 300 and Cook, 2001). 42 43 44 301 Small-mammal burrowing depth varies across species. Recorded species of vole and lemming 45 46 302 are characterized by shallow burrows that in Alaska are often located at the junction between 47 48 303 litter and mineral soil (Wolff and Lidicker, 1980), with average reported depths of 10 to 15 cm 49 50 51 304 below surface (Bee and Hall, 1956; Cole and Wilson, 2010; Wolff and Lidicker, 1980). Detailed 52 53 305 burrow measurements for Kentucky prairie voles (Microtus ochrogaster) indicate that 93% of 54 55 306 burrows volume occurs within the top 15 cm below surface (Davis and Kalisz, 1992). Jumping 56 57 58 13 59 60 Page 15 of 42

1 2 3 307 mice and ground squirrel, in contrast, are known to excavate deep burrows down to 60 and over 4 5 6 308 100 cm below surface, respectively (Nowak, 1999; Whitaker, 1972). 7 8 9 309 10 11 12 310 4.2. Methods of Study and Chronological Control 13 14 311 Differences in the modes of accumulation of small mammal remains between karstic and open- 15 16 17 312 air sites is readily seen in the relative density of documented remains, with karstic sites 18 19 313 preserving 2 to 3 orders of magnitude more remains than open-air sites. On the other hand, 20 21 314 similar rates of recovery between archaeological sites of the middle Tanana valley and Fairbanks 22 23 24 315 loess studies suggest that a screen mesh size of 3.2 mm, however incomplete, provides a fair 25 26 316 chance of recovery of vole-sized remains that is comparable to studies conducted with a smaller 27 28 317 (0.6 mm) screen mesh size. It is also more readily manageable in archaeological excavations 29 30 31 318 which primary goal is not to retrieve small mammal remains. The lack of recovered shrew 32 33 319 remains reflects the limitations of this sampling strategy for very small mammals with slender 34 35 320 and minute elements, in contrast to larger and more robust rodent m1, particularly when still 36 37 321 encased within mandible bones. 38 39 40 322 The fossorial mode of accumulation of the small-mammal assemblage impacts any date 41 42 43 323 estimation for their recovered skeletal remains. The jumping mouse dated to 5590-5330 cal B.P. 44 45 324 at the Swan Point site was located ~50 cm below the sedimentological layers deposited at that 46 47 325 time, which falls within the range of known burrow depths for jumping mice. The charcoal 48 49 50 326 fragments dated to 5580-5310 and 9540-9310 cal B.P. were located ~20 and ~35 cm, 51 52 327 respectively, below layers deposited at that time, which also falls within the depth range of 53 54 328 jumping mice burrows. These charcoal fragments provide a maximum rather than actual age for 55 56 57 58 14 59 60 Page 16 of 42

1 2 3 329 jumping mice populations, as individual mice likely incorporated charcoal from underlying 4 5 6 330 layers in the infilling of their burrows. 7 8 9 331 Date estimates therefore need to take into account the stratigraphic layers in which small- 10 11 332 mammal remains were recovered, radiocarbon dates, and potential burrow depths, ranging from 12 13 333 up to 20 cm for vole and lemming, and up to 60 cm for jumping mice (Figure 2). Despite these 14 15 334 limitations, open-air sites provide a more stable fossil environment than karstic sites, with higher 16 17 18 335 sedimentation rates and comparatively little bioturbation and stratigraphic mixing, much 19 20 336 reducing risks of contamination of asynchronous mammal communities. Dates for fossil 21 22 337 individuals can reliably be estimated within a couple millennia as opposed to entire geological 23 24 25 338 epochs (Table 4). The high occurrence of small mammal remains in anthropogenic features and 26 27 339 archaeological occupations in general also provide avenues for developing targeted sampling 28 29 340 protocols in these restricted areas. Open-air sites are routinely studied as part of archaeological 30 31 32 341 excavations and while methods used by archaeologists are not suitable for very small mammals 33 34 342 (e.g. shrews), archaeological research can contribute much to the biogeography and evolutionary 35 36 343 history of “larger small mammals” such as vole, lemming, and mice. 37 38 39 344 40 41 42 345 4.3. Biogeography and Regional Paleoenvironments 43 44 45 346 Recorded species of small mammals through the Pleistocene-Holocene transition reflect species 46 47 48 347 turnover that accompanied biome changes in Beringia/Alaska. Some species are only recorded 49 50 348 during the Pleistocene between 14,000 and ca. 12,000 cal B.P. (brown lemming, singing vole, 51 52 349 ground squirrel), others are only recorded in the Holocene after ca. 10,000-8000 cal B.P. 53 54 350 (jumping mouse, meadow vole), and some span both the Pleistocene and Holocene (red-backed 55 56 57 58 15 59 60 Page 17 of 42

1 2 3 351 vole, taiga vole, tundra vole). The presence of singing vole in a layer dated to 8000 cal B.P. at 4 5 6 352 the Hollembaek site likely relates to anthropogenic disturbance of Late Pleistocene layer (see 7 8 353 4.1.). 9 10 11 354 The presence of singing vole and brown lemming in Pleistocene contexts concurs with the small 12 13 355 mammal record in Full Glacial sediments in central Alaska, where these two species, along with 14 15 356 the arctic lemming, are dominant (Guthrie, 1968b; Repenning et al., 1964); as well as with the 16 17 18 357 karstic records from Lime Hills and Bluefish caves (Georgina, 2001; Morlan, 1989). Abundance 19 20 358 of singing vole in particular has led Guthrie to associate the species with the steppe-like 21 22 359 graminoid dominated biome that covered Beringia during much of the Late Pleistocene (Guthrie, 23 24 25 360 1968b). Both species are characteristic of open environments and are found today in alpine or 26 27 361 arctic tundra only (Bee and Hall, 1956; MacDonald and Cook, 2009). Likewise, ground squirrels 28 29 362 in the middle Tanana valley seem to have been a Pleistocene occurrence only until ca. 12,000 cal 30 31 32 363 B.P. Evidence for a later presence is lacking, but they may have persisted in local patches 33 34 364 exposed to high rates of disturbance until the Early Holocene. 35 36 37 365 Direct dates as well as interpretations based on burrowing habits suggest that the documented 38 39 366 jumping mice may be of Middle Holocene in age. They are today characteristic of shrubby or 40 41 367 herbaceous riparian zones (MacDonald and Cook, 2009) and along with the long-tailed and 42 43 44 368 meadow voles with whom they share ecological requirements, as well as the bog lemming, they 45 46 369 do not have any known Pleistocene occurrence in Beringia (MacDonald and Cook, 2009; 47 48 370 Morlan, 1989) and are considered based on genetic studies to be post-glacial immigrants (Conroy 49 50 51 371 and Cook, 2000; Cook et al., 2004; Malaney et al., 2013). Given these phylogeographic and 52 53 372 paleontological premises, as well as the remainder of the local record for the singing vole, 54 55 373 undetermined singing/meadow vole specimens recorded in ca. 8000 layers at the Hollembaek 56 57 58 16 59 60 Page 18 of 42

1 2 3 374 site are likely to be meadow vole, while similar specimens recorded in ca. 13,000-11,000 layers 4 5 6 375 at Upward Sun River are likely to be singing vole. 7 8 9 376 Species present throughout the period studied have known Pleistocene and Holocene occurrences 10 11 377 in Beringia (Georgina, 2001; Morlan, 1989) and are considered to have persisted in Beringia 12 13 378 throughout the last Glacial-Interglacial cycle (Conroy and Cook, 2005; Cook et al., 2004; 14 15 379 Galbreath and Cook, 2004). They are overall more generalist in their habitat requirements. The 16 17 18 380 tundra vole occurs in habitats as diverse as tundra, shrubland, meadow, and bogs, but does tend 19 20 381 to occur in habitats with little canopy (MacDonald and Cook, 2009). The red-backed and taiga 21 22 382 vole occur in habitats as diverse as boreal forest, tundra, shrubland, and meadow, but do tend to 23 24 25 383 occur in areas where trees are present (MacDonald and Cook, 2009). Indeed, the red-backed vole 26 27 384 is by far the dominant species of vole in the middle Tanana valley lowland boreal forest today 28 29 385 (MacDonald and Cook, 2001). 30 31 32 386 Overall, indicator small mammal species indicate environment openness as late as 11,000 cal 33 34 35 387 B.P. (Table 4), as well as the presence of trees as early as 14,000 cal B.P. Pollen and 36 37 388 macrobotanical records for the middle Tanana valley indicate that cottonwood/aspen Populus 38 39 389 spp. trees were present as early as 14,000 cal B.P. and white spruce Picea glauca by 10,000 cal 40 41 390 B. P. (Bigelow and Powers, 2001; Reuther et al., 2016). Trees at that time may have been sparse 42 43 44 391 on the landscape in the form of stands or in open woodlands interspersed with patches of open 45 46 392 grassland, rather than as the close-canopy forests characteristic of the current central Alaska 47 48 393 lowlands (Anderson and Brubaker, 1994; Bigelow, 1997; Bigelow and Powers, 2001). 49 50 51 394 Soils associated with mature boreal forest did not develop in the region until 8000 cal B.P. 52 53 54 395 (Dilley, 1998; Reuther, 2013). Likewise, carbon isotopic records for moose Alces sp. indicate 55 56 396 their use of a relatively open habitat until 8000 cal B.P., after which point their carbon isotope 57 58 17 59 60 Page 19 of 42

1 2 3 397 signature became more similar to ratios from boreal forest settings (Guthrie, 2006; Lanoë et al., 4 5 6 398 2017). The occurrence of jumping mice and meadow vole around that time may reflect a 7 8 399 biogeographic association with the boreal forest and a dispersal at that time from North America. 9 10 400 Species characteristic of old-growth forests such as flying squirrel and red squirrel (MacDonald 11 12 401 and Cook, 2009) may have arrived at that time too. Though absent in the record, they are not 13 14 15 402 fossorial and may have been present on the landscape without becoming part of the 16 17 403 paleontological assemblage. 18 19 20 404 21 22 23 405 4.4. Sites Paleoenvironmental Settings and Archaeological Implications 24 25 26 406 The high diversity of small mammal species observed in the middle Tanana valley during the 27 28 407 Terminal Pleistocene and Early Holocene points to the spatial and/or temporal diversity of 29 30 31 408 environments that were present at that time, including grasslands, wetlands, and open woodlands. 32 33 409 The association of species with different habitat requirements, such as singing vole and brown 34 35 410 lemming, may reflect the local contemporaneous presence of different micro-habitats close to the 36 37 411 site. Alternatively, accumulation of small-mammal remains by burrowing may have resulted in 38 39 40 412 palimpsests related to small-mammal populations distant by as much as several centuries, where 41 42 413 small-mammal remains recovered in the same layers may relate to successive communities that 43 44 414 did not overlap in their use of the site. In this case, the co-occurrence of species using different 45 46 47 415 micro-environments may indicate high landscape instability and rapid local environmental shifts. 48 49 50 416 Paleoenvironmental archives in the middle Tanana valley indicate that the Terminal Pleistocene 51 52 417 and Early Holocene were characterized by a dynamic landscape (Mann et al., 2018). High rates 53 54 418 of sediment input from aeolian or alluvial sources, frequent wildfires, megaherbivore grazing and 55 56 57 58 18 59 60 Page 20 of 42

1 2 3 419 trampling, as well as lagged dispersal of tree species all contributed to maintain the vegetation at 4 5 6 420 early successional stages, with little soil development until 8,000 cal B.P. (Bigelow and Powers, 7 8 421 2001; Reuther, 2013; Reuther et al., 2016). High rates of disturbance may have triggered rapid 9 10 422 and frequent local vegetation shifts to which small-mammal populations, highly dependent on 11 12 423 local environmental conditions, would have then adjusted by moving their home range. 13 14 15 424 The brown lemming is a species characteristic of wet environments, and is today primarily found 16 17 18 425 in patches of wet tundra (Bee and Hall, 1956; MacDonald and Cook, 2009). Its presence at the 19 20 426 Swan Point site ca. 14,000 cal B.P. and at the Upward Sun River site ca. 11,000 cal B.P. (Table 21 22 427 4) indicates the presence of wet micro-habitats in the vicinity of the sites. Both sites are currently 23 24 25 428 located above muskegs with thick peat development and extant and desiccated ponds (Dilley, 26 27 429 1998; Reuther, 2013). The archaeological assemblages in layers dated to 14,000-12,000 cal B.P. 28 29 430 include numerous remains of waterfowl such as swan, geese, and ducks (Lanoë and Holmes, 30 31 32 431 2016; Potter et al., 2011, 2008), as do nearby contemporary archaeological sites (Yesner, 1996). 33 34 432 The presence of brown lemming at the sites suggests that the wetlands currently located nearby 35 36 433 may have started developing as early as the Terminal Pleistocene as ponding and slack water 37 38 434 environments between dunes and in floodplains (Reuther, 2013). Waterfowl processed and 39 40 41 435 consumed by people may thus have been acquired locally (Potter et al., 2008). 42 43 44 436 The singing vole is a species characteristic of dry environments, and is today primarily found in 45 46 437 patches of well-drained tundra (Bee and Hall, 1956; MacDonald and Cook, 2009). The presence 47 48 438 of the singing vole at several sites of the Middle Tanana valley between 13,000 and 8,000 cal 49 50 51 439 B.P. (Table 4) suggests the persistence during the Late Glacial and Early Holocene of local 52 53 440 conditions reminiscent of the dry glacial steppes. Singing vole may have been particularly at ease 54 55 441 in xeric, unstable patches with high aeolian sediment deposition contributing to maintain 56 57 58 19 59 60 Page 21 of 42

1 2 3 442 vegetation in an early successional stage (Dilley, 1998; Reuther, 2013). Singing vole is indeed 4 5 6 443 present in layers dated to ~13,000-11,000 cal B.P. at the sites of Upward Sun River and Cook 7 8 444 (Table 4), which have yielded some of the highest known loess deposition rates of the region 9 10 445 (Reuther, 2013). Likewise, arctic ground squirrels are usually found in open areas with early 11 12 446 successional vegetation, primarily in the tundra, but also in patches within the lowlands that 13 14 15 447 maintain early successional vegetation due to frequent disturbance, such as flood plains (Bee and 16 17 448 Hall, 1956; MacDonald and Cook, 2009). 18 19 20 449 Regional conditions that favored environmental diversity and high alpha and/or beta small 21 22 450 mammal diversity likely favored human settlement as well. People living in the middle Tanana 23 24 25 451 valley during the Pleistocene-Holocene transition exploited resources from various habitats and 26 27 452 as diverse as large grazers, such as horse (Equus lambei), mammoth (Mammuthus primigenius), 28 29 453 bison (Bison priscus) and wapiti (Cervus canadensis/elaphus); waterfowl (Anatidae); gamefowl 30 31 32 454 (Tetraoninae); and anadromous fish (Salmonidae); as well as trees and berry shrubs (Choy et al., 33 34 455 2016; Halffman et al., 2015; Lanoë and Holmes, 2016; Potter et al., 2014, 2013; Yesner, 1996). 35 36 456 In parallel, patchy distribution of resources made their location and seasonality more predictable 37 38 457 than in more homogeneous landscapes, thereby reducing the search costs associated with hunting 39 40 41 458 and gathering (Lanoë et al., 2018b). These conditions for a productive and efficient hunting and 42 43 459 gathering economy may have been one of the dominant factors in the settling of the middle 44 45 460 Tanana valley during the Terminal Pleistocene and Early Holocene, and may help explain the 46 47 48 461 unusually rich archaeological record for this period. 49 50 51 462 52 53 54 463 5. Conclusion 55 56 57 58 20 59 60 Page 22 of 42

1 2 3 464 Small mammal remains at Terminal Pleistocene and Early Holocene archaeological sites of the 4 5 6 465 middle Tanana valley provide a welcome, if uncomplete, addition to the paleontological and 7 8 466 paleoenvironmental record of that region. Small mammal remains are routinely recorded as by- 9 10 467 products of archaeological projects that span several years of research, leading over time to the 11 12 468 collection of a sizeable number of specimens. Specimen quality in archaeological contexts is 13 14 15 469 similar to that observed in targeted paleontological projects for at least a portion of the small 16 17 470 mammal clades, namely rodents. Specimens from archaeological contexts also benefit from a 18 19 471 much tighter chronological control than typically observed at paleontological localities, 20 21 22 472 providing valuable contributions to the understanding of the evolution and biogeography of these 23 24 473 species. 25 26 27 474 Small mammal remains complement existing proxies documenting environmental change over 28 29 475 time. In contrast to proxies such as pollen or large mammal remains, small mammal remains 30 31 32 476 document more closely the environments local to archaeological sites. Such information is of 33 34 477 large interest to understand subsistence and mobility behavior of past hunter-gatherers, 35 36 478 particularly so in the characteristically heterogeneous environments of Terminal Pleistocene and 37 38 479 Early Holocene eastern Beringia and central Alaska. 39 40 41 480 A better understanding of the processes leading to the accumulation of small mammal remains 42 43 44 481 will benefit our knowledge of the formation and preservation of archaeological occupations in 45 46 482 aeolian context. The tight association of small mammal remains and probable burrows with 47 48 483 anthropogenic features raises questions of how post-depositional bioturbation may affect the 49 50 51 484 spatial integrity of archaeological occupations. Future research should in parallel explore how 52 53 485 the presence of small mammal remains may be indicative of human activities and anthropogenic 54 55 486 modification of place. 56 57 58 21 59 60 Page 23 of 42

1 2 3 487 4 5 6 488 Acknowledgments 7 8 9 489 Excavations of the sites reported in this paper were funded by grants from the National Science 10 11 12 490 Foundation to Lanoë (BCS-1504654), Reuther (PLR-1107631), Holmes (PLR-0540235), and 13 14 491 Potter (PLR- 1138811 and PLR-1223119). Additional funding was provided at various stages of 15 16 492 this research by the American Philosophical Society, the School of Anthropology of the 17 18 493 University of Arizona, the Alaska State Historic Preservation Office, and the University of 19 20 21 494 Alaska Museum of the North. 22 23 24 495 Access to the Hollembaek site was provided by the Hollembaek family. Excavation of the other 25 26 496 sites was permitted by the Alaska State Historic Preservation Office. Access to comparative 27 28 497 collections at the University of Alaska Museum of the North and the Arizona State Museum was 29 30 31 498 provided by Curator Link Olson and Collection Manager Aren Gunderson, and by Acting 32 33 499 Curators Nicole Mathwich and Matthew Rowe, respectively. 34 35 36 500 Logistical support of various sorts was provided by the Hollembaek family and by Barbara 37 38 501 Crass. Excavations and analysis of the materials benefitted from previous work by Cultural 39 40 502 Resources Consultants LLC (Michael Yarborough, Sarah Meitl, Aubrey Morrison, and Jason 41 42 43 503 Rogers), Holly McKinney, and from Carol Gelvin-Reymiller’s unpublished dissertation research 44 45 504 on the Bachner and Upward Sun River sites. An earlier version of this paper benefitted from 46 47 505 comments by Link Olson, Jonathan Driver, and three anonymous reviewers. 48 49 50 506 51 52 53 507 Data Availability Statement 54 55 56 57 58 22 59 60 Page 24 of 42

1 2 3 508 The data that support the findings of this study are available from tables and figures in the paper 4 5 6 509 as well as cited references. 7 8 9 510 10 11 12 511 References 13 14 512 Anderson, P. M., & Brubaker, L. B. (1994). Vegetation History of Northcentral Alaska: A 15 16 17 513 Mapped Summary of Late-Quaternary Pollen Data. Quaternary Science Reviews, 13, 71– 18 19 514 92. 20 21 22 515 Anderson, P. M., Edwards, M. E., & Brubaker, L. B. (2004). Results and Paleoclimate 23 24 516 Implications of 35 Years of Paleoecological Research in Alaska. Developments in 25 26 27 517 Quaternary Science, 1, 427–440. 28 29 30 518 Andrews, P. (1990). Owls, Caves and Fossils. University of Chicago Press. 31 32 33 519 Bee, J. W., & Hall, E. R. (1956). Mammals of Northern Alaska. Lawrence, Kansas: The Allen 34 35 520 Press. 36 37 38 521 Bennett, J. L. (1999). Thermal alteration of buried bone. Journal of Archaeological Science, 39 40 522 26(1), 1–8. 41 42 43 523 Bigelow, N. H. (1997). Late Quaternary Vegetation and Lake Level Changes in Central Alaska. 44 45 524 PhD Dissertation, University of Alaska Fairbanks. 46 47 48 525 Bigelow, N. H., & Powers, W. R. (2001). Climate, Vegetation, and Archaeology 14,000-9,000 49 50 51 526 Cal Yr B.P. in Central Alaska. Arctic Anthropology, 38(2), 171–195. 52 53 54 527 Blinnikov, M. S., Gaglioti, B. V., Walker, D. A., Wooller, M. J., & Zazula, G. D. (2011). 55 56 528 Pleistocene graminoid-dominated ecosystems in the Arctic. Quaternary Science Reviews, 57 58 23 59 60 Page 25 of 42

1 2 3 529 30(21–22), 2906–2929. 4 5 6 530 Choy, K., Potter, B. A., McKinney, H. J., Reuther, J. D., Wang, S. W., & Wooller, M. J. (2016). 7 8 9 531 Chemical profiling of ancient hearths reveals recurrent salmon use in Ice Age Beringia. 10 11 532 Proceedings of the National Academy of Sciences, 113(35), 9757–9762. 12 13 14 533 Cole, F. R., & Wilson, D. E. (2010). Microtus miurus (Rodentia: Cricetidae). Mammalian 15 16 534 Species, 42(855), 75–89. 17 18 19 535 Conroy, C. J., & Cook, J. A. (2000). Phylogeography of a Post-Glacial Colonizer: Microtus 20 21 536 longicaudus (Rodentia: Muridae). Molecular Ecology, 9(2), 165–175. 22 23 24 537 Conroy, C. J., & Cook, J. A. (2005). Molecular Systematics of a Holarctic Rodent ( Microtus : 25 26 27 538 Muridae). Journal of Mammalogy, 81(2), 344–359. 28 29 30 539 Cook, J. A., Runck, A. M., & Conroy, C. J. (2004). Historical Biogeography at the Crossroads of 31 32 540 the Northern Continents: Molecular Phylogenetics of Red-Backed Voles (Rodentia: 33 34 541 Arvicolinae). Molecular Phylogenetics and Evolution, 30(3), 767–777. 35 36 37 542 Davis, W. H., & Kalisz, P. J. (1992). Burrow systems of the prairie vole, Microtus ochrogaster, 38 39 543 in central Kentucky. Journal of Mammology, 73(3), 582–585. 40 41 42 544 Dilley, T. E. (1998). Late Quaternary Loess Stratigraphy, Soils, and Environments of the Shaw 43 44 45 545 Creek Flats Paleoindian Sites, Tanana Valley, Alaska. PhD Dissertation, University of 46 47 546 Arizona. 48 49 50 547 Edwards, M. E., Anderson, P. M., Brubaker, P. M., Ager, T. A., Andreev, A. A., Bigelow, N. H., 51 52 548 … Yu, G. (2000). Pollen-based biomes for Beringia 18,000, 6000 and 0 14C yr BP. Journal 53 54 55 549 of Biogeography, 27(146), 521–554. 56 57 58 24 59 60 Page 26 of 42

1 2 3 550 Elias, S. A. (2001). Mutual climatic range reconstructions of seasonal temperatures based on 4 5 6 551 Late Pleistocene fossil beetle assemblages in Eastern Beringia. Quaternary Science 7 8 552 Reviews, 20(1–3), 77–91. 9 10 11 553 Endacott, N. A. (2008). The Zooarchaeology of Lime Hills Cave: Paleoecological and 12 13 554 Taphonomic Insights. PhD Dissertation, Washington State University. 14 15 16 555 Galbreath, K. R., & Cook, J. A. (2004). Genetic consequences of Pleistocene glaciations for the 17 18 556 tundra vole (Microtus oeconomus) in Beringia. Molecular Ecology, 13, 135–148. 19 20 21 557 Georgina, D. M. (2001). The Small Mammals of Lime Hills Cave 1. In S. C. Gerlach & M. S. 22 23 24 558 Murray (Eds.), People and Wildlife in Northern North America (pp. 23–31). Oxford: BAR 25 26 559 International Series 944, Archaeopress. 27 28 29 560 Goebel, T., & Buvit, I. (Eds.). (2011). From the Yenisei to the Yukon: Interpreting Lithic 30 31 561 Assemblage Variability in Late Pleistocene/Early Holocene Beringia. College Station: 32 33 34 562 Texas A&M University Press. 35 36 37 563 Goebel, T., Waters, M. R., & O’Rourke, D. H. (2008). The late Pleistocene dispersal of modern 38 39 564 humans in the Americas. Science, 319(5869), 1497–502. 40 41 42 565 Graham, R. W., Belmecheri, S., Choy, K., Culleton, B. J., Davies, L. J., Froese, D., … Wooller, 43 44 566 M. J. (2016). Timing and causes of mid-Holocene mammoth extinction on St. Paul Island, 45 46 567 Alaska. Proceedings of the National Academy of Sciences, 113(33), 201604903. 47 48 49 568 Grayson, D. K. (1981). A Critical View of the Use of Archaeological Vertebrates in 50 51 52 569 Paleoenvironmental Reconstruction. Journal of Ethnobiology, 1, 28–38. 53 54 55 570 Guthrie, R. D. (1968a). Paleoecology of a Late Pleistocene Small Mammal Community from 56 57 58 25 59 60 Page 27 of 42

1 2 3 571 Interior Alaska. Arctic, 21(4), 223–244. 4 5 6 572 Guthrie, R. D. (1968b). Paleoecology of the Large-Mammal Community in Interior Alaska 7 8 9 573 during the Late Pleistocene. American Midland Naturalist, 79(2), 346–363. 10 11 12 574 Guthrie, R. D. (1982). Mammals of the Mammoth Steppe as Paleoenvironmental Indicators. In 13 14 575 D. M. Hopkins, J. V Matthews, C. E. Schweger, & S. B. Young (Eds.), Paleoecology of 15 16 576 Beringia (pp. 307–328). New York: Academic Press. 17 18 19 577 Guthrie, R. D. (1984). Mosaics, Allelochemics, and Nutrients: An Ecological Theory of Late 20 21 578 Pleistocene Megafaunal Extinctions. In P. S. Martin & R. G. Klein (Eds.), Quaternary 22 23 24 579 Extinctions: A Prehistoric Revolution (pp. 259–298). Tucson: University of Arizona Press. 25 26 27 580 Guthrie, R. D. (1990). Frozen Fauna of the Mammoth Steppe. University of Chicago Press. 28 29 30 581 Guthrie, R. D. (2001). Origin and Causes of the Mammoth Steppe: A Story of Cloud Cover, 31 32 582 Woolly Mammal Tooth Pits, Buckles, and Inside-Out Beringia. Quaternary Science 33 34 583 Reviews, 20, 549–574. 35 36 37 584 Guthrie, R. D. (2006). New Carbon Dates Link Climatic Change with Human Colonization and 38 39 585 Pleistocene Extinctions. Nature, 441, 207–9. 40 41 42 586 Halffman, C. M., Potter, B. a., McKinney, H. J., Finney, B. P., Rodrigues, A. T., Yang, D. Y., & 43 44 45 587 Kemp, B. M. (2015). Early Human Use of Anadromous Salmon in North America at 11,500 46 47 588 y ago. Proceedings of the National Academy of Sciences, 112(40), 12344–12348. 48 49 50 589 Harington, C. R., & Cinq-Mars, J. (2008). Bluefish Caves - Fauna and Context. Beringian 51 52 590 Research Notes, 19. 53 54 55 591 Haskell, J. P., Ritchie, M. E., & Olff, H. (2002). Fractal geometry predicts varying body size 56 57 58 26 59 60 Page 28 of 42

1 2 3 592 scaling relationships for mammal and bird home ranges. Nature, 418(6897), 527–530. 4 5 6 593 Hoffecker, J. F., Elias, S. A., Rourke, D. H. O., Scott, G. R., & Bigelow, N. H. (2016). Beringia 7 8 9 594 and the Global Dispersal of Modern Humans. Evolutionary Anthropology, 25, 64–78. 10 11 12 595 Holmes, C. E., VanderHoek, R., & Dilley, T. E. (1996). Swan Point. In F. H. West (Ed.), 13 14 596 American Beginnings: The Prehistory and Palaeoecology of Beringia (pp. 319–322). 15 16 597 University of Chicago Press. 17 18 19 598 Jones, M. C., & Yu, Z. (2010). Rapid deglacial and early Holocene expansion of peatlands in 20 21 599 Alaska. Proceedings of the National Academy of Sciences, 107(16), 7347–52. 22 23 24 600 Kaufman, D. S., Axford, Y. L., Henderson, A. C. G., McKay, N. P., Oswald, W. W., Saenger, 25 26 27 601 C., … Yu, Z. (2016). Holocene Climate Changes in Eastern Beringia (NW North America) 28 29 602 - A Systematic Review of Multi-Proxy Evidence. Quaternary Science Reviews, 147, 312– 30 31 603 339. 32 33 34 604 Kokorowski, H. D., Anderson, P. M., Sletten, R. S., Lozhkin, A. V., & Brown, T. a. (2008). Late 35 36 37 605 Glacial and Early Holocene Climatic Changes Based on a Multiproxy Lacustrine Sediment 38 39 606 Record from Northeast Siberia. Arctic, Antarctic, and Alpine Research, 40(3), 497–505. 40 41 42 607 Lanoë, F. B., & Holmes, C. E. (2016). Animals as Raw Materials in Beringia: Insights from the 43 44 608 Site of Swan Point CZ4b, Alaska. American Antiquity, 81(4), 682–696. 45 46 47 609 Lanoë, F. B., Reuther, J. D., Holloway, C. R., Holmes, C. E., & Kielhoffer, J. R. (2018). The 48 49 610 Keystone Dune Site, A Bølling-Allerød Hunting Camp in Eastern Beringia. PaleoAmerica, 50 51 52 611 4, 151–161. 53 54 55 612 Lanoë, F. B., Reuther, J. D., & Holmes, C. E. (2018). Task-Specific Sites and Paleoindian 56 57 58 27 59 60 Page 29 of 42

1 2 3 613 Landscape Use in the Shaw Creek Flats, Alaska. Journal of Archaeological Method and 4 5 6 614 Theory, 25, 818–838. 7 8 9 615 Lanoë, F. B., Reuther, J. D., Holmes, C. E., & Hodgins, G. (2017). Human Paleoecological 10 11 616 Integration in Subarctic Eastern Beringia. Quaternary Science Reviews, 175, 85–96. 12 13 14 617 Llamas, B., Fehren-Schmitz, L., Valverde, G., Soubrier, J., Mallick, S., Rohland, N., … Haak, 15 16 618 W. (2016). Ancient mitochondrial DNA provides high-resolution timescale of the peopling 17 18 619 of the Americas. Science Advances, 2, e1501385. 19 20 21 620 Lyman, R. L. (2008). Quantitative Paleozoology. Cambridge: Cambridge University Press. 22 23 24 621 MacDonald, S. O., & Cook, J. A. (2001). Mammal Inventory of Alaska’s National Parks and 25 26 27 622 Preserves: Yukon-Charley Rivers National Preserve. University of Alaska Museum. 28 29 30 623 MacDonald, S. O., & Cook, J. A. (2009). Recent Mammals of Alaska. Fairbanks: University of 31 32 624 Alaska Press. 33 34 35 625 Malaney, J. L., Conroy, C. J., Moffitt, L. A., Spoonhunter, H. D., Patton, J. L., & Cook, J. A. 36 37 626 (2013). Phylogeography of the Western Jumping Mouse (Zapus princeps) Detects Deep and 38 39 627 Persistent Allopatry with Expansion. Journal of Mammalogy, 94(5), 1016–1029. 40 41 42 628 Mann, D. H., Groves, P., Gaglioti, B. V., & Shapiro, B. A. (2018). Climate-driven ecological 43 44 45 629 stability as a globally shared cause of Late Quaternary megafaunal extinctions: the Plaids 46 47 630 and Stripes Hypothesis. Biological Reviews, doi: 10.1111/brv.12456. 48 49 50 631 Mann, D. H., Groves, P., Reanier, R. E., Gaglioti, B. V, Kunz, M. L., & Shapiro, B. (2015). Life 51 52 632 and extinction of megafauna in the ice-age Arctic. Proceedings of the National Academy of 53 54 55 633 Sciences, 112, 14301–14306. 56 57 58 28 59 60 Page 30 of 42

1 2 3 634 Meltzer, D. J. (2009). First People in a New World: Colonizing Ice Age America. Berkeley/Los 4 5 6 635 Angeles: University of California Press. 7 8 9 636 Merritt, J. F. (2010). The Biology of Small Mammals. Baltimore, Maryland: John Hopkins 10 11 637 University Press. 12 13 14 638 Moreno-Mayar, J. V., Potter, B. A., Vinner, L., Steinrücken, M., Rasmussen, S., Terhorst, J., … 15 16 639 Willerslev, E. (2018). Terminal Pleistocene Alaskan genome reveals first founding 17 18 640 population of Native Americans. Nature, 553, 203–7. 19 20 21 641 Morlan, R. E. (1989). Paleoecological Implications of Late Pleistocene and Holocene Microtine 22 23 24 642 Rodents from the Bluefish Caves, Northern Yukon Territory. Canadian Journal of Earth 25 26 643 Sciences, 26, 149–156. 27 28 29 644 Morrison, P., & Galster, W. (1975). Patterns of hibernation in the arctic ground squirrel. 30 31 645 Canadian Journal of Zoology, 53(9), 1345–1355. 32 33 34 646 Nowak, R. (1999). Mammals of the World: Ground Squirrels or Susliks. Baltimore, Maryland: 35 36 37 647 John Hopkins University Press. 38 39 648 Oswald, W. W., Brubaker, L. B., Hu, F. S., & Kling, G. W. (2014). Late Quaternary 40 41 42 649 Environmental and Ecological History of the Arctic Foothills, Northern Alaska. In J. E. 43 44 650 Hobbie & G. W. Kling (Eds.), Alaska’s Changing Arctic: Ecological Consequences for 45 46 651 Tundra, Streams, and Lakes (pp. 81–89). New York: Oxford University Press. 47 48 49 652 Potter, B. A. (2011). Late Pleistocene and Early Holocene Assemblage Variability in Central 50 51 52 653 Alaska. In T. Goebel & I. Buvit (Eds.), From the Yenisei to the Yukon: Interpreting Lithic 53 54 654 Assemblage Variability in Late Pleistocene/Early Holocene Beringia (pp. 215–233). 55 56 57 58 29 59 60 Page 31 of 42

1 2 3 655 College Station, TX: Center for the Study of the First Americans. 4 5 6 656 Potter, B. A., Baichtal, J. F., Beaudoin, A. B., Fehren-Schmitz, L., Haynes, C. V., Holliday, V. 7 8 9 657 T., … Surovell, T. A. (2018). Current Evidence Allows Multiple Models for the Peopling of 10 11 658 the Americas. Science Advances, 4, eaat5473. 12 13 14 659 Potter, B. A., Holmes, C. E., & Yesner, D. R. (2013). Technology and Economy among the 15 16 660 Earliest Prehistoric Foragers in Interior Eastern Beringia. In K. E. Graf, C. V Ketron, & M. 17 18 661 R. Waters (Eds.), Paleoamerican Odyssey (pp. 81–103). College Station, TX: Center for the 19 20 21 662 Study of the First Americans. 22 23 24 663 Potter, B. A., Irish, J. D., Reuther, J. D., Gelvin-Reymiller, C., & Holliday, V. T. (2011). A 25 26 664 terminal Pleistocene child cremation and residential structure from eastern Beringia. 27 28 665 Science, 331(6020), 1058–62. 29 30 31 666 Potter, B. A., Irish, J. D., Reuther, J. D., & McKinney, H. J. (2014). New insights into Eastern 32 33 34 667 Beringian mortuary behavior: A terminal Pleistocene double infant burial at Upward Sun 35 36 668 River. Proceedings of the National Academy of Sciences, 111(48), 17060–17065. 37 38 39 669 Potter, B. A., Reuther, J. D., Bowers, P. M., & Gelvin-Reymiller, C. (2008). Little Delta Dune 40 41 670 Site : A Late-Pleistocene Multicomponent Site in Central Alaska. Current Research in the 42 43 671 Pleistocene, 25, 132–135. 44 45 46 672 Raghavan, M., Steinrucken, M., Harris, K., Schiffels, S., Rasmussen, S., DeGiorgio, M., … 47 48 49 673 Willerslev, E. (2015). Genomic evidence for the Pleistocene and recent population history 50 51 674 of Native Americans. Science, 349, aab3884. 52 53 54 675 Repenning, C. A., Hopkins, D. M., & Rubin, M. (1964). Tundra rodents in a late Pleistocene 55 56 57 58 30 59 60 Page 32 of 42

1 2 3 676 fauna from the Tofty Placer District, Central Alaska. Arctic, 17((3)), 177–197. 4 5 6 677 Reuther, J. D. (2013). Late Glacial and Early Holocene Geoarchaeology and Terrestrial 7 8 9 678 Paleoecology in the Lowlands of the Middle Tanana Valley, Subarctic Alaska. PhD 10 11 679 Dissertation, University of Arizona. 12 13 14 680 Reuther, J. D., & Lanoë, F. B. (2018). Stratigraphic Profiles of the Hollembaek Site. On File at 15 16 681 the University of Alaska Museum of the North. 17 18 19 682 Reuther, J. D., Potter, B. A., Holmes, C. E., Feathers, J. K., Lanoë, F. B., & Kielhofer, J. (2016). 20 21 683 The Rosa-Keystone Dunes Field: The Geoarchaeology and Paleoecology of a Late 22 23 24 684 Quaternary Stabilized Dune Field in Eastern Beringia. The Holocene, 26(12), 1939–1953. 25 26 27 685 Semken, H. A., & Wallace, S. C. (2002). Key to Arvicoline (“Microtine” rodents) and 28 29 686 Arvicoline-like Lower First Molars Recovered from Late Wisconsinan and Holocene 30 31 687 Archaeological and Palaeontological Sites in Eastern North America. Journal of 32 33 34 688 Archaeological Science, 29(1), 23–31. 35 36 37 689 Sikora, M., Pitulko, V. V, Sousa, V. C., Allentoft, M. E., Vinner, L., Rasmussen, S., Margaryan, 38 39 690 A., de Barros Damgaard, P., de la Fuente, C., Renaud, G., Yang, M. A., Fu, Q., Dupanloup, 40 41 691 I., Giampoudakis, K., Nogués-Bravo, D., Rahbek, C., Kroonen, G., Peyrot, M., McColl, H., 42 43 692 Vasilyev, S. V, Veselovskaya, E., Gerasimova, M., Pavlova, E. Y., Chasnyk, V. G., 44 45 46 693 Nikolskiy, P. A., Gromov, A. V, Khartanovich, V. I., Moiseyev, V., Grebenyuk, P. S., 47 48 694 Fedorchenko, A. Y., Lebedintsev, A. I., Slobodin, S. B., Malyarchuk, B. A., Martiniano, R., 49 50 695 Meldgaard, M., Arppe, L., Palo, J. U., Sundell, T., Mannermaa, K., Putkonen, M., 51 52 53 696 Alexandersen, V., Primeau, C., Baimukhanov, N., Malhi, R. S., Sjögren, K.-G., Kristiansen, 54 55 697 K., Wessman, A., Sajantila, A., Lahr, M. M., Durbin, R., Nielsen, R., Meltzer, D. J., 56 57 58 31 59 60 Page 33 of 42

1 2 3 698 Excoffier, L., & Willerslev, E. (2019). The population history of northeastern Siberia since 4 5 6 699 the Pleistocene. Nature, 570, 182–188. 7 8 9 700 Stiner, M. C., Kuhn, S. L., Weiner, S., & Bar-Yosef, O. (1995). Differential Burning, 10 11 701 Recrystallization, and Fragmentation of Archaeological Bone. Journal of Archaeological 12 13 702 Science, 223–237. 14 15 16 703 Surovell, T., Waguespack, N., & Brantingham, P. J. (2005). Global archaeological evidence for 17 18 704 proboscidean overkill. Proceedings of the National Academy of Sciences, 102(17), 6231– 19 20 21 705 6236. 22 23 24 706 Vachula, R. S., Huang, Y., Longo, W. M., Dee, S. G., Daniels, W. G., & Russell, J. M. (2019). 25 26 707 Evidence of Ice Age humans in Eastern Beringia Suggests Early Migration to North 27 28 708 America. Quaternary Science Reviews, 205, 35-44. 29 30 31 709 Whitaker, J. O. (1972). Zapus hudsonius. Mammalian Species, 11, 1–7. 32 33 34 710 Willerslev, E., Davison, J., Moora, M., Zobel, M., Coissac, E., Edwards, M. E., … Taberlet, P. 35 36 37 711 (2014). Fifty thousand years of Arctic vegetation and megafaunal diet. Nature, 506, 47–51. 38 39 712 Wolff, J. O., & Lidicker, W. Z. (1980). Population Ecology of the Taiga Vole, Microtus 40 41 42 713 xanthognathus, in Interior Alaska. Canadian Journal of Zoology, 58(10), 1800–1812. 43 44 45 714 Wooller, M. J., Saulnier-Talbot, É., Potter, B. A., Belmecheri, S., Bigelow, N. H., Choy, K., … 46 47 715 Williams, J. W. (2018). A new terrestrial palaeoenvironmental record from the Bering Land 48 49 716 Bridge and context for human dispersal. Royal Society Open Science, DOI: 50 51 52 717 10.1098/rsos.180145. 53 54 55 718 Yesner, D. R. (1996). Human Adaptation at the Pleistocene-Holocene (circa 13,000 to 8,000 BP) 56 57 58 32 59 60 Page 34 of 42

1 2 3 719 Boundary in Eastern Beringia. In L. G. Straus, B. V Eriksen, J. M. Erlandson, & D. R. 4 5 6 720 Yesner (Eds.), Humans at the End of the Ice Age (pp. 255–276). New York: Plenum Press. 7 8 9 721 Zazula, G. D., Froese, D. G., Elias, S. a., Kuzmina, S., & Mathewes, R. W. (2007). Arctic ground 10 11 722 squirrels of the mammoth-steppe: paleoecology of Late Pleistocene middens (∼24000– 12 13 723 29450 14C yr BP), Yukon Territory, Canada. Quaternary Science Reviews, 26(7–8), 979– 14 15 16 724 1003. 17 18 725 Zimov, S. A., Chuprynin, V. I., Oreshko, A. P., Chapin III, F. S., Reynolds, J. F., & Chapin, C. 19 20 21 726 (1995). Steppe-Tundra Transition : A Herbivore-Driven Biome Shift at the End of the 22 23 727 Pleistocene. The American Naturalist, 146(5), 765–794. 24 25 26 728 27 28 29 729 Figure Captions 30 31 32 730 Figure 1. Map of the Middle Tanana valley and sites mentioned in text. 33 34 35 731 Figure 2. Simplified stratigraphy and location of recovered small-mammal remains at 36 37 732 archaeological sites in the lowlands of the Middle Tanana valley. Dots indicate the depth of 38 39 40 733 recovered remains (averaged for depth ranges), with arrows indicating the maximal reported 41 42 734 burrow depth for these species. Numbers refer to the NISP for the associated species and depth. 43 44 735 All radiocarbon dates are previously reported in the literature (references in Table 1) except for 45 46 736 Hollembaek’s, which are listed in Table 3. Starred (*) dates relate to averages of 47 48 49 737 contemporaneous dates for the same occupation, while sets of dates relate to distinct occupations 50 51 738 in compressed parts of the stratigraphies. LT: brown lemming; MM: singing vole; MMP: singing 52 53 739 vole or meadow vole; MO: tundra vole; MP: meadow vole; MR: red-backed vole; MX: taiga 54 55 56 740 vole; ZH: jumping mouse. 57 58 33 59 60 Page 35 of 42

1 2 3 741 Figure 3. Rodent burrows exposed in the stratigraphy of the Swan Point site (A) and provenience 4 5 6 742 of one of the radiocarbon dates (Beta-215328) associated with a rodent burrow (B). 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 34 59 60 Page 36 of 42

1 2 3 Elevation Excavated Site Local Other Vegetation Site Sediment Site Vegetation References 4 (m asl) Area (m2) Landform Landform Present Locally 5 well-drained mixed forest of white black spruce forest; 6 Bachner bedrock spruce, aspen, poplar, and paper lakeshore (Reuther, 2013; Wooller 310 3 loess lake 7 (XBD-155) bluff birch; patches of dry grassland with vegetation with et al., 2012); this paper 8 sedges, grasses, and sagebrush willow and alder 9 well-drained mixed forest of white Cook loess (90%) bedrock (Reuther, 2013); this 10 310 10 spruce, aspen, poplar, and paper floodplain black spruce forest (XBD-72) sand (10%) bluff paper 11 birch 12 well-drained mixed forest of white 13 Hollembaek bedrock spruce, aspen, poplar, and paper 360 12 loess floodplain black spruce forest this paper 14 (XBD-376) knoll birch - patches of dry grassland 15 with sedges, grasses, and sagebrush 16 Keystone well-drained mixed forest of white (Lanoë et al., 2018b, alluvial 17 Dune 350 22 sand sand dune spruce, aspen, poplar, and paper black spruce forest 2018a; Reuther et al., terraces 18 (XBD-363) birch 2016) 19 (Dilley, 1998; Hirasawa 20 well-drained mixed forest of white muskeg with black and Holmes, 2017; Swan Point loess (80%) bedrock alluvial 21 320 86 spruce, aspen, poplar, and paper spruce, larch, and Holmes, 2014, 2011; (XBD-146) sand (20%) knoll terraces 22 birch tussocks Lanoë and Holmes, 23 2016) 24 Upward well-drained mixed forest of white dune field (Potter et al., 2014, muskeg with black 25 Sun River 260 240 loess sand dune spruce, aspen, poplar, and paper and 2011, 2008; Reuther, spruce and tussocks 26 (XBD-298) birch floodplain 2013) 27 Table 1. Sites geomorphological setting and present environment (aspen: Populus tremuloides; black spruce: Picea mariana; grass: 28 29 Poaceae; larch: Larix laricina; paper birch: Betula papyrifera; poplar: Populus balsamifera; sagebrush: Artemisia spp.; sedge: 30 Cyperaceae; white spruce: Picea glauca). 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 37 of 42

1 2 3 Excavated Density Density 4 Site NISP 2 -2) -2 -3 5 Area (m ) (NISP.m (NISP.m .yr ) 6 Bachner 6 1 6.00 0.92 7 Cook 1 10 0.10 0.02 8 9 Keystone Dune 0 22 0.00 0.00 10 Hollembaek 42 12 3.50 0.54 11 Swan Point 26 86 0.30 0.05 12 13 Upward Sun River 22 240 0.09 0.01 14 Total 97 371 0.26 0.04 15 16 Table 1. Recorded vole, lemming, and jumping mice lower first molars. 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 38 of 42

1 2 3 Table 3 : Radiocarbon dates for the Hollembaek Site. Calibration conducted with 4 5 IntCal13 and OxCal 4.3 (Bronk Ramsey, 1994; Reimer et al., 2013) 6 7 8 Laboratory 14C Date Calibrated Age B.P. δ13C 9 Material 10 Number B.P. (2σ) (‰) 11 Beta-294400 collagen, Cervus canadensis 7330 ± 40 8290-8020 -20.8 12 13 UGAMS 26195 collagen, Cervus canadensis 7020 ± 30 7940-7790 -21.3 14 UGAMS 26196 charcoal 7690 ± 30 8540-8420 -24.3 15 UGAMS 30764 charcoal 10,220 ± 25 12,090-11,810 -23.4 16 17 UGAMS 30765 charcoal 6680 ± 20 7590-7510 -23.1 18 UGAMS 30766 charcoal 7530 ± 20 8390-8340 -24.7 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 39 of 42

1 2 3 Keystone Upward Sun Bachner Cook Hollembaek Swan Point 4 Dune River 5 Brown lemming 14,000 11,000 6 7 Ground squirrel > 12,000 > 12,000 > 12,000 > 12,000 > 13,000-11,000 8 Jumping mouse < 8000 < 9,500 9 Cf. meadow vole 8000 10 8000 Red-backed vole 14,000-12,000 11,000 11 > 12,000 12 11,000 Singing vole 12,000 > 12,000 13 13,000 14 8000 15 16 Taiga vole 12,000 14,000-12,000 17 > 12,000 18 11,000 Tundra vole 12,000 > 12,000 14,000-12,000 19 11,000-13,000 20 21 22 Table 1. Estimated ages cal B.P. of for small mammal presence at archaeological sites in the 23 24 lowlands of the middle Tanana valley. The likely occurrence of meadow vole at Hollembaek is 25 discussed in section 4.3. 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 40 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Figure 1 : Map of the Middle Tanana valley and sites mentioned in text. 24 25 84x43mm (300 x 300 DPI) 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 41 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Peer Review 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 Figure 2. Simplified stratigraphy and location of recovered small-mammal remains at archaeological sites in 45 the lowlands of the Middle Tanana valley. Dots indicate the depth of recovered remains (averaged for depth 46 ranges), with arrows indicating the maximal reported burrow depth for these species. Numbers refer to the 47 NISP for the associated species and depth. All radiocarbon dates are previously reported in the literature 48 (references in Table 1) except for Hollembaek’s, which are listed in Table 3. Starred (*) dates relate to averages of contemporaneous dates for the same occupation, while sets of dates relate to distinct 49 occupations in compressed parts of the stratigraphies. LT: brown lemming; MM: singing vole; MMP: singing 50 vole or meadow vole; MO: tundra vole; MP: meadow vole; MR: red-backed vole; MX: taiga vole; ZH: 51 jumping mouse. 52 53 54 55 56 57 58 59 60 Page 42 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Peer Review 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 Figure 2 continued. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 43 of 42

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Figure 3 : Rodent burrows exposed in the stratigraphy of the Swan Point site (A) and provenience of one of 23 the radiocarbon dates (Beta-215328) associated with a rodent burrow (B). 24 25 192x90mm (300 x 300 DPI) 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60