Determining Dietary Variation and its Palaeoecological Implications of Mammuthus primigenius Using Mesowear Angle Analysis Master’s Programme in Geology and Geophysics Palaeontology and Global Change Master's Thesis Author: Regan Douglas Supervisor(s): Juha Saarinen 28.5.2024 Helsinki Faculty: Faculty of Science Degree programme: Master’s Programme in Geology and Geophysics Study track: Palaeontology and Global Change Author: Regan Douglas Title: Determining Dietary Variation and its Palaeoecological Implications of Mammuthus primigenius Using Mesowear Angle Analysis Level: Master’s Thesis Month and year: June 2024 Number of pages: 85 Keywords: Pleistocene, Proboscidea, Mammoth, Mesowear, Palaeoecology Supervisor or supervisors: Juha Saarinen Where deposited: Additional information: Abstract: Mesowear, a method that scores the wear of teeth to determine the amount of abrasive material in the diet, has long been used to understand palaeoecology though the diet of herbivores. Until recently, proboscidean teeth could not be used for these studies. The method of mesowear angle analysis introduced by Saarinen et al. in 2015 has made this possible by measuring the relative angle between the enamel ridges and dentine valleys of the lophs of proboscidean teeth to account for wear. This study compares the average mesowear angles of 428 specimens of Pleistocene Mammuthus to determine geospatial variation across the genus as well as within the species M. primigenius. These results are then corroborated with previous studies of other palaeoecological proxies to ensure they truly reflect a means to determine palaeoecology through proboscidean mesowear. Overall, this study finds that significant geospatial variation and little interspecific variation of Mammuthus proves that mammoths were highly adaptable herbivores capable of surviving in a wide array of the harshest habitats and browsing or grazing habits were not determined by species morphologies, but the environments they inhabited. Table of Contents 1 Introduction 5 2 Background 6 2.1 Taxonomy 6 2.2 Teeth 10 2.3 Diet 12 2.4 Woolly Mammoths in Their Habitats 12 2.4.1 Morphological Adaptations 12 2.4.2 Pleistocene Habitats 14 2.5 Palaeoecological Proxies 17 2.5.1 Mesowear 17 2.5.2 Stable Isotope Analysis 19 2.5.3 Pollen Analysis and Plant Fossils 20 2.5.4 Vertebrate Assemblages 21 2.5.5 Mollusca Assemblages 21 3 Materials and Method 23 3.1 Materials 23 3.2 Method 23 3.2.1 Mesowear Angle Analysis 23 3.2.2 Other Measurements 24 3.2.3 Statistical Analyses 26 4 Results 27 5 Discussion 38 5.1 United Kingdom 39 5.2 Hungary 41 5.3 Germany 42 5.4 Italy 42 5.5 Finland 43 5.6 Estonia 44 5.7 Russia 45 5.8 Alaska 47 5.9 Canada 48 5.10 Montana 49 5.11 South Dakota 50 5.12 California 50 5.13 Texas 51 5.14 Colorado 52 5.15 New Mexico 52 5.16 Nebraska 54 5.17 Kansas 55 5.18 Iowa 55 5.19 Missouri 56 5.20 Mexico 57 5.21 Illinois 57 5.22 Kentucky 57 6 Conclusion 59 7 Acknowledgements 60 References 61 Appendices 74 5 1 Introduction Despite their reputation as one of the most commonly known prehistoric animals of the Pleistocene, variations within the woolly mammoth’s (Mammuthus primigenius, Blumenbach 1799) diet remain a mystery. Some fossil finds have preserved stomach contents or faecal matter that give a broad sense of the diet. Isotope data has been used in the past to narrow down the dietary habits of Mammuthus primigenius but anomalous results have made interpretations of these studies difficult. Mesowear— the effect of types of consumed vegetation on the tooth wear pattern of herbivorous mammals—may provide new insights or expand upon these previous studies. Due to the unique shape and wear pattern of proboscidean teeth, the mesowear angle analysis of M. primigenius is achieved by measuring the angle between the dentine valley and enamel ridges of the average worn lamellae of each molar, which then serves as a proxy for average wear of the tooth (Saarinen et al. 2015). Blunt mesowear angles reflect the dominance of graze material in the diet whereas sharp mesowear angles reflect feeding on predominantly browse material. The aim of this study is to compile mesowear data from a wide sample of individuals in order to approximate diet by measuring the effect of their food on the worn occlusal shape of their molars. This dataset includes Eurasian and North American samples of M. primigenius, M. trogontherii, M. meridionalis, M. columbi, and indeterminate Mammuthus specimens. This data is then compared to palaeoecological reconstructions in order to determine if the mesowear measurements agree with other palaeoecological proxies such as pollen data, plant macrofossils, vertebrate assemblages, and invertebrate assemblages. By sampling both M. primigenius as well as other Mammuthus species, comparisons between species can be made to investigate the possibility of differing approaches to environmental variations. Covering a large geographical range allows for, in the case of this study, a detailed look at which dental traits are species driven, which are most susceptible to change within a species between geographical populations, an d which are simply changed through time and evolution through other means. 6 2 Background 2.1 Taxonomy The oldest proboscidean fossil is thought to be from approximately 55 million years ago in Morocco, belonging to Phosphatherium escuilliei (Baleka et al. 2022). The evolution of proboscideans then saw three major radiations over the next 55 million years. The first of these was the Late Paleocene-Eocene diversification of primitive proboscideans, followed by the second radiation in the Early Miocene which diversified the families Gomphotheriidae, Mammutidae, and Stegodontidae. The final major radiation occurred in the Late Miocene-Early Pliocene and diversified the Elephantidae family, in which Mammuthus is included (Baleka et al. 2022). The earliest true elephantid fossils are represented by genera such as Stegotetrabelodon and Primelephas from the Late Miocene of East and Central Africa. The earliest Mammuthus species, M. subplanifrons, could already be found by 6 million years ago in Kenya (Sanders 2023). Mammuthus did not reach Europe until the Late Pliocene, represented at this time only by M. rumanus. By the Early Pleistocene, M. meridionalis had spread across Europe (Lister et al. 2005). The evolution of Mammuthus primigenius is thought to have followed gradual changes along the lineage of three Eurasian Pleistocene chronospecies: M. meridionalis—M. trogontherii—M. primigenius (Lister & Sher 2001). In Eurasia, M. primigenius and M. trogontherii may have coexisted at certain points of the Pleistocene (Maschenko 2002). The total number of lophs (lamellae) as well as the number of lamellae over a 10 cm length of the crown of a molar, known as lamellar frequency, is often used to distinguish between members of Elephantidae. The woolly mammoth, the most derived form in this lineage, had the highest lamellar frequency. This species lived through the coldest and most arid climate of these three Mammuthus species and thus required teeth that were capable of processing the dry and coarse vegetation available during Late Pleistocene (Boeskorov et al. 2016). Through time, the teeth in this lineage evolved towards increasing hypsodonty (tooth crown height), thinner enamel, an increase in the number of lamellae and lamellar frequency, and changes in the shape of the cranium and mandible (Lister & Sher 2001). A dramatic shift towards hypsodonty is seen globally across most herbivore groups during the Miocene, largely attributed to the spread of grasses and open, dry 7 environments. This is not only in response to vegetation but also partially attributed to aeolian mineral dust from these relatively drier environments increasing wear on the teeth (Damuth & Janis, 2011; Fortelius et al. 2002). Early elephant relatives were no exception—their crown height saw a threefold increase during the Late Miocene of Africa (Lister 2013). At the same time, these elephants greatly increased the number of lophs present in their teeth, which is another adaptation to prevent excessive dental wear due to the incorporation of more abrasive material in the diet (e.g. Saarinen and Lister 2023). In Africa, this morphology is widely seen by approximately 5 million years ago, just 3 million years after the spread of C4 grasses began (Lister 2013). The more derived morphology of woolly mammoth teeth is believed to be the most extreme form of these adaptations for grazing in the steppe environments of the Late Pleistocene. This type of habitat, particularly in the far North, would have been severe, cold, and arid, providing many challenges for large herbivores to find the necessary resources to survive (Boeskorov et al. 2016). From the earliest forms of mammoth to the latest forms, M3 plates counts increased from 8 to 27+, lamellar frequency increased from 3.0 to 11.0, and enamel thickness decreased from 5.5 to 1.0 or less (Maglio 1973). Between 1.2-0.8 million years ago, mammoths in Siberia had shifted to an herb- or grass-dominated diet due to climatic and environmental conditions (Lister & Sher 2001). Subsequently, the woolly mammoth began to diverge from the steppe mammoth by approximately 800,000 years ago. These mammoths then began to radiate through these difficult habitats to reach other areas, such as Europe and North America, in a series of immigration events. Lister and Sher (2001) suggest that Siberia served as a factory-style source for these immigrants multiple times during the evolution of the woolly mammoth, as the severe conditions there forced grazing adaptations that were then able to persevere in other regions of the North Hemisphere. Similarly, the development of the M. columbi lineage occurred through two separate events according to mitogenome analysis. These events, where an M. primigenius lineage and an ancestral M. trogontherii lineage of eastern Eurasia intermixed, created a hybrid speciation event that gave rise to North American M. columbi (van der Valk et al. 2021). Phylogenetic relationships between North American mammoths have a long and contested history which is still unresolved as of today. M. meridionalis, M. columbi, 8 M. imperator, M. jeffersonii, M. primigenius, and M. exilis are the most commonly named species, though others such as M. hayi or M. haroldcooki are purported to exist (Lister 2017). In 1973, Maglio considered seven global species of Mammuthus and stated that a larger number of species and subspecies within the genus was not justified with the level of diversity seen in the species and that many of these forms were simply successions along a lineage. Throughout the 1900s, multiple characters such as Henry Fairfield Osborn and Oliver Perry Hay erected conflicting species, referred species to various genera, and argued over the priorities of species names (Lister 2017). Remnants of this debate are still ongoing today and the challenging overlapping diagnostic characters between molars, one of the hardiest and most common fossils of mammoths, only add to the confusion. The most definitive and comprehensive guide outlining diagnostic measurements for Elephantidae is Maglio’s 1973 study on evolutionary relationships within Elephantidae. However, even here North American mammoths go largely unstudied—measurements for M. primigenius are provided although only in reference to Eurasian members of the species, and M. imperator and M. columbi are both only briefly discussed with no listed measurements. M. meridionalis is also considered as a primitive North American mammoth in this discussion, though the designation of the earliest mammoths in North America is still unclear and it is unlikely that M. meridionalis ever migrated to the New World (Lister 2017). M. imperator is described by Maglio as being a robust form that is morphological comparable to M. armeniacus (syn. M. trogontherii) of Europe. It was a contested species from the start, even by the erector of the species who later noted that it was similar to other specimens he had seen before (Leidy 1869; Falconer 1863). Maglio (1973) hints at the inadequacy of the holotype of M. imperator for establishing a new species but Lister (2017) goes into far more detail, outlining the incomplete nature of the holotype and the overlap of M. columbi and M. imperator characteristics in both the holotype and the neotype. Even so, some researchers have kept the species, such as Aguirre (1969), Graham (1986), Madden (1981), while Maglio (1973) and others (e.g., Kurtén & Anderson 1980; Agenbroad 2003, 2005; Lister 20017) have considered it a synonym of M. columbi. 9 Even between more widely accepted species, relationships remain enigmatic. Mitogenomic sequencing of M. columbi showed that not only was this particular specimen closely related to M. primigenius, but the mitogenome actually belonged within the M. primigenius mitogenome (Enk et al. 2011). This has many speculative implications for the evolution of these two North American species, but more importantly it has implications of potential interbreeding between the populations throughout their history in North America. Should the more primitive M. columbi have interbred with the more advanced M. primigenius, the morphological characters of their offspring may have displayed transitional properties that make species designation difficult, leading to the creation of several species and subspecies designations that have little solid evidence or justification. While specimens referred to M. imperator very well may represent sexual or ontogenetic variations within M. columbi, they are often extreme cases that would be considered outliers for the diagnostic characteristics of the species. For the purposes of this study, specimens historically referred to M. imperator have been kept separate from M. columbi due to their distinct morphology of thick enamel, large width, and excessive cover of cementum and will be referred to as M. ?imperator here onwards. Similarly, the DNA analysis that revealed interbreeding between M. columbi and M. primigenius showed that many of those intermediate forms had been designated M. jeffersonii. (Enk et al. 2011). The M. jeffersonii designation is problematic for many reasons and most of these specimens most likely represent morphological variations between geographically distinct populations of M. columbi, M. primigenius, or both. Maglio (1973) wrote that advanced M. jeffersonii could not be separated from M. primigenius based on dental characters alone, and indeed all measurements he describes are identical to woolly mammoths. He further suggests that these advanced forms were simply evolutionary responses to the tundra and boreal forests of the north, and may be a polytypic species requiring deeper study. Like M. imperator, M. jeffersonii has its supporters (Kurtén & Anderson 1980; Madden 1981; Saunders et al. 2010) and its critics (Maglio 1973; Lister 2017). Because the less-advanced forms overlap with M. columbi and the more-advanced forms overlap with M. primigenius, specimens in this study that had been historically referred to M. jeffersonii and could not be reidentified to either M. columbi or M. primigenius almost certainly belong to 10 one of these two species but should be considered only at the genus level and will be referred to as M. ?jeffersonii here onwards. 2.2 Teeth Mammoth teeth, particularly the large and robust third molars, are the most plentiful of woolly mammoth remains. Mammoth dentition is significantly reduced. Mammuthus does retain six molariform teeth—three deciduous premolars and three permanent molars—in its dentition (Maschenko 2002). Fourth premolars have been retained in a small number of fossil mammoths, though it is a mutation likely related to ancestral traits due to its relatively rare occurrence (Sanders 2018). The formula is as follows for Mammuthus deciduous dentition (left) and permanent dentition (right): 1.0.3.0 0.0.3.0 1.0.0.3 0.0.0.3 (Maschenko & Kalmykov 2009) Each molar has three main roots, with deciduous premolars ranging from one to three roots. These teeth consist of a series of plates formed by dentine covered with a layer of enamel. The gaps between the plates are filled by cementum, which is also present around the outside of the crown. As the teeth wear on the occlusal surface, the plates are worn to expose the dentine valley in the center of each plate. Left and right cheek teeth can be distinguished by curvature, as the crowns of upper teeth are concave lingually and the crowns of lower teeth are concave buccally (Maschenko 2002). Distinguishing between upper and lower teeth can be more difficult in late stages of wear, however upper molars tend to have shorter roots, higher crowns, and a convex profile on the worn occlusal surface, whereas lower molars have the corresponding opposites to these traits. The evolution of hypsodont multi-ridged teeth was primarily in response to the expansion of open, dry environments throughout the Cenozoic (Saarinen & Lister 2023). An increase in lamellae provides more enamel ridges per tooth and therefore more shearing surfaces that can be used to process tough and abrasive materials (Saarinen & Lister 2023). The plate formula of the woolly mammoth is the most advanced in terms of plate count and is as follows: 11 𝑀3 20 − 27 20 − 27′ 𝑀2 15 − 17 15 − 16′ 𝑀1 12 − 14 11 − 15′ 𝑑𝑝4 9 − 13 10 − 11′ 𝑑𝑝3 8 8′ (Maglio 1973) The first evidence of the typical ‘conveyor-belt’ form of tooth replacement in proboscideans dates back to approximately 26 million years ago (Sanders 2018). It is achieved through the anterior movement of teeth during use, where only one or two cheek teeth are in occlusion at any given time, until their subsequent release from the jaw and loss (Sanders 2018). The second deciduous premolar differs from the other 5 molariform teeth in both formation and replacement, as it replaces vertically unlike the typical horizontal replacement of proboscidean teeth (Maschenko & Kalmykov 2009). This method of replacement improves the longevity of proboscidean teeth as a whole by limiting the number of teeth in use and allowing each tooth to be used to its full potential before replacement (Sanders 2018). This was achieved through earlier eruption times—especially of the final three molars—as well as longer formation times for the crown (Maschenko 2002). This formation period was so prolonged that the anterior end of the tooth entered wear before the posterior plates had finished forming fully. This, in turn, lengthens the lifetime of proboscideans. Prolonging the use of permanent teeth and thereby prolonging life of the individual not only allows for the use of a wider range of food but also allows for more breeding cycles, allowing the species to spread geographically and produce more offspring. The incorporation of a wider range of food includes those that are gritty and abrasive by nature, which are tougher on the teeth and wear them more rapidly, as well as those that have become drier or coarser in response to climatic conditions and may increase wear on a smaller scale. These adaptations relate to the dry, coarse vegetation of the Late Pleistocene environments that this species more often inhabited compared to ancestral species. Third molars are the best diagnostic for distinguishing between late species of Mammuthus. For the M. meridionalis—M. trogontherii—M. primigenius lineage of Eurasia, lamellae counts are a reliable way to determine species. The average number 12 of lamellae begins with 13 in M. meridionalis and increases to 19 for M. trogontherii and again to 24 in M. primigenius (Lister & Sher 2015). 2.3 Diet Preserved carcasses of the woolly mammoth in places such as Siberia and Canada have given direct evidence of their diets. Stomach contents have shown diets including leaves and branches of shrubs, grasses, mosses, and branches and bark of trees (Boeskorov et al. 2016; Maschenko 2002, Shpanksy 2021). Much of their diet was low in nutritional quality, necessitating larger quantities of food overall. One such well-known carcass, the mammoth calf Lyuba from the Yamal Peninsula of Russia, also had the faeces of adult mammoths in its stomach (Boeskorov et al. 2016). This is a known practice for the young of some species to establish gut microbiomes and, in the case of roughage, utilizes resources to their highest ability by extracting nutrients that had not been extracted the first time through the digestive system (van Geel et al. 2011). It may also represent the scarcity of food in general. The daily mass of vegetation for an adult mammoth weighing between 2.5-3.5 tons would have ranged between 150-200 kg (Maschenko 2002). The seasonal variability in the diet has also been observed from preserved stomach carcasses. Summer diets predominantly comprised of sedges, grasses, and mosses with smaller proportions of bushes and trees compared to winter counterparts (Maschenko 2002; Shpansky 2021). In Siberia, these often represented a lack of arid plants, reflecting the general moisture of the summer season through the Pleistocene in the region (Shpansky 2021). There is a lack of direct data on winter diet of mammoths, though it can be inferred based on knowledge of winter vegetation in severe northern Pleistocene environments that they would have been forced to eat dry grass and woody parts of bushes and trees (Maschenko 2002). 2.4 Woolly Mammoths in Their Habitats 2.4.1 Morphological Adaptations Since approximately 2.6 million years ago, mammoths have been continuously present in Eurasia (Lister & Sher 2001). The woolly mammoth had a geographical range spanning Europe, Asia, and North America in the Pleistocene. They could be found across Eurasia except for in mountainous areas (Ukkonen et al. 2011). 13 Fennoscandian finds even suggest that woolly mammoths were not only present in interglacial periods, but also during the last glacial in times when a thick continental ice sheet was thought to be covering the landscape, except for the LGM when Fennoscandia was fully glaciated (Ukkonen et al. 2007). Many of these environments were severe and difficult to survive in, especially for large herbivores that demand huge intakes of vegetation. The woolly mammoth had evolved several traits to help it thrive in such unforgiving habitats. The woolly fur that gives this species its namesake was an adaptation to ward off the cold. This wool formed a thick undercoat, primarily for warmth, and a sparse overcoat of guard hairs to keep them dry. The reduced ear size and tail length diminished the surface area of vulnerable extremities and helped retain heat (Boeskorov et al. 2016). Along with the fur, perhaps the most important adaptation of this species to cold environments was its thick skin and layers of fat (Boeskorov et al. 2016). Cave art shows woolly mammoths consistently depicted with humped backs, likely a form of fat storage for the animal. This fat storage would have not only been a source of warmth for the animal but also would have helped it survive long, harsh winters with little high-quality vegetation until the summer growing season (Boeskorov et al. 2016). Despite this, early analyses of fat deposits on mammoth carcasses showed no particular concentration of fat and instead showed a uniform layer around 8-9 cm thick (Boeskorov et al. 2016; Kubiak 1982). Later studies, however, found three specimens with concentrations of fat deposits—all three concentrated between the base of the skull and the withers, thus supporting their appearance in cave art (Boeskorov et al. 2016). Woolly mammoth teeth are often regarded as the most ‘advanced’ form when it comes to plate count and enamel thinness, which are both responses to the dry, tough vegetation of its environments. The early eruption and subsequent heavy wear of the deciduous third premolar in the first year of woolly mammoth calves’ lives suggests that they were already reliant on a source of combined food by six or seven months (Maschenko 2002). The fast growth of mammoth calves is likely an adaptation to short summers, ensuring that calves were prepared and able to survive long, brutal winters and difficult migrations (Boeskorov et al. 2016; Maschenko 2002). 14 These adaptations led to woolly mammoths being the champions of the mammoth steppe. It has been suggested that the drier environments of western Beringia after the Last Glacial Maximum acted as a barrier that prohibited faunal exchange, except in species such as the woolly mammoth that were already adapted to these challenging environments (Szpak et al. 2010) or in the case of more mesic conditions, that the environment would have filtered out grazing species such as the woolly rhino Coelodonta antiquitatis (Guthrie 2001). 2.4.2 Pleistocene Habitats By the Last Glacial Maximum, the so-called ‘mammoth steppe’ had become the most extensive ecosystem, stretching across nearly all of the North Hemisphere (Zimov et al. 2012). As the name suggests, this ecosystem was strongly related to M. primigenius and the related taxa that made up the associated mammoth steppe fauna. Unfortunately, there is no modern analogue for this ecosystem, as it disappeared alongside its namesake by the Holocene. However, analyses of pollen and plant fossils, vertebrate fossils, and frozen soils have revealed much about the characteristics of the mammoth steppe. It was primarily open grassland, though it also contained herbs such as Artemisia, Elymus, and Silene, and shrubs such as Salix (Zimov et al. 2012). Many other Pleistocene ecosystems that the woolly mammoth lived in or near still persist today. Forests, tundras, wetlands, and grasslands are often found in or near the same regions they were in during the Late Pleistocene. The International Union for Conservation of Nature (IUCN) has a Habitat Classification Scheme that defines modern ecosystems and organizes them into a standard to simplify the modelling and tracking of ecosystems globally (IUCN 2012). Other studies, such as Tarasov et al. (2000), have utilized a similar approach of biomization to reconstruct the dominant plant types of Pleistocene habitats—in the case of their 2000 study, in Eurasia. The boreal forest is defined by the IUCN as a region between 50-60°N with a continuous stand of trees, which can range from a closed forest to a more open woodland. Forests north of 60°N are referred to as subarctic forests in this classification scheme. In the modern definition, this includes taiga, deciduous, and coniferous forests (IUCN 2012). Conifers dominate boreal forests due to their resistance to cold, dry habitats compared to deciduous trees (Cole 1986). Tarasov et 15 al. (2000) concluded that the dominant plant types of the taiga were arctic and boreal dwarf shrubs, boreal evergreen conifers, boreal and boreal-temperate summergreen shrubs, eurythermic conifers, and heath. Other boreal forests, either deciduous, coniferous, or mixed, contained many of the same elements to different degrees. The IUCN (2012) defines the temperate forest similarly to the boreal forest, only restricted to temperate zones. It is noted that temperate and boreal forests exist on a north to south gradient, where the lower latitude temperate forest gradually transitions to the higher latitude boreal forest. Temperate forests are heavily influenced by the seasonality of the temperate regions and typically experience warm, wet summers and mild winters (Cole 1986). Tarasov et al. (2000) again defines this ecosystem type as containing several types of summergreen shrubs, conifers, and heath as well as deciduous trees and warmer varieties of shrubs and coniferous trees. Typical trees of the temperate forests globally in the Pleistocene were spruce, pine, larch, and hardwood trees (Cole 1986). Grasslands are generally defined by the IUCN as occurring in semi-arid areas with warm growing seasons. They are dominated by grasses and herbaceous plants with very little trees and woody plants (IUCN 2012). In North America, extensive tall- grass, short-grass, and mixed prairies made up the majority of grassland environments in the Pleistocene with steppe-grasslands in the far north. Modern Eurasia is home to a diverse range of steppes that were still present in the Pleistocene. These include forest-steppes, mountain-steppes, sagebrush-grass steppes, and semi-desert steppes, to name a few (Cole 1986). Steppes are usually found in sunny, dry regions with an extremely continental climate (Tarasov et al. 2000). Tundras are included in the grassland habitats of the IUCN. Here, a tundra is defined as a grassland that has developed over permafrost with cold desert-like conditions. However, grasses are not typically the most abundant species of a tundra environment. The more specific plant types identified by Tarasov et al. (2000) in the Pleistocene tundra include arctic, alpine, and boreal dwarf shrubs, grasses, heath, and sedges. Due to the severe conditions of the far north, tundra vegetation are highly specialized plants that have adapted to survive in brief growing periods and long, bitter winters. As such, mosses and lichens replace much of the typical vegetation. 16 Tundras are low-diversity environments in general and diversity further decreases with latitude (Cole 1986). Pleistocene savannas were not included in the 2000 biomization study, likely due to the study region of Eurasia. However, modern savannas are defined by the IUCN as a transitional habitat between grasslands and forests. In general, savannas are open environments dominated by grasses but including trees and shrubs (IUCN 2012). However, due to their transitional nature and the global distribution of savannas, several regional names are given to this ecosystem depending on the types of vegetation present, the aridity, and the openness of the canopy (Archibold 1995). Cole (1986) defined five main types of savannas—savanna woodland, savanna parkland, savanna grassland, low tree and shrub savanna, and thicket and scrub savanna. The parkland, woodland, and grassland savannas are mainly differentiated based on the amount and density of trees forming the canopy cover. The latter two categories of savanna are, as the names suggest, based on the type of vegetation present —in this case, low trees and shrubs or scrubs and thickets. Wetlands include bogs, fens, marshes, peatlands, and swamps, as well as other environments with saturated soils (Cole 1986; IUCN 2012). The habitat itself is not currently defined by the IUCN, but the systems are. Modern bogs are defined as low- nutrient peat systems fed only by rainwater. Fens have more decomposed peat than bogs and include groundwater as a source. Swamps are higher nutrient systems with a higher water table that typically lies above the substrate throughout the year . The water is sourced from surface runoff and groundwater (IUCN 2012). They include a number of aquatic species of plants (Cole 1986). Marshes are ecosystems with large fluctuations of water level and typically consist of reeds, sedges, rushes, and grassy meadows (Cole 1986, IUCN 2012). Few, if any, woolly mammoths would have made it far enough south to live in arid desert environments despite their southern ancestry. Some of the contemporaneous Pleistocene Mammuthus species, especially M. columbi and M. ?imperator, are known to have inhabited these environments as far south as Mexico in North America. Open shrublands and scrublands with tough vegetation such as sagebrush, thorny scrubs, and juniper or desert willow trees could be found in the southwestern United States and Mexico already by the Pleistocene (Cole 1986). 17 Unfortunately, it may have partly been the mammoth’s specializations that led to its extinction. As in modern elephants, mammoths were capable of shaping their environment to fit their needs as large herbivores covering large territories. By the end of the Pleistocene, these mammoth-steppes and other regions with high concentrations of mammoth populations changed drastically with the climate— perhaps too quickly for the mammoths to keep up (Maschenko 2002). This, paired with possible unfavourable interactions with early humans, may have been a leading cause for their extinction (Maschenko 2002, Agenbroad 2005). 2.5 Palaeoecological Proxies There are many methods of estimating and reconstructing palaeoecology or palaeoenvironments, each with their own benefits and drawbacks. Multi-proxy approaches are the best solution to reduce the uncertainties of any given method and strengthen interpretations of the data across multiple sources. This study relies on previous palaeoecological investigations from literature in order to establish whether the mesowear results reported here correspond with the vegetation of their palaeoenvironments. 2.5.1 Mesowear Mesowear analysis can be used among herbivores to distinguish between browsing and grazing diets, and a continuum of mixed diets in between these dietary extremes. The original mesowear method of Fortelius and Solounias (2000) was intended to be used on fossil ungulates and was based on selenodont or trilophodont tooth morphology. In this original method, teeth were scored based on occlusal relief (low or high) and cusp shape (sharp, rounded, or blunt). The combinations of these characteristics imply different dietary niches based on the abrasiveness of food and the resulting wear of the teeth. Lower occlusal relief and blunter cusp shapes are associated with grazing behaviours due to the more abrasive foods in the diet. In contrast, the softer materials in browsing diets lead to attrition (tooth-to-tooth) dominated wear, which produces higher occlusal relief and sharper cusp shapes (Fortelius & Solounias 2000). This method is enticing to many researchers because it is a quick-and-easy approach that requires no equipment and can be done with the naked eye or under low 18 magnification—this also means that it causes no damage to the sample, unlike isotope analyses. Additionally, mesowear scores tend to be consistent across different observers (Green & Croft 2018). The only notable drawback to mesowear analysis is the lack of detail; it describes the abrasiveness of the food but provides no specifics on the makeup of the diet (Green & Croft 2018). Cross-referencing mesowear observations and other palaeoecological proxies is therefore an important way to reduce speculation. Due to the unique lamellar morphology and wear patterns of elephant teeth, they were long considered unsuitable candidates for mesowear analysis. Their occlusal surface is more or less flat and provides little perspective on relief and the lack of proper cusps makes it impossible to score cusp shape. Despite the obstacles these adaptations provide for mesowear analysis, a new method was developed and tested using measurements of the angles between dentine valleys and enamel ridges to quantify the wear patterns in a manner applicable to the specialized morphology of proboscidean teeth (Saarinen et al. 2015). This method is based measuring the relief of enamel ridges (rather than cusps) on worn proboscidean molars as angles measured from the bottom of dentine valleys between the enamel ridges. The functional principle behind this is that increasingly abrasive diets will abrade the top of the enamel ridges making their relief progressively lower, which is reflected in progressively blunter mesowear angles (Saarinen et al. 2015). Thus, the mesowear angle analysis provides a more quantitative dietary signal than the traditional mesowear because it gives a continuous numeric value (mean mesowear angle) that is correlated with proportion of grass in diet (Saarinen et al. 2015). Prior to this development, the macroscopic wear patterns of elephant teeth were considered too challenging to analyse because of their lack of distinct wear facet development (attrition). This initial study compared fossil proboscideans Deinotherium bozasi, Elephas ekorensis, Elephas recki, Loxodonta adaurora, Loxodonta sp., Stegotetrabelodon orbus and Mammuthus columbi to two extant elephant species, Elephas maximus and Loxodonta africana. Stable carbon isotope data was used to correlate the mesowear measurements with dietary information for all species except Mammuthus columbi. The hypothesis for this method of measuring mesowear was that abrasive material would still cause bluntness in proboscidean teeth the same way it does for 19 other herbivores, only that the bluntness is seen in the relationship between the enamel ridges of the lophs and their dentine valleys instead of in the relief of the cusps. Proboscidean mesowear is reliable due to its complete lack of anatomical bias. Their teeth do not show the same types of lineage-based adaptation to grazing and browsing as other herbivores—proboscidean adaptations come in the form of hypsodonty and loph count, not cusp shape. Subsequent studies have reinforced this method (Saarinen & Lister 2016, 2023, Loponen 2020). In addition, elephant mesowear is shown to correlate with local vegetation structure, likely because elephants are large non-selective feeders that shift their dietary composition according to available vegetation (Cerling 1999; Saarinen & Lister 2016, 2023; Abraham et al. 2024) 2.5.2 Stable Isotope Analysis Stable isotope analysis of carbon has long been used to differentiate the types of plants eaten by herbivores and consequently, whether they tend to be browsers, grazers, or mixed feeders. Carbon-13 to carbon-12 stable isotope ratios can be used to determine whether an animal is eating primarily C3 or C4 terrestrial plants. C4 plants are lacking at high latitudes and, as such, the method is not as robust at differentiating diets between grazers and browsers in these regions. However, carbon isotope ratios at high latitudes may still show small-scale differences relating to openness of environment. The distinct photosynthetic processes and fractionation methods that differ between C3 and C4 plants result in δ13C enrichment or depletion and are direct reflections of the environmental conditions in which the plants lived (Ehleringer & Monson 1993). These δ13C values are found by analysing samples from bones or teeth and comparing the resulting values to the plants that would have contributed either enriched or depleted carbon to the diet (DeNiro & Epstein 1978). Lower temperatures are also thought to lead to more negative δ13C values due to reduced CO2 uptake by the animal (Szpak et al. 2010). Additionally, carbon uptake is different between bone collagen and fat deposits, therefore the differences in δ13C values relative to other herbivores could be reflective of different metabolic process and fat storage (Szpak et al. 2010). Stable isotope analysis of nitrogen is also used in determining dietary habits between herbivores, omnivores, and carnivores. Much like its carbon counterpart, this 20 analysis relies on the ratio between stable isotopes nitrogen -15 and nitrogen-14. These values are seen as a reflection of the basal δ 15N values of plants and the subsequent increase of δ15N values with each trophic level in the food chain. Thus, carnivores typically show higher values than herbivores, with omnivores occupying the range between them (Michener & Lajtha 2008). Woolly mammoths, however, show anomalously high levels of δ15N in their bone collagen. This places them squarely within the range of carnivores, if not higher. Several behavioural habits have been considered to produce these values, such as distinct metabolic processes, coprophagy, eating trampled or winter-killed plants, or eating arid-stressed plants (Schwartz-Narbonne et al. 2015). As of yet, there have been no decisive results that precisely determine which plants or habits produce the high δ15N values in M. primigenius. We know from preserved stomach contents that woolly mammoths, at least during certain points of their lives, eat their own faeces or the faeces of other mammoths (van Geel et al. 2008; Fisher et al. 2012). Whether or not this is substantial enough to account for such dramatic δ 15N values has yet to be proven. Species that excrete concentrated urea tend to have elevated δ15N values due to the recycling of nitrogen. Nutritional- and protein-related stresses may elevate δ15N values, and the rough and often low-nutrition diet of mammoths—at the very least, during winters—likely played a role in their anomalous δ15N results, though it is still unclear how and to what degree (Szpak et al. 2010). 2.5.3 Pollen Analysis and Plant Fossils Pollen samples are typically analysed for palaeoecological reconstructions by counting the number of pollen and spore grains for each taxonomic group in a sample under a microscope (Mander & Punyasena 2018). These counts are then used either as a representative of the relative abundance of each plant type in a sample, or as a representation of the influx of pollen in relation to sedimentation rates. Pollen data is an important starting place for understanding vegetation assemblages but it can be troublesome to compare between studies due to the variability in interpretations of the data based on the method used. At the very least, pollen assemblages represent known taxa in the study area that narrow down possible habitat types and their palaeoecological interpretations can be supported through multi-proxy approaches. 21 Plant macrofossils can be an extreme advantage when reconstructing past vegetation regimes, as they often reflect plants that are otherwise missing from the pollen record—either through poor preservation of fragile pollen or an absence/low abundance of pollen altogether (Mander & Punyasena 2018). While pollen often cannot be identified below genus or family level, macrofossils are often identifiable to the species level and thus provide more accurate data on the types of plants as well as the number of species within a genus being represented in pollen records. 2.5.4 Vertebrate Assemblages Vertebrate assemblages are used in palaeoenvironmental reconstructions by taking into account the unique adaptations of each species within a community for the use of different resources and the inhabitation of different environments. For example, reptiles can be used reliably to define temperature ranges or precipitation values due to many taxa’s intolerance of cold or dry environments, herbivores can be used to determine vegetation types based on their diets, and ungulates can be used to estimate the openness or closedness of an environment based on the morphology of their limbs (Cruz et al. 2024; Kovarovic et al. 2018, Saarinen et al. 2016, Saarinen & Lister 2016). Microfauna provide more accurate local data than megafauna due to their restricted range and higher reliance on local resources (Wells et al. 1987). However, not all assemblages are suitable for community analysis. A certain number of taxa must be preserved in order to make interpretations on palaeoenvironments, as it is the space where different taxa overlap in habitat characteristics that narrows down the interpretations. The community analysis method relies on the extrapolation of interpreted functional morphology from modern analogues to fossil taxa. This is based on an assumption of similar functional response to environmental conditions across time based on modern taxa’s responses to their environments. Furthermore, these behaviours exist on a spectrum and can be difficult to categorize in terms of palaeoecological reconstructions due to their variable nature (Kovarovic et al. 2018)— as seen in the highly adaptable nature of mammoths. 2.5.5 Mollusca Assemblages Molluscs, particularly gastropods, are a valuable resource for palaeoecological reconstructions due to their high quality of preservation and, thus, abundance in the 22 fossil record. The calcium-carbonate shells of gastropods preserve their life history through an incremental growth pattern and can be used for stable isotope analysis for palaeoclimatic data (Graumlich 2020). Assemblages of molluscs, terrestrial or aquatic, can be used to determine dominant forest type in much the same way as plants or vertebrates. The relative abundance of species adapted for forests, wetlands, grasslands, and all other biomes can be calculated and compared to reconstruct palaeoenvironments. As with vertebrate assemblages, this is typically done through analogies with the tolerance of modern species in range and habitat conditions (Graumlich 2020; Wells et al. 1987). 23 3 Materials and Methods 3.1 Materials The material for this study comprises fossil molars of species within the Mammuthus genus, including M. primigenius, M. columbi, M. trogontherii, M. meridionalis, as well as indeterminate M. sp. and the previously mentioned unresolved species of M. ?jeffersonii and M. ?imperator. Some data was provided from literature (Saarinen et al. 2015; Saarinen & Lister 2016) and some were measured for the purpose of this study. Newly measured teeth were provided by the Finnish Museum of Natural History, the Estonian Museum of Natural History, the Kierikki Stone Age Centre, the University of Nebraska State Museum, and the University of Iowa Museum of Natural History. The total number of specimens is 428, with 249 newly measured specimens and 179 previously measured specimens. Information on the origin of these specimens and whether or not they were previously published data can be found in Table 3 in the appendix. 3.2 Methods 3.2.1 Mesowear Angle Analysis Mesowear angles were measured according to the method introduced by Saarinen et al. (2015). This method of measurement entails using a digital angle meter (0.1° of precision) to measure the angle formed by the enamel ridges and dentine valleys on the occlusal surface of the tooth. This is done by placing the vertex of the angle meter in the deepest part of the dentine valley and opening the device until the arms touch the enamel ridges of the measured loph. The angle meter was modified with attached metal plates to create extendable arms that can fit comfortably within the dentine valley. For this study, all upper and lower molars were considered. Due to the unique ‘conveyor-belt’ method of tooth replacement in proboscideans, measurements were taken from the central lophs of each occlusal surface to ensure angles represented a true average of the wear on the tooth and were not skewed towards browsing or grazing signals by measuring less worn or more worn areas, respectively. Likewise, measurements were taken between the margin and the median expansion that forms the central loop of each loph to restrict the relative angle of the measurements to the most consistent width of the lophs. 24 When possible, up to five measurements were taken per specimen and averaged. Although providing a continuous quantified proxy for the proportion of grass in diet, the mean mesowear angles can also be used to classify the molars into on e of five categories based on dietary habit: 1. Pure browsers 2. Browse-dominated mixed feeders 3. True mixed feeders 4. Graze-dominated mixed feeders 5. Pure grazers This system and the corresponding angles that bracket the categories as seen in Figure 1 were introduced by Saarinen et al. in 2015 and later updated by Saarinen and Lister in 2023 to reflect corrections given by larger sampling. Here, correlation with δ13C values has proven that these angles can be used to distinguish between the feeding habits of browsing (less than 10% grass in diet), browse-dominated mixed feeding (between 10-30% grass in diet), mixed feeding (between 30-70% grass in diet), graze-dominated mixed feeding (between 70-90% grass in diet), and grazing (more than 90% grass in diet). Figure 1. Spectrum of mesowear values and their respective feeding categories. Modif ied f rom Saarinen et al. 2015 and Saarinen & Lister 2023. 3.2.2 Other Measurements An array of measurements was taken from each specimen using digital calipers (0.01° of precision) to clarify species and tooth position when possible. These included crown height (CH), maximum length (L), maximum width (W), lamellar frequency (LF), loph length (LL), lamellar distance (LD), enamel thickness (ET), and total loph 25 count (TL). Illustrations detailing the manner in which these measurements were taken are shown in Figure 2. Crown height was measured from unworn plates only; if all plates are in wear, then no crown height can be estimated. Maximum length was taken from the most protruding end of the anterior plate to the most protruding end of the posterior plate or talonid. Likewise, if the anterior plates have already been lost to wear, then no length can be estimated. Maximum width was measured across the tooth lingually- buccally at its widest point. Lamellar frequency is the number of plates that fit completely within a 10 cm length of the occlusal surface. Loph length is the width of each plate, or loph, measured from the outside of one enamel ridge to the outside of the opposite enamel ridge. Lamellar distance is the length between the centerpoints of two adjacent plates, measured by placing the calipers in the deepest part of the dentine valleys. Enamel thickness is taken from worn enamel surfaces and varies across the tooth, particularly in the most derived M. primigenius. As such, multiple measurements were taken per tooth when possible and averaged. Total loph count is simply the number of plates per tooth. If the anterior plates and root are lost to wear, no total loph count can be estimated. However, if the anterior root and marker plate are present, any plates in front of the root lost to wear can be estimated by species (Lister & Sher 2015). The marker plate is the first plate behind the unpaired root. Figure 2. Lef t: Buccal view of M. primigenius lower molar demonstrating the measurements of crown height (CH), enamel thickness (ET), and maximum length (L). Enamel and dentine depicted in light grey, cementum depicted in dark grey. Right: Occlusal view of M. primigenius lower molar demonstrating the measurements of maximum width (W), lamellar distance (LD), loph length (LL), and lamellar f requency (LF). Enamel and dentine depicted in light grey, cementum depicted in dark grey. 26 3.2.3 Statistical Analyses Each tooth was assigned a population by geographical location, split into countries or states in the case of the United States of America. The averages of all reported mean mesowear angles for M. primigenius populations as well as general Mammuthus populations were calculated for each geographical division to determine average mesowear signals geospatially and allow for comparisons between species within geographic divisions. The averages of all mesowear angles from all locations were calculated for each species to make this comparison. A nonparametric pairwise Wilcoxon test was run using SAS JMP Pro ver. 17.2 based on population (species and location), average mesowear angle, and a simplified set of eight palaeoenvironments. These palaeoenvironments were chosen based on the reconstructions provided for the study areas from literature, as seen in the discussion. 27 4 Results Comparison between the average mesowear signal of different species (M. primigenius, M. trogontherii, M. meridionalis, M. columbi, M. ?jeffersonii, M. ?imperator, and M. sp.) show no significant differences between species, seen in Figure 3. Each species covers a broad range of values from approximately 100-130° with median values all within a smaller range of 114.6-121.7°. As observed in the initial proboscidean mesowear study, no difference between upper and lower molars was observed in these measurements (Saarinen et al. 2015). Figure 4 shows the average mesowear angle of all species at a location, which shows variance geospatially. Hungary had the population with the lowest mean mesowear angle (106.4°) and Montana had the highest (134.2°). Figure 3 Box and whisker plot showing the mean and range values of average mesowear angles for each species. 28 Figure 4 Box and whisker plot showing the median and range values of average mesowear angles. These values cover all species within Mammuthus and are separated by locality, arranged f rom western Eurasia to eastern North America. The comparison between global population, separated here by country or state for the United States of America, show distinct variations geospatially. Figures 5 and 6 show the geospatial variation when considering all species within Mammuthus. Hungary is the only population to represent a pure-browse diet. Four populations represent a browse-dominated diet, six populations represent a mixed-feeding diet, seven populations represent a graze-dominated diet, and three population represent a pure- graze diet. Figures 7 and 8 shows the geospatial variation of only M. primigenius. Here, every population falls into a graze-dominated category besides the state of Iowa, which indicates a mixed-feeding signal. For both woolly mammoth populations (Fig. 5 and 6) and general Mammuthus populations (Fig. 7 and 8), the highest mean mesowear angles were seen in the northern populations of Eurasia and North America. For North America, this comparison to southern populations and their decreasing mesowear values can be made clearly amongst contemporaneous populations. The lower mean mesowear angles of Central Europe are affected by temporal variation, but show decreasing values as well (Fig. 5 and 6). 29 Figure 5. Global depiction of average mesowear results for Mammuthus. Figure 6. Graphical depiction of average mesowear results for Mammuthus. 30 Figure 7. Global depiction of average mesowear results for M. primigenius. Figure 8. Graphical depiction of average mesowear results for M. primigenius. Table 1 is a summary of the average mesowear angle measurements by species and population. Of the total samples, 63% came from North America. The central region of North America is highly represented with 48% of the total samples coming from the three states of Iowa, Missouri, and Nebraska. In Eurasia, 26% of the total samples of the study came from the United Kingdom. 123 samples were M. primigenius. Table 2 in the appendix details all information of the measurements of samples and their sources. 31 Table 1. Summary of average mesowear angles by species within populations. Number in parentheses indicates specimen counts for each average measurement. Population Number of Specimens Total Population Feeding Habit M. primigenius M. trogontherii M. meridionalis M. columbi M. ?imperator M. ?jeffersonii M. sp. United Kingdom (UK) 111 122.1° Graze- dominated 126.2° (21) 123.6° (54) 117.4° (36) Hungary (HUN) 4 104.9° Pure Browse 104.9° (4) Germany (GER) 11 111.1° Browse- dominated 111.1° (11) Italy (ITA) 7 112.0° Browse- dominated 112.0° (7) Finland (FIN) 3 119.6° Graze- dominated 119.6° (3) Estonia (EST) 4 119.5° Graze- dominated 119.5° (4) Russia (RUS) 20 119.4° Graze- dominated 121.7° (12) 116.6° (8) Alaska (AK) 11 127.0° Graze- dominated 127.0° (11) Canada (CAN) 1 132.0° Pure Graze 132.0° (1) Montana (MT) 3 134.2° Pure Graze 134.2° (3) 32 Population Number of Specimens Total Population Feeding Habit M. primigenius M. trogontherii M. meridionalis M. columbi M. ?imperator M. ?jeffersonii M. sp. South Dakota (SD) 1 124.0° Graze- dominated 124.0° (1) California (CA) 25 132.5° Pure Graze 132.5° (25) Texas (TX) 6 120.1° Graze- dominated 120.2° (6) Colorado (CO) 1 110.1° Browse- dominated 110.1° (1) New Mexico (NM) 3 110.6° Browse- dominated 97.2° (1) 117.3° (2) Nebraska (NE) 126 116.1° Mixed Feeding 120.9° (14) 116.9° (33) 115.6° (34) 115.0° (3) 114.4° (42) Kansas (KS) 5 115.7° Mixed Feeding 114.7° (2) 118.1° (1) 115.4° (2) Iowa (IA) 56 116.2° Mixed Feeding 114.1° (38) 119.6° (13) 122.8° (5) Missouri (MO) 24 116.1° Mixed Feeding 117.2° (19) 111.8° (5) Mexico (MEX) 2 115.5° Mixed Feeding 118.2° (1) 112.8° (1) Illinois (IL) 1 106.9° Browse- dominated 106.9° (1) Kentucky (KY) 3 113.3° Mixed Feeding 113.3° (3) 33 Table 2 summarizes the palaeoenvironmental reconstructions from literature and the proxies used. These reconstructions were then simplified into eight categories to run the nonparametric pairwise Wilcoxon test to determine statistically significant diffe rences between the mesowear signals associated with each palaeoenvironment. Table 2 Palaeoenvironmental reconstructions gathered f rom literature. Population Locality Palaeoenvironmental Reconstruction Simplified Palaeoenvironment Proxy Type Source United Kingdom (UK) Norwich Crag Open grassland Grassland/steppe Pollen analysis and plant fossils Richards et al. 1999 Cromer Forest Bed Predominantly closed woodlands Closed forest Pollen analysis and plant fossils Saarinen & Lister 2016 Ilford Open savanna Savanna Pollen analysis and plant fossils West et al. 1964 Kent’s Cavern Mammoth steppe Grassland/steppe Pollen analysis and plant fossils Campbell 1977 Lea Valley Park-tundra Tree/shrub tundra Pollen analysis and plant fossils Allison et al. 1952 Hungary (HUN) Carpathian Basin Closed woodlands Vertebrate assemblages Kovács et al. 2015 Southern Hungary Open woodlands and shrublands Vertebrate assemblages Botka & Mézáros 2018 Germany (GER) Mosbach Open riparian forests Open graze environment with riparian forests Vertebrate assemblages Maul 2000 Mauer Woodland-shrubland Stable isotope analysis Pushkina et al. 2014 Italy (ITA) Val d’Arno Closed forest interspersed with marshlands Mosaic Pollen analysis and plant fossils Bertini 2013 Finland (FIN) Sokli Open tundra Open tundra Pollen analysis and plant fossils Ukkonen et al. 2010 Ruunaa Open forest Open forest Pollen analysis and plant fossils Helmens et al. 2009 Laatokka Open taiga Open forest Pollen analysis and plant fossils Hubberten et al. 2004 Estonia (EST) Open shrub tundra-parkland Tree/shrub tundra Pollen analysis and plant fossils Ukkonen et al. 2011 Russia (RUS) Novaya Zemlya Open meadow-tundra Open tundra Stable isotope analysis; Pollen analysis and plant fossils Genoni et al. 1998; Serebryanny & Malyasova 1998 34 Population Locality Palaeoenvironmental Reconstruction Simplified Palaeoenvironment Proxy Type Source Kucheryayevka Open tundra or tundra- steppe Open tundra Vertebrate assemblages Ratnikov 2019 Gorelovka Open tundra or tundra- steppe Open tundra Vertebrate assemblages Shpansky 2021 Wrangel Island Sedge-dominated tundra or grassland Open tundra Pollen analysis and plant fossils Lozhkin et al. 2001, Bryson et al. 2010 Alaska (AK) Fairbanks Steppe or tundra with interglacial forests Pollen analysis and plant fossils Gaglioti et al. 2011, Matheus et al. 2003 Point Barrow Open grassland or meadow with shrubs Tree/shrub tundra Pollen analysis and plant fossils Lapointe et al. 2017 Canada (CAN) British Columbia Open grassland or steppe Grassland/steppe Pollen analysis and plant fossils Pendleton 2017 Montana (MT) Open marshland Open tundra Pollen analysis and plant fossils Hill 2001, Hill 2005 California (CA) Rancho La Brea Possible mosaic of open and closed environments Mosaic Vertebrate assemblages; Stable isotope analysis Coltrain et al. 2004; Fuller et al. 2020, Jones & Desantis 2017 Texas (TX) Brazos County Mosaic of grassland and woodland Mosaic Stable isotope analysis Nordt et al. 2015 San Felipe de Austin Mosaic of grassland and woodland Mosaic Pollen analysis and plant fossils; Stable isotope analysis Lundelius Jr. et al. 2019 New Mexico (NM) Jal Grassland/savanna with riparian forests Open graze environment with riparian forests Vertebrate assemblages; Pollen analysis and plant fossils Morgan et al. 2006; Morgan & Lucas 2005 Bernalillo County Montane coniferous forests Closed forest Pollen analysis and plant fossils Morgan & Lucas 2005 Nebraska (NE) Kearney Spruce forest interspersed with sedge wetlands Mosaic Pollen analysis and plant fossils Dillon et al. 2018 North-central Nebraska Upland open grassland with riparian forests Open graze environment with riparian forests Mollusca assemblages Frankel 1956 Southeastern Nebraska Upland open grassland with riparian forests Open graze environment with riparian forests Pollen analysis and plant fossils Baker 2009 35 Population Locality Palaeoenvironmental Reconstruction Simplified Palaeoenvironment Proxy Type Source Kansas (KS) Northeastern Kansas Coniferous forests and aspen parklands Closed forest Mollusca assemblages; Pollen analysis and plant fossils Wells et al. 1987; Baker 2009 Central Kansas Open meadows with riparian forests Open graze environment with riparian forests Vertebrate assemblages McMullen 1975 Iowa (IA) Mosaic of tundra and boreal forest Mosaic Pollen analysis and plant fossils Baker 1986; Baker 2009 Missouri (MO) Upland open grassland with riparian forests Open graze environment with riparian forests Pollen analysis and plant fossils Frankel 1956 Mexico (MEX) Tamaulipas Thorn scrub forest or scrubland Open forest Vertebrate assemblages Rzedowski & Huerta 1978, Ceballos et al. 2010 Illinois (IL) Northeastern Illinois Open woodlands Open forest Pollen analysis and plant fossils Saunders et al. 2010 Kentucky (KY) Big Bone Lick Mosaic of open and closed environments Mosaic Stable isotope analysis Tankersley et al. 2015 36 Figure 9 Average mesowear angle across all species and populations plotted by simplif ied palaeoenvironment. The simplified palaeoenvironments were compared with average mesowear angles and then analysed for statistical differences. Figure 9 is a box and whisker plot showing the means and ranges for each simplified palaeoenvironment regardless of locality and species. Figure 10 contains the results of the statistical analysis conducted on these data. The Wilcoxon test results show that there is a statistically significant difference in the mean mesowear angles of Mammuthus populations from wooded environments (open and closed forests) and those from grass-dominated open environments (savanna, grassland steppe, and tundra). However, there is no statistically significant difference in Mammuthus mesowear between the various open environment categories (savanna, grassland/steppe, and tundra). There is, however, a significant difference between “open graze environments with riparian forests” and other kinds of open environments. 37 Figure 10 Nonparametric pairwise Wilcoxon test results comparing palaeoenvironments and their associated mesowear measurements. Signif icant values are bolded and shown in red (less than 0.05) and orange (less than 0.01). 38 5 Discussion The lack of distinction between dietary habits of different species of mammoth is untroubling, as mammoths and their relatives are notoriously large animals that require vast amounts of vegetation to survive and thus are less selective about their feeding by necessity. As such, mammoths would have needed to eat any vegetation available to them and would have covered a wide range of dietary habits that depended solely on the vegetation of the surrounding environment (see e.g. Saarinen & Lister 2016). Population variations between mammoths came back as expected when compared to other palaeoecological proxies from literature. There is a general latitudinal trend in the results that corresponds with the environmental tendencies of the Pleistocene to have sparse, treeless habitats to the north that gradually incorporated more woodlands in lower latitudes (Blinnikov et al. 2011, Boeskorov et al. 2016). As such, northern populations such as Russia, Finland, Estonia, Alaska, and Canada show the expected incorporation of mostly abrasive vegetation in their diets. South of these populations, in the European countries of Germany, Italy, and Hungary as well as states below approximately 45°N, the mesowear angle begins to shift to lower values and indicate more browse being incorporated into the diets. These populations may also represent temporal variations in mammoth diets, as they are some of the earliest specimens included in this study. Exceptions to this latitudinal trend include the southern states of California and Texas. When excluding other species to analyse the variation between populations of woolly mammoths, the same results are seen as in the general Mammuthus results. This is expected, as there was very little variation between species. Whereas the general Mammuthus populations are subject to larger temporal variations that may make comparison difficult across populations, the M. primigenius populations are far more restricted in time and are stronger evidence of this latitudinal trend in dietary variation. This proves that the difference between mesowear signals in mammoths is solely dependent on their environment and the manner in which they interact with it, and any morphological differences between species has little to no factor in their feeding preferences. Guthrie (1982) described the mammoth steppe environment as a mosaic that varied locally but were broadly connected through the faunal 39 assemblages they supported. Therefore, we expect to find variation in dietary signals of mammoths because the very nature of their habitat is variability. The statistical analysis of the simplified palaeoenvironments show that nearly half of the comparisons (13 of 28) show statistically significant differences in respect to their average mesowear measurements. Of these, ten comparisons had a p-value less than 0.01. These results indicate that palaeoenvironments have an effect on the eventual mesowear of their inhabitants in species of Mammuthus. The difference between the wooded environments and open environments indicates that mammoths had less grass-dominated diets in the forest/woodland environments than in the grass- dominated open environments, as expected. The lack of variation between open environments (grassland/steppe, tundra, and savanna) suggests availability of similar, grass-rich dietary resources in those environmental categories. The difference between these open environments and the “open graze environment with riparian forests” may be due to more wooded environment or undergrowth dominated by other kinds of herbs than grass. The following subsections discuss the patterns of diet and palaeoenvironments for mammoth populations in different parts of their geographic range in detail. 5.1 United Kingdom An overlapping study by Saarinen and Lister in 2016 has already analysed the correlation between proboscidean mesowear and local pollen data. These specimens have been split into six palaeopopulations for the interpretations in this study: Norwich Crag, Forest Bed, Crayford, Ilford, Kent’s Cavern, and Lea Valley. Of these sites, only the Late Pleistocene Kent’s Cavern and Lea Valley have M. primigenius fossils. The aforementioned study included proboscideans outside the Mammuthus genus, which are not considered here. Norwich Crag is an Early Pleistocene site on the east coast of England with M. meridionalis fossils. It has a pollen assemblage dominated by grass species with a few herbaceous species. Macrofossils also indicate a dry, open grassland with only one Alnus fruit, reflecting the low pollen count of trees in general even in its most abundant species. The lack of arboreal pollen indicates that the trees present in the record are largely non-local populations that have been carried by wind (Richards et 40 al. 1999). The average mesowear signal for this site placed these seven individuals in the graze-dominated feeding category, indicative of an open environment such as this. The Forest Bed population contains specimens from the Bacton, Mundesley, Overstrand, Sidestrand, Trimingham, East Runton, and Pakefield sites. It also spans a larger range of time, from the Early to Middle Pleistocene, and introduces M. trogontherii. Altogether the Forest Bed specimens are comprised of thirty-nine M. meridionalis individuals and forty-eight M. trogontherii individuals. As the name suggests, the area has been interpreted as a generally forested woodlands environment with locally open areas of grasslands, such as those in West Runton (Saarinen & Lister 2016). As Forest Bed encompasses a larger span of time and multiple sites, some variation in the data is to be expected. For example, M. meridionalis populations had measured mesowear angles ranging from 111.6° at Sidestrand to an average of 119.4° at East Runton. Likewise, M. trogontherii populations ranged from an average of 118.3° at Bacton to an average of 129° at East Runton. Altogether, each site averaged within the mixed-feeding or graze-dominated feeding categories regardless of species. Taking into consideration the presence of the browsing proboscidean, Paleoloxodon antiquus at many of these sites, Saarinen and Lister posited that a form of niche partitioning took place that allowed Paleoloxodon and Mammuthus species to take advantage of browse and graze material, respectively (2016). Ilford, slightly south of these localities, has a pollen assemblage with a high count of grass as well as Artemisia and related herbaceous species. This open vegetation with scattered trees and shrubs indicates an open savanna-like environment (West et al. 1964). The site is Middle Pleistocene in age and no longer represents M. meridionalis, only M. trogontherii. This population of twenty-one individuals represents a graze-dominated feeding category, agreeing with the interpretation of an open environment with high amounts of grass. Kent’s Cavern is the most western United Kingdom site in this study, approximately 280 km away. The six M. primigenius individuals at Kent’s Cavern had an average measurement of 137.6° and represent a pure-graze feeding habit. Pollen analysis of this site showed typical mammoth steppe vegetation with high amounts of grasses 41 and sedges, followed by a moderate amount of shrub species and very few trees (Campbell 1977). This open landscape is one that would support such an abrasive diet that would lead to blunt mesowear signals like the ones measured here. Lea Valley is near Ilford and covers the Late Pleistocene. Unsurprisingly, the environment is still quite open at this time and is associated with a park -tundra vegetation (Allison et al. 1952). The mammoth population at this site was comprised solely of nineteen M. primigenius individuals which had an average mesowear angle of 126.1°, placing it in the graze-dominated feeding category—though this is mostly due to a wide range of measurements from 103.7° to 146.5°. The dietary signal is, on average, grass-dominated. This is similar to other mammoth steppe populations, although the high range of values may indicate a higher-than-usual dietary variation. This could be due to, for example, changes in the palaeoenvironment through time, but this is difficult to prove based on the resolution of available data. Palaeoenvironmental proxies were missing from the Middle Pleistocene Crayford site. Crayford is situated near Ilford and both sites are from the same interglacial period, thus it potentially represents a likewise open savanna-like environment. Mesowear results are remarkably similar: the average measurement of eighteen M. trogontherii individuals at Crayford is 126.6° compared to Ilford’s 128.0°, placing both in the graze-dominated feeding category. 5.2 Hungary Four M. meridionalis specimens from the Early Pleistocene of an unknown locality in Hungary were measured and found to be the only population of mammoths in this study with a pure-browse diet. It is also important to note that this population, along with the population from Italy, is represented solely by a much earlier member of the genus Mammuthus than the woolly mammoth. As such, it does not exhibit such special grazing adaptations as the woolly mammoth, for example very high molar tooth crowns and large number of lamellae (Lister et al. 2005) although it does show remarkable dietary variation that follows patterns of local vegetation (Saarinen & Lister 2016). While an exact estimation from outside proxies of the environment they inhabited is difficult to produce without location data, it is assumed that a mesowear signal such as this could only come from a closed environment of some sort. Indeed, faunal assemblages of the Early Pleistocene Carpathian Basin indicate closed 42 woodland environments with proboscidean browsers such as Anancus arvernensis and Mammut borsoni alongside other browsing species like Stephanorhinus jeanvireti and Tapirus arvernensis. The stable isotope analysis done on teeth from these species reveal that they relied heavily on C3 plants (Kovács et al. 2015). A reliance on C3 plants does not rule out grazing on C3 grasses, however. Furthermore, the small mammal faunal assemblage at an Early Pleistocene site in southern Hungary suggests a more open environment with woodlands and shrublands restricted only to surrounding water sources (Botka & Mészáros 2018). Regardless, it is obvious that these individuals must have inhabited woodlands or shrublands to some degree and utilized browse material regularly. 5.3 Germany The eleven German specimens in this study all represent a Middle Pleistocene population of M. trogontherii from Mosbach. The site of Mosbach 2 has been interpreted as an open environment covered with water, with riparian forests sparsely incorporated (Maul 2000). This interpretation comes from the deposits themselves as well as examination of the small mammal fauna of the site. A more regional palaeoenvironment can be drawn from local sites of comparable age such as Mauer—a nearby site where stable isotope analysis of large ungulate populations has revealed a woodland-shrubland environment capable of supporting both typical browsers and typical grazers (Pushkina et al. 2014). The Mosbach specimens have an average measurement of 111.1°, placing them in the browse- dominated category. The range of measurements covers pure-browsing (99.4°) to graze-dominated mixed feeding (121.8°). With woodlands providing predominantly browse and shrublands providing both browse and graze material, this range of mixed-feeders skewed slightly towards browse is in agreement with their purported habitat. 5.4 Italy The Val d’Arno Basin in Northern Italy saw extensive and fluctuating changes throughout the last 3 million years as the end Pliocene warming led in to the glacial/interglacial cycles of the Pleistocene. The relatively open, herbaceous environment of the latest Pliocene persisted until approximately 2.6 million years ago 43 when the reintroduction of forest and marsh pollen taxa replaced the environment with closed temperate forest species interspersed with local marshlands (Bertini 2013). Similar to Hungary, the seven Early Pleistocene Val d’Arno M. meridionalis specimens in this study have an average mesowear signal that correlates with a browse-dominated feeding habit. The mammoths here likely took advantage of the browse material available while still utilizing the sparse marshlands for grazing whenever possible. 5.5 Finland Of the mammoth remains in Finland, two are adult M. primigenius molars with acceptable preservation for mesowear studies—one from Iijoki in northwestern Finland and one from Nilsiä in eastern Finland. The molar found from Laatokka is in modern-day Russia but has been grouped with the Finnish specimens of this study due to its proximity and similarity in habitat reconstruction. Ukkonen et al. (2007, 2011) have studied Finnish mammoth fossils extensively and conducted studies on the palaeoenvironments they might have inhabited throughout the Pleistocene. Prior to the Last Glacial Maximum, between 50-40 thousand years ago, the environment of northern Finland much resembled the modern treeless tundra found in Lapland (Ukkonen et al. 2010). Iijoki is slightly to the south of Sokli, from which this record was taken. The molar from this site has a blunt mesowear angle, measuring at 130.8° and placing it at the lower end of the pure-graze feeding category. This would indeed be reflective of a treeless, open tundra environment but must be interpreted with caution as it is only one sample. In general, there is a lack of data in Finland towards the end of the Last Glacial Maximum. It is thought that the environment across Finland throughout much of the Late Pleistocene was relatively open with few trees (Ukkonen et al. 2007). At Ruunaa, an environmental reconstruction based on pollen data from the Middle Weichselian— 50-25 thousand years ago—suggests an open forest dominated by birch with stands of spruce (Helmens et al. 2009). Moving forward to the Last Glacial Maximum, the landscape had indeed opened up farther into a periglacial tundra environment with sparse vegetation (Ukkonen et al. 2011). Also a tentative observation from a small sample size, the Nilsiä molar shows a pure mixed feeding signal with a measurement 44 of 115.1°. Radiocarbon dating of the specimen has placed it between these records, at approximately 22 thousand years old (Ukkonen et al. 1999) It is unlikely that there was much graze material present in a periglacial tundra, let alone enough to support mammoth populations, and as such it is likely that this specimen comes from an open forest environment similar to that found at Ruunaa, perhaps already in the stages of transition towards a tundra. Lake Laatokka near Saint Petersburg is part of a huge lake system in modern times. The tooth included in this study is dated to over 45,800 years before present (Lõugas et al. 2002). During this time, macrofossil remains from the northwestern Russian Plain to the southeast of Lake Laatokka indicates a colder permafrost environment supporting taiga vegetation (Hubberten et al. 2004). Studies of analogous age from sites along the margins of the ice sheet in Northern Russia repeatedly point towards open forests in these areas despite the proximity to the ice sheet (Henriksen et al. 2008; Schirrmeister et al. 2002). These environments would have been more than sufficient to support a browsing diet. This specimen’s mesowear signal of 113.0° places it on the browse-dominated end of mixed feeding. These three specimens form a spectrum of grazing in the west at Iijoki and nearly browsing in the east at Laatokka with Nilsiä at the halfway point geographically. 5.6 Estonia The Late Pleistocene record of Estonia shows an open environment with shrub tundra vegetation and the incorporation of arboreal elements such as dwarf birch and juniper (Ukkonen et al. 2011). Arctic species such as Dryas octopelta are also noted in the pollen assemblage. The climate was cold and dry. Approaching the Holocene, this environment is replaced with more closed forests until it reaches its present -day assemblage. The four Estonian M. primigenius specimens of this study have an average mesowear measurement of 119.5°, all within a small range of the graze-dominated feeding category. Two of the specimens are from the same location in Puurmani, central Estonia, and one northern specimen comes from Ihasalu on the coast. One specimen did not have associated location data. This mesowear signal corresponds with an interpretation of a relatively open, predominantly graze-covered environment with available browse material—such as the aforementioned shrub tundra. 45 5.7 Russia Accurate location data of Russian specimens included in this study is lacking. Fifteen specimens have been given a broad location of Siberia, making it difficult to provide accurate palaeoecological proxies to corroborate with these results. Schirrmeister et al. (2002) undertook a comprehensive analysis of Ice Complexes in northern Siberia that demonstrate a wide variety of habitats across the region in the Pleistocene forming a mosaic of tundra-steppe biomes, marsh-tundra biomes, and submersed wetlands. These habitats likely had a complex relationship with advancing and retreating ice sheets that caused constant changes and fluctuations in vegetation throughout the Pleistocene. The Russian specimens lacking location data agree with this interpretation of the northern environment, though without knowing more exact locations it is impossible to use this data to support it. These specimens ranged from a dramatically pure-browse diet (92.2°) to a dramatically pure-graze diet (147°). Such a high variation can be related to variable conditions through time or within local ranges. This dietary variation is also supported by analyses of woolly mammoth faeces preserved in the permafrost in Siberia, with some specimens showing heavily grass-dominated diets and others a high proportion of shrub and forb consumption (e.g. van Geel 2008; Polling et al. 2021). However, five specimens included in this study have been given locations of Novaya Zemlya (Arkhangelsk Oblast), Kucheryayevka (Voronezh Oblast), Gorelovka (Tomsk Oblast), and Wrangel Island (Chukotka Autonomous Okrug). While only two teeth came from Kucheryayevka and one each from the other sites, these specimens cover the breadth of Russia and allow for more specific analyses to be done tentatively and on a smaller scale. A δ18O isotope analysis on mammoth and reindeer material by Genoni et al. in 1998 interpreted the climate of Novaya Zemlya and other Siberian sites to be periodically seasonal with warmer and wetter summers as well as colder winters than seen in modern times in the region. The authors observed that these results would have been necessary in the far north, where more severe climatic conditions would have made it impossible for the vegetation to thrive enough to support mammoth herds. Fossil pollen analysis determined that throughout the entirety of the Pleistocene and into the Holocene, forest-type vegetation was never present on Novaya Zemlya 46 (Serebryanny & Malyasova 1998). The specimen from Novaya Zemlya has a pure graze mesowear signal, indicating an animal that relied solely on the grasses of arctic meadows and tundras present on the island. The Quaternary record of Voronezh is a periodical cycle of warming/cooling and aridification/humidification (Ratnikov et al. 2019). Without accurate dating, it is impossible to tell which of these periods the specimen included in this study can be attributed to. Ratnikov et al. does, however, compare Voronezh to the general habitats in Western Siberia in the Pleistocene. It was described as a tundra or tundra- steppe and that the mammoths of the region would have eaten mainly herbaceous plants. The two teeth from Kucheryayevka analysed here have mesowear signals that indicated graze-dominated diets, and thus support the idea of a relatively open environment. In the south-central region of Tomsk Oblast, several mammoth finds and archaeological sites have been studied over the years—but none from Gorelovka. δ13C and δ15N analyses on other mammoths from Tomsk show the typical C3 based diet of high latitudes (Puchkovskaya & Shpansky 2023). Furthermore, the morphologies of the faunal assemblage found in Tomsk provide hints at the type of environment there. Very few forest-dwelling species compared to a wide diversity of open- landscape grazing species suggest a tundra or tundra-steppe environment. The hooves of the ungulates found in Tomsk sites are wider than those found in other regions—an adaptation for stability when walking across waterlogged soil (Shpansky 2021). We can then infer that during the late Pleistocene period when these fauna coexisted, south-central Siberia was a cool, wet, and open environment capable of sustaining a variety of large mammals. The Gorelovka specimen of this study had an average measurement of 117.7°. Interpreting this result from a single data point is made more difficult by its central position on the spectrum of dietary habits. Given the palaeoenvironmental interpretation of Gorelovka, mammoths in this area likely had a diverse range of vegetation to feed on that may have been subject to further diversification from seasonal variations in climate in this more southern locality. Wrangel Island is famous for being the last refugium of mammoths. Here, they saw a dramatic reduction in size before their extinction (Kirillova et al. 2020). While the Wrangel Island specimen included in this study is undated, the reduction in size of this right upper molar narrows down the most likely age in terms of the Pleistocene. 47 The specimen was heavily weathered, making precise measurements impossible. A lamellar frequency of 10, an enamel thickness of 1.22 mm, and a loph length of 6.72 mm were measured. The approximate width of the tooth was most likely between 70- 80 mm. Given this reduced size compared to expected woolly mammoth molar dimensions—for example, those found in Maglio (1973)—we can infer that it is quite late in the Pleistocene near Holocene boundary. Beyond 10,000 years ago, Wrangel Island was characterized by a sedge-dominated tundra with woody shrubs (Lozhkin et al. 2001). The presence of herbaceous plants that grow in moist environments suggest it was warmer and wetter than in modern times. Another study found that the early Holocene on Wrangel Island saw the formation of grasslands due to precipitation and temperature changes in the region (Bryson et al. 2010). This data agrees with the mesowear signal given by the specimen of this study, which indicates a graze-dominated diet suitable for either of these environments. 5.8 Alaska Six of the Alaskan specimens provided no locality data. Three specimens came from the Fairbanks area, one from eastern Alaska near the Canadian border, and one from Point Barrow. Much palaeoenvironmental work has been done surrounding Fairbanks. Fossil ground squirrel nests and caches can reveal palaeoecological data through plant micro- and macrofossil analysis. One such study found that the interior region of Alaska near Fairbanks consisted of dry steppe or mesic tundra vegetation between 44,008-21,299 years ago (Gaglioti et al. 2011). In the earlier parts of the Pleistocene of interior Alaska, plant macrofossils from peat beds reveal intermittent foresting during interglacial periods as far back as 2 million years ago (Matheus et al. 2003). While mammoth bones have been found at the site of this analysis, they do not extend beyond the late Pleistocene into the earlier parts of the period. It is unknown whether this is simply a lack of preservation, a sign that mammoths had avoided this part of Beringia at the time, or whether they simply preferred a different habitat cannot be stated definitively. The three specimens from Fairbanks in this study had a wide range of mesowear measurements (112.1°, 127.7°, and 132.4°) for an average of 124.1°. This range extends from mixed feeding to pure grazing signals and may be a reflection of the variability of the forest and tundra biomes in the Late Pleistocene of Alaska, or perhaps the specimens act as representatives of populations 48 that inhabited individual biomes for much of their lives. A larger sample size with more accurate dating could resolve these hypotheses. Palaeoecological studies of the Arctic Coastal Plain in Alaska are rare. One study site, on the border between the Arctic Coastal Plain and the Arctic Foothills, contains a comprehensive record of pollen data from approximately 40,000 years ago to the present day. This site is over 300 km southeast of Point Barrow and thus does not provide local data for the specimen used in this study. However, Lapointe et al. (2017) concluded that the vegetation changes in this site compared to contemporaneous changes in the Fairbanks region to the south indicated that the boreal forest never stretched northward to the Itkillik River site in the Late Pleistocene. Therefore, the even more northern site at Point Barrow should have seen no boreal vegetation as well. Their study showed a gradual change from grassland and meadow to the incorporation of shrubs towards the beginning of the Holocene. The Point Barrow specimen of this study has a graze-dominated mesowear signal that agrees with an interpretation of an open environment with some woody material, such as shrubs. 5.9 Canada Accurate location and age data were lost for the singular woolly mammoth specimen from British Columbia, Canada. Comprehensive ecological review of western Canada, namely British Columbia and western Alberta, suggest open grasslands or steppes across the region at the Pleistocene-Holocene transition (Pendleton 2017). After the onset of the Holocene, Betula palynological data and the introduction of forest species indicate the introduction of the boreal forest biome seen in much of the region today. This study reviewed the sites of Peace Lake (central British Columbia), Tse’K’wa National Historic Site (eastern British Columbia), Grande Prairie (western Alberta), and Wally’s Beach Site (southern Alberta). An earlier specimen dated to approximately 34,000 years ago from central British Columbia is paired with botanical evidence for a shrub tundra in the region (Harington et al. 1974). The extensive Wisconsin glaciation unfortunately destroyed Early Pleistocene deposits in the region, so not much is known about earlier glaciations and few fossils are found from that time. The specimen included in this study has a mesowear measurement of 132.0° and indicates a pure graze diet—entirely possible for any of these steppe, 49 tundra, or open grassland environments across the region until the onset of the Holocene. 5.10 Montana Three plaster casts of mammoth teeth (M. sp.) from west-central Montana were measured, all belonging to the same individual. Sites in western Montana are closely entwined with the glacial fluctuations of the Late Pleistocene. The floodwaters of melting glaciers created swampy or marshy environments east of the Rocky Mountains. Pleistocene deposits yielding mammoth bones and teeth in this region are largely palustrine or lacustrine and tied to these open marshy environments. Examples include the Jefferson River Drainage to the south, with palu strine deposits dating to 58,000-32,000 years ago (Hill 2005) or the Sun River palustrine deposits dating to approximately 11,500 years ago (Hill 2001). Hill, in his 2006 paper detailing mammoth remains across all of Montana, often used faunal assemblages to determine the palaeoecology of the sites and rarely mentions other proxies such as pollen or plant fossils. One such assumption made in this paper is that M. primigenius must have inhabited boreal, tundra, or steppe environments along the margins of ice sheets and M. columbi or other reported Mammuthus species must have inhabited open parklands and grasslands. This is an unreliable way to determine palaeoenvironments, as seen by the results of this study. Each species of Mammuthus is highly adaptable to a variety of habitats and cannot be used alone to discern between them. As such, accurate multi-proxy palaeoenvironmental reconstructions are lacking in the region which make corroborating the mesowear information from the individual in this study to local sites. All three molars from this individual indicate a pure graze diet. While this does agree with the o pen, marshy environments of the mammoth-bearing palustrine deposits in the region, it is still unknown whether the alluvial and fluvial mammoth-bearing deposits in the region reflect the same kind of open habitat. Furthermore, these samples are casts of the original specimen and all reflect the diet of only one individual, essentially restricting the sample size to one and making it difficult to make any observations definitively. 50 5.11 South Dakota The South Dakota specimen provides very little information to the study—it is undetermined at the species level, has no location data beyond the state of South Dakota, and no age data. It is a fragmentary posterior portion of a molar. It provides excellent mesowear measurements with an average of 124.0°, but as a standalone specimen with no information it cannot be determined whether this is valuable for palaeoecological reconstructions. The well-known site of Mammoth Hot Springs in South Dakota may provide future possibilities for accurate research in this area. 5.12 California The twenty-five Californian specimens were M. columbi individuals from the well- known site of Rancho La Brea. Fuller et al. (2020) investigated the stable isotope ratios of several taxa from Rancho La Brea in order to determine the age, palaeoecology, and dietary habits of the species there. They found that the δ13C values reflected a predominantly C3-based diet for all species. The study measured one Columbian mammoth specimen but did not include it in the discussion. This specimen had a reported δ13C value of -19.4‰ and a δ15N value of 9.4‰ (Fuller et al. 2014). This is within the bounds of other herbivores at the site, albeit on the most positive edge of the range in respects to both isotopic values. Comparing stable isotope values at Rancho La Brea to comparable modern European fauna suggests an open environment that may have had cool, wet winters and dry summers (Coltrain et al. 2004). However, dental microwear analysis of the site’s most common herbivores (Camelops hesternus, Bison antiquus, and Equus occidentalis) indicates mostly browsing and mixed-feeding across the faunal assemblage, even in typical grazers such as bison and horses (Jones & Desantis 2017). The stable isotope values of the rattlesnake (Crotalus sp.) were also interpreted by Fuller et al. (2020) to be indicative of a diet of prey from a closed or wooded environment. Previous studies had analysed the migratory patterns of the herbivores at this site and found that some had established, localized populations whereas others took advantage of a wider geographical range (Coltrain et al. 2004). The nature of microwear analysis is such that it only records the very last meals of an animal whereas mesowear is averaged over the course of the animal’s life. Jones and Desantis posit that the inclusion of C4 plants and C3 or C4 graze material 51 in the diets of Rancho La Brea herbivores reflect a complex and nuanced diet within a changing climate, or may be a side-effect of the migratory natures of some of these species that merely came to Rancho La Brea to die (2017). The mesowear results of the Rancho La Brea mammoths of this study overwhelmingly represent a pure-graze diet. One possibility for this result is that the area immediately surrounding Rancho La Brea may have been a mosaic of open and closed environments that supported a high diversity of taxa and allowed for multiple approaches to resource use and may also be representative of niche partitioning between concurrent browsing and grazing proboscidean species (Calandra et al. 2008, Rivals et al. 2015, Saarinen & Lister 2016). Several studies have shown that the American mastodon (Mammut americanum) was a browser across its range (e.g. Green et al. 2017) and it is possible that the mammoths at Rancho La Brea and similar sites where mammoths and mastodonts lived sympatrically and shared resources in the environment could have shifted to a more grass-dominated diet to avoid competition with the browsing mastodonts. Unfortunately, microwear or other mesowear studies have not been performed on Rancho La Brea proboscideans to date and several of these studies on contemporary herbivores did not incorporate mammoth or mastodon diets into their palaeoecological interpretations (Fuller et al. 2014; Fuller et al. 2020; Jones & Desantis 2017). Whether the mesowear signals presented here represent the migratory nature of this mammoth population, dietary niche-partitioning, or are an anomalous result amongst an assemblage of browsers is not possible to determine without further investigation and the incorporation of multi-proxy approaches. 5.13 Texas Six M. columbi specimens from two locations in Texas were used for this study. The lone specimen from Brazos County was found in the Brazos River, but more specific location information is unavailable. The nearby Waco Mammoth Site, approximately 130 km to the south, provides an excellent glimpse at the behavioural habits of Columbian mammoths in the southern United States during the Pleistocene (Nordt et al. 2015). Stable carbon isotope analysis shows a diet of 70-80% C4 grasses for this herd. The authors reconstructed a mosaic environment of grasslands and woodlands and a cool and moist glacial climate for this herd. 52 Approximately 225 km south of Waco, five specimens were taken from San Felipe de Austin, Texas. These specimens fall within a range of 112–123° compared to the 128° measurement of the Brazos County specimen. Only 35 km away, a palaeoecological study was done on proboscidean remains from the Deweyville Formation of the Late Pleistocene (Lundelius Jr. et al. 2019). These authors determined that the site was also a mixture of open grasslands and woodlands at the time. Stable isotope analysis showed all mammoth specimens plotting within a mixed C3/C4 diet towards the C4 end of the range, whereas other proboscideans from the site—the gomphothere Cuvieronius—plotted towards the C3 end of mixed-feeders or as C3 feeders. Notably, there is a distinct lack of overlap between the two genera, even within the mixed feeders. It is possible, then, that a form of niche partitioning occurred in this locality that pushed both animals to the extreme ends of available vegetation. There were no strict C4 feeders at the site, and the Columbian mammoths plot in complete overlap with Equus. 5.14 Colorado The only Colorado specimen included in the study is an indeterminate species with no location or age data beyond the state of Colorado. It is a fragment of the middle portion of a tooth, making any species identification difficult. As such, not much can be said in the way of palaeoecological reconstruction. The Rocky Mountains and its foothills provide vastly different environments that would have responded to environmental changes very differently in glacial and interglacial stages. Pleistocene sites with Mammuthus remains in Colorado can be from alpine environments in the Rocky Mountains (Agenbroad & Mead 1989; Mandryk 1998, Meyer & Weber 1995, Miller et al. 2014; Saunders 1999) or open environments in their foothills (Graham 1981; Hager 1974). It is entirely possible that the specimen in this study comes from one of these alpine sites, as its mesowear signal indicates a browse-dominated diet, but the possibility of it coming from the foothills cannot be ruled out and no conclusions can be drawn with any degree of certainty. 5.15 New Mexico This study includes two M. sp. specimens from the well-studied site of Jal, New Mexico and one M. columbi specimen from Bernalillo County. The faunal assemblage at Jal is predominantly large ungulates that are typically grazers (Mammuthus, 53 Equus, Bison) with only a few species of browsers (Camelops, Mammut) (Morgan & Lucas 2005). The authors suggest that the addition of Mammut americanum is an indication of nearby forested areas. While this alone may not be a reliable indicator of palaeoenvironment, as the browsing mastodon was certainly capable of adapting to a variety of habitats, the presence of montane vegetation in similar sites in southeastern New Mexico as well as the presence of fauna that live in modern alpine environments supports this palaeoenvironmental hypothesis (Morgan et al. 2006). Furthermore, the wetter conditions that are posited to have allowed montane vegetation to spread to lower altitudes in southeastern New Mexico at this time are also said to have supported riparian forests surrounding the streams and rivers of low-lying areas in the region (Morgan & Lucas 2005). The authors’ conclusion is that southeastern New Mexico in the Pleistocene was dominated by grasslands or savannas with interspersed forested areas alongside bodies of water. The more northern and central site outside of Albuquerque corresponds with other sites from Bernalillo County that were part of this palaeoecological study. Here, the authors found that Bernalillo County and northeastern New Mexico was made up of montane coniferous forests (Morgan & Lucas 2005). The specimens of this study agree nicely with these interpretations—particularly the one from Bernalillo County. The two Jal specimens are both categorized as mixed feeders, something that is supported by an open environment with pockets of forested areas. While some analyses of the region based on faunal assemblage point to open grasslands due to the number of not only grazing species, but high numbers of individuals within those species in the assemblage. The results from these specimens suggest that the browsers found at Jal were not adapting to an open environment despite their browsing preferences, but likely taking advantage of the scattered parklands or savannas dispersed alongside water sources. The Bernalillo specimen is well within the pure browse category with a measurement of 97.2°, supporting the interpretation of a closed environment providing a diet of almost exclusively woody plants. 54 5.16 Nebraska Modern Nebraska is a semiarid, open prairie with little to no trees across the landscape. In the past, however, wetter conditions have supported arboreal species. Pollen and macrofossils from the Kearney area of central Nebraska show that the Late Pleistocene environment in the area was that of a spruce forest interspersed with sedge wetlands (Dillon et al. 2018). While pollen analysis has rarely been done in the region, partially due to poor preservation, mollusc analyses have been performed repeatedly across the Great Plains region. One such study found that four sites in central Nebraska and one in the north showed two distinct habitats—herbs, shrubs, and grasses growing on open plains and closed forests growing in valleys (Frankel 1956). A later study across the entire Great Plains region supported this interpretation, stating that prairie gastropod taxa indicated a widespread prairie vegetation similar to the modern environment, but with scattered areas of trees and shrubs along the valleys (Frye 1957). The Baker et al. (2009) study of southeastern Nebraska sites showed the same pattern of upland prairie vegetation with little-to-no arboreal elements and valley riparian woodlands and marshlands. Another study looking into the latest Pleistocene, post-glaciation, found that mollusc taxa almost entirely indicated open prairie except for a few species, such as Discus whitneyi and D. shimekii, that require at least partial arboreal cover (Rossignol et al. 2004). Nebraska provides one of the highest resolution samples for this study, with one hundred and twenty-six specimens from M. primigenius, M. columbi, M. ?jeffersonii, M. ?imperator, and undetermined members of the genus. These specimens come from the entirety of the state, largely as individual loose teeth from non-associated sites. The average mesowear measurement of the state is 116.1° and all five categories of feeding habits are represented. The majority of the specimens exhibit graze- dominated feeding habits (58), followed by browse-dominated (38), true mixed- feeding (19), pure-browse (8), and pure graze (3). This distribution supports the diverse yet predominantly open palaeoenvironments put forth by mollusc analyses, where herbaceous plants and grasses likely made up a majority of the preferred diets of proboscideans and shrubs and woody plants were available for supplementation. 55 5.17 Kansas Of the five Kansas specimens used in this study, only three had location data. Two came from the northern border of Kansas, one M. columbi and one M. ?imperator. One M. columbi individual came from central Kansas. Kansas was included in Frankel’s analysis of the Pleistocene palaeoecology of the Great Plains in which he described open upland environments and forested valleys and streams (1956). Both northern and central molluscan fossil deposits show assemblages that are, today, only found in arboreal environments (Wells et al. 1987). Paired with plant macrofossils, pollen data, and small mammal fauna, this assemblage is interpreted to represent a mixture of coniferous forests and aspen parklands. The authors equate this environment to the modern Rocky Mountains west of Kansas. Pol len data from two sites in northeastern Kansas suggested that following approximately 29,000 years ago, the environment in the region had begun to close into a glacial spruce forest (Baker et al. 2009). A study on the distribution of shrew species at a site in central Kansas further supports the idea of a relatively wooded environment. This site in Ellis County suggests riparian deciduous or coniferous forests were interspersed with meadows in the region (McMullen 1975). All five Kansan specimens are categorized in the three mixed-feeding categories based on mesowear. Their average mesowear measurement is 115.7°. Mammoths, of course, give a much less local ecological signal than shrews due to their larger range, but these specimens are a strong addition to the argument that Kansas was a mosaic of locally open and closed environments with both graze and browse material available. 5.18 Iowa The preservation of organic material in Iowa during the Early Pleistocene was hindered both by glacial advances and a generally more arid environment. Pollen and plant macrofossils dating back to approximately 16,000-18,000 years ago reveal tundra species such as sedges occurring alongside spruce, pine, and larch trees (Baker et al. 1986). Insects from the same study are found in the transition zone between tundra-forest environments today, such as those found in Yukon and Alaska. 56 The authors concluded that this environment in the Late Pleistocene would have tundra patches interspersed with forest elements resembling the modern subarctic and subalpine tree lines that are stunted by cold temperatures and high winds. The later study by Baker et al. (2009) found that an eastern site in Iowa also had a pollen assemblage supporting boreal elements and C3 plants. An examination of Pleistocene palaeosols in Iowa indicates that, similar to Nebraska, more forested environments gradually shifted to a system where upland environments were open and prairie-like grasslands with the forests restricted to valleys and alongside water sources by the Late Pleistocene (Woida & Thompson 1993). Not only is the average mesowear measurement for Iowa nearly identical to that of Nebraska at 116.2°, the distribution of the specimens across all five feeding categories is highly comparable. Here, the best-represented category was again graze-dominated feeding habits (26), followed by browse-dominated (11), true mixed-feeding (10), pure browse (6), and pure graze (3). The two populations seem to have inhabited similar environments and responded in like manners, further corroborating both the palaeoenvironmental analyses and the assumption of proboscidean behaviour in response to those environments. 5.19 Missouri Palaeoenvironmental analyses for northern Missouri are lacking, as much of the focus has been on better preserved deposits in the Ozarks region of southern Missouri. However, as the five M. columbi and nineteen M. primigenius specimens in this study come from the northern border of Missouri alongside Nebraska and Iowa, extrapolation from nearby studies is possible. Frankel (1956) included Missouri in his broad analysis of the Great Plains Pleistocene palaeoecology and concluded that it was part of the same system. Nearly the same pattern is repeated here: the 24 specimens have an average mesowear measurement of 116.1° and the majority of the specimens fall into the graze-dominated category (10), followed by true mixed-feeding (6), pure browse (5), pure graze (2), and browse dominated (1). It is thus likely that the environment here 57 matched the surrounding plains with a mixture of open grasslands and pockets of forested areas near water sources. 5.20 Mexico Only two Mexican specimens were measured for this study, one M. columbi and one M. sp. from Tamaulipas in eastern Mexico. A potential reconstruction by Rzedowski and Huerta in 1978 was further elaborated on in a study of biogeographical corridors from Mexico. This reconstruction shows that much of Tamaulipas in the Pleistocene was covered by thorn scrub forest or scrubland (Ceballos et al. 2010). These scrub forests are generally open and characterized by thorny trees, shrubs, and dry grasslands. These two specimens had measurements of 118.2° and 112.8°, making the average of the two 115.5°. While a small sample size, it does reflect this potential environmental reconstruction that would necessitate mixed feeding in a landscape of difficult-to-eat vegetation. 5.21 Illinois Illinois mammoths are underrepresented in this study, with only one M. columbi individual from the northern part of the state. Its average mesowear angle of 106.9° indicates a browse-dominated diet. The exact age of this specimen is unknown and makes correlating these results with palaeoecological data difficult. Saunders et al. investigated the environment partitioning between concurrent species of proboscideans between 21,228–12,944 years ago in northeastern and central Illinois (2010). The authors determined that both M. primigenius and M. ?jeffersonii were inhabiting open spruce or black ash woodlands at this time. A single data point is not enough for palaeoenvironmental interpretations on its own, but it does not disagree with Saunders et al. interpretation of an environment that provided browse material. 5.22 Kentucky Stable carbon isotope data taken from bulk organic matter within the assemblage containing the mammoths at Big Bone Lick, Kentucky show an environment dominated by C3 plants (Tankersley et al. 2015). Pollen data is lacking for the site, but nearby archaeological remains spanning the Pleistocene and Holocene support an environment of mixed C3-C4 vegetation that is dominated by C3 plants. These authors suggest that this represents a suite of plants that reacted to climate change in 58 various ways and possibly at various intensities that would have transformed the landscape into a dynamic mosaic of vegetation throughout time. The three M. columbi specimens from Big Bone Lick in this study have an average mesowear angle of 113.3°, placing them at the margin of browse-dominated and mixed feeding. This is consistent with an environment dominated by browse plants, such as spruce and conifers, with a limited availability of graze material. 59 6 Conclusion Mesowear has long been a staple of palaeoecological studies involving large herbivores. Its easy-to-use, non-damaging, and robust nature make it an excellent option for studying the diets of prehistoric animals. In the case of this study, mesowear angle analysis has proven yet again to provide valuable insights on the palaeoecology of the herbivores of the past. These results indicate that mesowear angle analysis can be used as part of a multi- proxy approach to reconstruct palaeoenvironments and past vegetation regimes. For all instances where the sample size was large enough or specimen age and location information was accurate enough, the results of this study agreed with previously published studies involving stable isotope analysis, faunal community analysis, pollen data, plant macrofossils, and even mesowear studies involving other taxa. Furthermore, by comparing the results of these populations to each other on a global scale, this study was able to draw conclusions regarding the geospatial variation in mammoth diets during the Pleistocene. Clear differences between populations reflect differences in environment and highlight the adaptability of mammoths, M. primigenius in particular, to a wide array of habitats. Comparing M. primigenius to other Mammuthus species showed no variation in the adaptability of these species, implying that all of them were capable of utilizing the resources across different habitats to some degree. The geospatial variations and lack of interspecific variations found through this analysis are expected of a genus that is known for being able to adapt to survive some of the harshest environments of the last several million years by doing—and eating—whatever is necessary. 60 7 Acknowledgements My deepest thanks to all of the museum workers who made this possible—old friends and new. George Corner, Shane Tucker, and Ash Poust of the University of Nebraska State Museum—thanks for the warm welcome back and the enthusiastic support. The UNSM team will always have a special place in my heart. Tiffany Adrain of the University of Iowa Museum of Natural History—I was so glad to have the chance to work with you, and look forward to doing so again someday. Björn Kröger and Janne Granroth of the Finnish Museum of Natural History—special thanks for all you’ve done to support me during my time in Helsinki, but with this research in particular. Sami Viljanmaa of Kierikki Stone Age Centre—thank you for the lovely tour and accommodation for our strange one-day roadtrip to Oulu. And finally, Karin Truuver of the Estonian Museum of Natural History—thanks for the last-minute planning and pleasant company. I would be remiss not to thank my parents for their continued support, not only in moving halfway across the world for a degree, but also in this research. I’m glad you always checked in to make sure I was still alive and well (and sane) throughout this process. Above all, I want to thank my supervisor, Juha Saarinen, for his role as a mentor, critic, travel companion, and fellow mammoth enthusiast. I’ve learned so much from you and I couldn’t imagine anyone better to have led this project. 61 References 2012. Habitats Classification Scheme, Version 3.1. IUCN, 1-14. Abraham, J.O., Rowan, J., O'Brien, K., Sokolowski, K.G. and Faith, J.T. 2024. Environmental context shapes the relationship between grass consumption and body size in African herbivore communities. Ecology and Evolution, 14(2). https://doi.org/10.1002/ece3.11050 Agenbroad, L.D. 2003. New absolute dates and comparisons for California’s Mammuthus exilis. Deinsea 9, 1-16. Agenbroad, L.D. 2005. North American proboscideans: Mammoths: The state of knowledge, 2003. Quaternary International 126, 73-92. https://doi.org/10.1016/j.quaint.2004.04.016 Agenbroad, L.D. and Mead, J.I. 1989. Quaternary geochronology and distribution of Mammuthus on the Colorado Plateau. Geology 17, 861-864. Aguirre, E. 1969. Evolutionary History of the Elephant: A tentative phylogeny of Elephantidae based on morphological and quantitative analysis is given. Science 164, 1366-1376. https://doi.org/10.1126/science.164.3886.1366 Allison, J., Godwin, H., Warren, S., and Hinton, M.A.C. 1952. Late-glacial deposits at Nazeing in the Lea valley, North London. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences 236, 169-240. https://doi.org/10.1098/rstb.1952.0002 Archibold, O. 1995. Mediterranean ecosystems. Ecology of world vegetation, Springer Dordrecht, 131-164. https://doi.org/10.1007/978-94-011-0009-0 Baker, R., Bettis III, E.A., Mandel, R., Dorale, J., and Fredlund, G. 2009. Mid- Wisconsinan environments on the eastern Great Plains. Quaternary Science Reviews 28, 873-889. https://doi.org/10.1016/j.quascirev.2008.12.021 Baker, R., Bettis III, E.A., Schwert, D., Horton, D., Chumbley, C., Gonzalez, L.A., and Reagan, M. 1996. Holocene paleoenvironments of northeast Iowa. Ecological Monographs 66, 203-234. https://doi.org/10.2307/2963475 Baker, R., Sanders Rhodes II, R., Schwert, D.P., Ashworth, A.C., Frest, T.J., Hallberg, G.R., and Janssens, J.A. 1986. A full-glacial biota from southeastern Iowa, USA. Journal of Quaternary Science 1, 91-107. https://doi.org/10.1002/jqs.3390010202 62 Baleka, S., Varela, L., Tambusso, P.S., Paijmans, J.L., Mothé, D., Stafford, T.W., Fariña, R.A., and Hofreiter, M. 2022. Revisiting proboscidean phylogeny and evolution through total evidence and palaeogenetic analyses including Notiomastodon ancient DNA. Iscience 25, 1-19. https://doi.org/10.1016/j.isci.2021.103559 Bertini, A. 2013. Climate and vegetation in the Upper Valdarno Basin (central Italy) as a response to Northern Hemisphere insolation forcing and regional tectonics in the late Pliocene-early Pleistocene. Italian Journal of Geosciences 132, 137-148. https://doi.org/10.3301/IJG.2012.18. Boeskorov, G.G., Mashchenko, E.N., Plotnikov, V.V., Shchelchkova, M.V., Protopopov, A.V., and Solomonov, N.G. 2016. Adaptation of the woolly mammoth Mammuthus primigenius (Blumenbach, 1799) to habitat conditions in the glacial period. Contemporary Problems of Ecology 9, 544-553. https://doi.org/10.1134/s1995425516050024. Botka, D. and Mészáros, L. 2018. Taxonomic and palaeoecological review of the Soricidae (Mammalia) fauna from the late Early Pleistocene Somssich Hill 2 locality (Villány Hills, Southern Hungary). Fragmenta Palaeontologica Hungarica 35, 143-151. https://doi.org/10.17111/FragmPalHung.2018.35.143 Bryson, R.A., Agenbroad, L.D., and DeWall, K.M. 2010. Paleoclimate modeling and paleoenvironmental interpretations for three instances of island dwelling mammoths. Quaternary International 217, 6-9. https://doi.org/10.1016/j.quaintt.2009.09.028 Calandra, I., Göhlich, U.B. and Merceron, G. 2008. How could sympatric megaherbivores coexist? Example of niche partitioning within a proboscidean community from the Miocene of Europe. Naturwissenschaften, 95, 831-838. https://doi.org/10.1007/s00114-008-0391-y Campbell, J. B. 1977. The Upper Palaeolithic of Britain: a study of man and nature in the Late Ice Age. Ceballos, G., Arroyo-Cabrales, J., and Ponce, E. 2010. Effects of Pleistocene environmental changes on the distribution and community structure of the mammalian fauna of Mexico. Quaternary Research 73, 464-473. https://doi.org/10.1016/j.yqres.2010.02.006. Cerling, T.E. 1999. Paleorecords of C4 plants and ecosystems. C4 Plant Biology, 445- 469. 63 Cole, M.M. 1986. The savannas, biogeography and geobotany. Academic Press, 438 pp. Coltrain, J.B., Harris, J.M., Cerling, T.E., Ehleringer, J.R., Dearing, M.-D., Ward, J., and Allen, J. 2004. Rancho La Brea stable isotope biogeochemistry and its implications for the palaeoecology of late Pleistocene, coastal southern California. Palaeogeography, Palaeoclimatology, Palaeoecology 205, 199- 219. https://doi.org.10.1016/j.palaeo.2003.12.008 Cruz, J.A., Wooldrich-Piña, G.A., Basanta, M.D., García-Castillo, M.G., and Parra- Olea, G. 2024. Reptiles as Paleoenvironmental Proxies and Their Association with the Climate. Past Environments of Mexico: Unveiling the Past Environments of a Megadiverse Country Through its Fossil Record. Springer, 273-287. https://doi.org/10.1007/978-3-031-51034-2_12 Damuth, J., and Janis, C. M. 2011. On the relationship between hypsodonty and feeding ecology in ungulate mammals, and its utility in palaeoecology. Biological Reviews, 86(3), 733-758. https://doi: 10.1111/j.1469- 185X.2011.00176.x DeNiro, M.J. and Epstein, S. 1978. Influence of diet on the distribution of carbon isotopes in animals. Geochimica et cosmochimica acta 42, 495-506. https://doi.org/10.1016/0016-7037(78)90199-0 Dillon, J.S., Stolze, S., and Larsen, A.K. 2018. Late Pleistocene pollen and plant macrofossils from a buried wetland deposit in the Platte River Valley, south- central Nebraska. Great Plains Research 28, 173-184. https://doi.org/10.1353/gpr.2018.0028 Ehleringer, J.R. and Monson, R.K. 1993. Evolutionary and ecological aspects of photosynthetic pathway variation. Annual Review of Ecology and Systematics 24, 411-439. Enk, J., Devault, A., Debruyne, R., King, C., Treangen, T., O'Rourke, D., Salzberg, S., Fisher, D., MacPhee, R., and Poinar, H. 2011. Complete Columbian mammoth mitogenome suggests interbreeding with woolly mammoths. Genome biology 12, 1-8. https://doi.org/10.1186/gb-2011-12-5-r51 Falconer, H. 1863. On the American fossil elephant of the regions bordering the Gulf of Mexico, (Elephas columbi, Falc.) with general observations on the living and extinct species. Natural History Review 3, 43-114. 64 Fortelius, M. and Solounias, N. 2000. Functional characterization of ungulate molars using the abrasion-attrition wear gradient: a new method for reconstructing paleodiets. American Museum Novitates 2000, 1-36. https://doi.org/10.1206/0003- 0082(2000)301%3C0001:FCOUMU%3E2.0.CO;2 Fortelius, M., Eronen, J., Jernvall, J., Liu, L., Pushkina, D., Rinne, J., Tesakov, A., Vislobokova, I., Zhang, Z. and Zhou, L. 2002. Fossil mammals resolve regional patterns of Eurasian climate change over 20 million years. Evolutionary Ecology Research, 4(7), 1005-1016. Frankel, L. 1956. Pleistocene geology and paleoecology of parts of Nebraska and adjacent areas. University of Nebraska-Lincoln, 277 pp. Frye, J.C. 1957. Ecological interpretations of Pliocene and Pleistocene stratigraphy in the Great Plains region. Virtual Landscapes of Texas, 1-11. http://dx.doi.org/10.26153/tsw/4945 Fuller, B.T., Fahrni, S.M., Harris, J.M., Farrell, A.B., Coltrain, J.B., Gerhart, L.M., Ward, J.K., Taylor, R., and Southon, J.R. 2014. Ultrafiltration for asphalt removal from bone collagen for radiocarbon dating and isotopic analysis of Pleistocene fauna at the tar pits of Rancho La Brea, Los Angeles, California. Quaternary Geochronology 22, 85-98. http://dx.doi.org/10.1016/j.quageo.2014.03.002 Fuller, B.T., Southon, J.R., Fahrni, S.M., Farrell, A.B., Takeuchi, G.T., Nehlich, O., Guiry, E.J., Richards, M.P., Lindsey, E.L., and Harris, J.M. 2020. Pleistocene paleoecology and feeding behavior of terrestrial vertebrates recorded in a pre- LGM asphaltic deposit at Rancho La Brea, California. Palaeogeography, Palaeoclimatology, Palaeoecology 537, 109383. http://dx.doi.org/10.1016/j.palaeo.2019.109383 Gaglioti, B.V., Barnes, B.M., Zazula, G.D., Beaudoin, A.B., and Wooller, M.J. 2011. Late Pleistocene paleoecology of arctic ground squirrel (Urocitellus parryii) caches and nests from Interior Alaska's mammoth steppe ecosystem, USA. Quaternary Research 76, 373-382. http://dx.doi.org/10.1016/j.yqres.2011.08.004 Genoni, L., Iacumin, P., Nikolaev, V., Gribchenko, Y., and Longinelli, A. 1998. Oxygen isotope measurements of mammoth and reindeer skeletal remains: an archive 65 of Late Pleistocene environmental conditions in Eurasian Arctic. Earth and Planetary Science Letters 160, 587-592. Graham, R.W. 1981. Preliminary report on late Pleistocene vertebrates from the Selby and Dutton archeological/paleontological sites, Yuma County, Colorado. University of Wyoming, Laramie, Contributions to Geology 20, 33-56. Graham, R.W. 1986. The Colby Mammoth Site: taphonomy and archaeology of a Clovis kill site in northern Wyoming (GC Frison and LC Todd, editors). University of New Mexico Press, 165-229. Graumlich, E.K. 2020. Paleoenvironmental Analysis Through Gastropods at the Page-Ladson (8JE591) Inundated Terrestrial Site, Northwest Florida, USA. Florida State University, 1-72. Green, J.L., DeSantis, L.R. and Smith, G.J., 2017. Regional variation in the browsing diet of Pleistocene Mammut americanum (Mammalia, Proboscidea) as recorded by dental microwear textures. Palaeogeography, Palaeoclimatology, Palaeoecology, 487, 59-70. https://doi.org/10.1016/j.palaeo.2017.08.019 Green, J.L. and Croft, D.A. 2018. Using Dental Mesowear and Microwear for Dietary Inference: A Review of Current Techniques and Applications. Methods in Paleoecology, 53-73. http://dx.doi.org/10.1007/978-3-319-94265-0_5 Guthrie, R.D. 1982. Mammals of the mammoth steppe as paleoenvironmental indicators. Paleoecology of Beringia, 307-326. Guthrie, R.D. 2001. Origin and causes of the mammoth steppe: a story of cloud cover, woolly mammal tooth pits, buckles, and inside-out Beringia. Quaternary Science Reviews, 20(1-3), 549-574. https://doi.org/10.1016/S0277- 3791(00)00099-8 Hager, M.W. 1974. Late Pliocene and Pleistocene history of the Donnelly Ranch vertebrate site, southeastern Colorado. University of Wyoming, 62 pp. Harington, C.R., Tipper, H.W., and Mott, R.J. 1974. Mammoth from Babine Lake, British Columbia. Canadian Journal of Earth Sciences 11, 285-303. https://doi.org/10.1139/e74-025. Helmens, K.F., Risberg, J., Jansson, K.N., Weckström, J., Berntsson, A., Tillman, P.K., Johansson, P.W., and Wastegård, S. 2009. Early MIS 3 glacial lake evolution, ice-marginal retreat pattern and climate at Sokli (northeastern Fennoscandia). Quaternary Science Reviews 28, 1880-1894. https://doi.org/10.1016/j.quascirev.2009.03.001 66 Henriksen, M., Mangerud, J., Matiouchkov, A., Murray, A.S., Paus, A., and Svendsen, J.I. 2008. Intriguing climatic shifts in a 90 kyr old lake record from northern Russia. Boreas 37, 20-37. https://doi.org/ 10.1111/j.1502-3885.2007.00007.x Hill, C.L. 2001. Pleistocene mammals of Montana and their geologic context. Museum of the Rockies Occasional Paper 3, 127-144. Hill, C. 2005. Late Pliocene and Pleistocene Geology and Paleontology in the Three Rivers Basin, Montana. Museum of the Rockies Occasional Paper 3, 113-125. Hubberten, H.W., Andreev, A., Astakhov, V.I., Demidov, I., Dowdeswell, J.A., Henriksen, M., Hjort, C., Houmark-Nielsen, M., Jakobsson, M., and Kuzmina, S. 2004. The periglacial climate and environment in northern Eurasia during the Last Glaciation. Quaternary Science Reviews 23, 1333-1357. https://doi.org.10.1016/j.quascirev.2003.12.012 Jones, D.B. and Desantis, L.R. 2017. Dietary ecology of ungulates from the La Brea tar pits in southern California: a multi-proxy approach. Palaeogeography, Palaeoclimatology, Palaeoecology 466, 110-127. http://dx.doi.org/10.1016/j.palaeo.2016.11.019 Kirillova, I.V., Borisova, O.K., Chernova, O.F., Van Kolfschoten, T., Van Der Lubbe, J.H., Panin, A.V., Pečnerová, P., Van Der Plicht, J., Shidlovskiy, F.K., and Titov, V.V. 2020. ‘Semi‐dwarf’ woolly mammoths from the East Siberian Sea coast, continental Russia. Boreas 49, 269-285. https://doi.org/10.1111/bor.12431 Kovács, J., Szabó, P., Kocsis, L., Vennemann, T., Sabol, M., Gasparik, M., and Virág, A. 2015. Pliocene and Early Pleistocene paleoenvironmental conditions in the Pannonian Basin (Hungary, Slovakia): Stable isotope analyses of fossil proboscidean and perissodactyl teeth. Palaeogeography, Palaeoclimatology, Palaeoecology 440, 455-466. https://doi.org/10.1016/j.palaeo.2015.09.019. Kovarovic, K., Su, D.F., and Lintulaakso, K. 2018. Mammal community structure analysis. Methods in paleoecology: Reconstructing Cenozoic terrestrial environments and ecological communities, 351-372. http://dx.doi.org/10.1007/978-3-319-94265-0_16 Kubiak, H. 1982. Morphological characters of the mammoth: an adaptation to the arctic-steppe environment. Paleoecology of Beringia 1982, 281-289. https://doi.org/10.1016/B978-0-12-355860-2.50028-4 67 Kurtén, B. and Anderson, E.A. 1980. Pleistocene Mammals of North America. Columbia University Press, New York, 442 pp. Lapointe, L., Talbot, J., Fortier, D., Fréchette, B., Strauss, J., Kanevskiy, M., and Shur, Y. 2017. Middle to late Wisconsinan climate and ecological changes in northern Alaska: Evidences from the Itkillik River yedoma. Palaeogeography, palaeoclimatology, palaeoecology 485, 906-916. https://doi.org/10.1016/j.palaeo.2017.08.006 Leidy, J. 1869. The extinct Mammalian fauna of Dakota and Nebraska. Journal of the Academy of Natural Sciences of Philidelphia 2, 257-401. Lister, A. and Sher, A. 2015. Evolution and dispersal of mammoths across the Northern Hemisphere. Science 350, 805-809. http://dx.doi.org/10.1126/science.aac5660 Lister, A.M. 2013. The role of behaviour in adaptive morphological evolution of African proboscideans. Nature 500, 331-334. https://doi.org/10.1038/nature12275. Lister, A.M. 2017. On the type material and evolution of North American mammoths. Quaternary International 443, 14-31. https://doi.org/10.1016/j.quaint.2017.02.027 Lister, A.M. and Sher, A.V. 2001. The origin and evolution of the woolly mammoth. Science 294, 1094-1097. https://doi.org/10.1126/science.1056370. Lister, A.M., Sher, A.V., van Essen, H., and Wei, G. 2005. The pattern and process of mammoth evolution in Eurasia. Quaternary International 126-128, 49-64. https://doi.org/10.1016/j.quaint.2004.04.014. Loponen, L. 2020. Diets of Miocene proboscideans from Eurasia. 54 pp. Lõugas, L., Ukkonen, P., and Jungner, H. 2002. Dating the extinction of European mammoths: new evidence from Estonia. Quaternary Science Reviews 21, 1347-1354. http://dx.doi.org/10.1016/S0277-3791(01)00098-1 Lozhkin, A., Anderson, P., Vartanyan, S., Brown, T., Belaya, B., and Kotov, A. 2001. Late Quaternary paleoenvironments and modern pollen data from Wrangel Island (Northern Chukotka). Quaternary Science Reviews 20, 217-233. http://dx.doi.org/10.1016/S0277-3791(00)00121-9 Lundelius Jr, E.L., Thies, K.J., Graham, R.W., Bell, C.J., Smith, G.J., and DeSantis, L.R. 2019. Proboscidea from the big cypress creek fauna, Deweyville 68 formation, Harris County, Texas. Quaternary International 530, 59-68. http://dx.doi.org/10.1016/j.quaint.2019.11.018 Madden, C.T. 1981. Mammoths of North America. University of Colorado at Boulder, 271 pp. Maglio, V.J. 1973. Origin and evolution of the Elephantidae. Transactions of the American Philosophical Society 63, 1-149. Mander, L. and Punyasena, S.W. 2018. Fossil pollen and spores in paleoecology. Methods in paleoecology: Reconstructing Cenozoic terrestrial environments and ecological communities, 215-234. http://dx.doi.org/10.1007/978-3-319- 94265-0_11 Mandryk, C.A. 1998. A geoarchaeological interpretation of the Lamb Spring site, Colorado. Geoarchaeology: An International Journal 13, 819-846. Maschenko, E. 2002. Individual development, biology and evolution of the woolly mammoth. Cranium 19, 4-120. Maschenko, E.N. and Kalmykov, N.P. 2009. Homology of the Elephantid dentition in the context of elephants (Mammalia, Proboscidea) evolution. Doklady Biological Sciences 425, 157-159. https://doi.org/10.1134/s0012496609020215. Matheus, P., Begét, J., Mason, O., and Gelvin-Reymiller, C. 2003. Late Pliocene to late Pleistocene environments preserved at the Palisades site, central Yukon River, Alaska. Quaternary Research 60, 33-43. http://dx.doi.org/10.1016/S0033-5894(03)00091-7 Maul, L.C. 2000. Arvicola mosbachensis (Schmidtgen 1911) of Mosbach 2: a basic sample for the early evolution of the genus and a reference for further biostratigraphical studies: with 6 text-figures, 2 tables and 5 plates. Senckenbergiana lethaea: wissenschaftliche Mitteilungen der Senckenbergischen Naturforschenden Gesellschaft 80, 129-147. http://dx.doi.org/10.1007/BF03043667 McMullen, T. 1975. Shrews from the Late Pleistocene of Central Kansas, with the Description of a New Species of Sorex. Journal of Mammalogy 56, 316. https://doi.org/10.2307/1379363. Meyer, H.W. and Weber, L. 1995. Florissant Fossil Beds: National Monument: Preservation of an Ancient Ecosystem. Rocks & Minerals 70, 232-239. 69 Michener, R. and Lajtha, K. 2008. Stable isotopes in ecology and environmental science. John Wiley & Sons, 316 pp. http://dx.doi.org/10.1002/9780470691854.ch10 Miller, I.M., Pigati, J.S., Anderson, R.S., Johnson, K.R., Mahan, S.A., Ager, T.A., Baker, R.G., Blaauw, M., Bright, J., and Brown, P.M. 2014. Summary of the Snowmastodon Project Special Volume: A high-elevation, multi-proxy biotic and environmental record of MIS 6–4 from the Ziegler Reservoir fossil site, Snowmass Village, Colorado, USA. Quaternary Research 82, 618-634. https://doi.org/10.1016/j.yqres.2014.07.004 Morgan, G.S. and Lucas, S.G. 2005. Pleistocene vertebrate faunas in New Mexico from alluvial, fluvial, and lacustrine deposits. Bulletin of the New Mexico Museum of Natural History and Science 28, 185-248. Morgan, G.S., Lucas, S.G., Land, L., Lueth, V., and Raatz, W. 2006. Pleistocene vertebrates from southeastern New Mexico. New Mexico Geological Survey Guidebook 57, 317-336. https://doi.org/10.56577/FFC-57.317 Nordt, L., Bongino, J., Forman, S., Esker, D., and Benedict, A. 2015. Late Quaternary environments of the Waco Mammoth site, Texas USA. Quaternary Research 84, 423-438. https://doi.org/10.1016/j.yqres.2015.10.003 Pendleton, R.A. 2017. Microfauna at Tse’K’wa: Paleoenvironmental Reconstruction in the Peace River Region, Northeast British Columbia. Simon Fraser University, 1-167. Polling, M., ter Schure, A.T., van Geel, B., van Bokhoven, T., Boessenkool, S., MacKay, G., Langeveld, B.W., Ariza, M., Van Der Plicht, H., Protopopov, A.V. and Tikhonov, A. 2021. Multiproxy analysis of permafrost preserved faeces provides an unprecedented insight into the diets and habitats of extinct and extant megafauna. Quaternary Science Reviews, 267. https://doi.org/10.1016/j.quascirev.2021.107084 Puchkovskaya, V. and Shpansky, A. 2023. Finds of skeletons of mammoths (Mammuthus primigenius) in the territory of the Tomsk Oblast. 65-70 [in Russian]. Pushkina, D., Bocherens, H., and Ziegler, R. 2014. Unexpected palaeoecological features of the Middle and Late Pleistocene large herbivores in southwestern Germany revealed by stable isotopic abundances in tooth enamel. Quaternary 70 International 339-340, 164-178. https://doi.org/10.1016/j.quaint.2013.12.033. Ratnikov, V. and Negrobov, O. 2019. Historical paleoecology of the Voronezh Oblast. VSU Publishing, Voronezh. 96 pp. [in Russian]. Richards, A., Gibbard, P., and Pettit, M. 1999. The sedimentology and palaeoecology of the Westleton member of the Norwich crag formation (early Pleistocene) at Thorington, Suffolk, England. Geological Magazine 136, 453-464. https://doi.org/10.1017/S0016756899002782 Rivals, F., Mol, D., Lacombat, F., Lister, A.M. and Semprebon, G.M. 2015. Resource partitioning and niche separation between mammoths (Mammuthus rumanus and Mammuthus meridionalis) and gomphotheres (Anancus arvernensis) in the Early Pleistocene of Europe. Quaternary International, 379, 164-170. Rossignol, J., Moine, O., and Rousseau, D.D. 2004. The Buzzard's Roost and Eustis mollusc sequences: comparison between the paleoenvironments of two sites in the Wisconsinan loess of Nebraska, USA. Boreas 33, 145-154. http://dx.doi.org/10.1111/j.1502-3885.2004.tb01136.x Rzedowski, J. and Huerta, L. 1978. Vegetación de México. Editorial Limusa, México. 432 pp. [in Spanish]. Saarinen, J., Karme, A., Cerling, T., Uno, K., Säilä, L., Kasiki, S., Ngene, S., Obari, T., Mbua, E., Manthi, F.K., and Fortelius, M. 2015. A New Tooth Wear–Based Dietary Analysis Method for Proboscidea (Mammalia). Journal of Vertebrate Paleontology 35. https://doi.org/10.1080/02724634.2014.918546. Saarinen, J. and Lister, A.M. 2016. Dental mesowear reflects local vegetation and niche separation in Pleistocene proboscideans from Britain. Journal of Quaternary Science 31, 799-808. http://dx.doi.org/10.1002/jqs.2906 Saarinen, J. and Lister, A.M. 2023. Fluctuating climate and dietary innovation drove ratcheted evolution of proboscidean dental traits. Nature ecology & evolution 7, 1490-1502. https://doi.org/10.1038/s41559-023-02151-4 Sanders, W. J. 2018. Horizontal tooth displacement and premolar occurrence in elephants and other elephantiform proboscideans. Historical Biology, 30(1-2), 137-156. Sanders, W. J. 2023. Evolution and Fossil Record of African Proboscidea. CRC Press. 71 Saunders, J. 1999. Morphometrical analyses of Mammuthus columbi from the Dent site, Weld county, Colorado. Deinsea 6, 55-78. Saunders, J.J., Grimm, E.C., Widga, C.C., Campbell, G.D., Curry, B.B., Grimley, D.A., Hanson, P.R., McCullum, J.P., Oliver, J.S., and Treworgy, J.D. 2010. Paradigms and proboscideans in the southern Great Lakes region, USA. Quaternary International 217, 175-187. https://doi.org/10.1016/j.quaint.2009.07.031 Schirrmeister, L., Siegert, C., Kuznetsova, T., Kuzmina, S., Andreev, A., Kienast, F., Meyer, H., and Bobrov, A. 2002. Paleoenvironmental and paleoclimatic records from permafrost deposits in the Arctic region of Northern Siberia. Quaternary International 89, 97-118. http://dx.doi.org/10.1016/S1040- 6182(01)00083-0 Schwartz-Narbonne, R., Longstaffe, F.J., Metcalfe, J.Z., and Zazula, G. 2015. Solving the woolly mammoth conundrum: amino acid (1)(5)N-enrichment suggests a distinct forage or habitat. Scientific Reports 5, 9791. https://doi.org/10.1038/srep09791. Serebryanny, L. and Malyasova, E. 1998. The Quaternary vegetation and landscape evolution of Novaya Zemlya in the light of palynological records. Quaternary International 45, 59-70. https://doi.org/10.1016/S1040-6182(97)00007-4 Shpansky, A. 2021. Of mammoths and their companions: paleoecology of mammoth fauna. Fiton XXI, Moscow. 152 pp. [in Russian]. Szpak, P., Gröcke, D.R., Debruyne, R., MacPhee, R.D.E., Guthrie, R.D., Froese, D., Zazula, G.D., Patterson, W.P., and Poinar, H.N. 2010. Regional differences in bone collagen δ13C and δ15N of Pleistocene mammoths: Implications for paleoecology of the mammoth steppe. Palaeogeography, Palaeoclimatology, Palaeoecology 286, 88-96. https://doi.org.10.1016/j.palaeo.2009.12.009. Tankersley, K.B., Murari, M.K., Crowley, B.E., Owen, L.A., Storrs, G.W., and Mortensen, L. 2015. Quaternary chronostratigraphy and stable isotope paleoecology of Big Bone Lick, Kentucky, USA. Quaternary Research 83, 479- 487. https://doi.org/10.1016/j.yqres.2015.01.009 Tarasov, P., Volkova, V., Webb III, T., Guiot, J., Andreev, A., Bezusko, L., Bezusko, T., Bykova, G., Dorofeyuk, N., and Kvavadze, E. 2000. Last glacial maximum biomes reconstructed from pollen and plant macrofossil data from northern 72 Eurasia. Journal of Biogeography 27, 609-620. https://doi.org/10.1046/j.1365-2699.2000.00429.x Ukkonen, P., Aaris-Sørensen, K., Arppe, L., Clark, P., Daugnora, L., Lister, A., Lougas, L., Seppä, H., Sommer, R., and Stuart, A. 2011. Woolly mammoth (Mammuthus primigenius Blum.) and its environment in northern Europe during the last glaciation. Quaternary Science Reviews 30, 693-712. http://dx.doi.org/10.1016/j.quascirev.2010.12.017 Ukkonen, P., Arppe, L., Houmark-Nielsen, M., Kjær, K.H., and Karhu, J.A. 2007. MIS 3 mammoth remains from Sweden—implications for faunal history, palaeoclimate and glaciation chronology. Quaternary Science Reviews 26, 3081-3098. https://doi.org/10.1016/j.quascirev.2007.06.021. Ukkonen, P., Lunkka, J.P., Jungner, H., and Donner, J. 1999. New radiocarbon dates from Finnish mammoths indicating large ice‐free areas in Fennoscandia during the Middle Weichselian. Journal of Quaternary science 14, 711-714. https://doi.org/10.1002/(SICI)1099-1417(199912)14:7%3C711::AID- JQS506%3E3.0.CO;2-E van der Valk, T., Pečnerová, P., Díez-del-Molino, D., Bergström, A., Oppenheimer, J., Hartmann, S., Xenikoudakis, G., Thomas, J.A., Dehasque, M., Sağlıcan, E. and Fidan, F.R. 2021. Million-year-old DNA sheds light on the genomic history of mammoths. Nature, 591(7849), 265-269. https://doi.org/10.1038/s41586- 021-03224-9 van Geel, B., Aptroot, A., Baittinger, C., Birks, H.H., Bull, I.D., Cross, H.B., Evershed, R.P., Gravendeel, B., Kompanje, E.J., and Kuperus, P. 2008. The ecological implications of a Yakutian mammoth's last meal. Quaternary Research 69, 361-376. http://dx.doi.org/10.1016/j.yqres.2008.02.004 van Geel, B., Fisher, D.C., Rountrey, A.N., van Arkel, J., Duivenvoorden, J.F., Nieman, A.M., van Reenen, G.B.A., Tikhonov, A.N., Buigues, B., and Gravendeel, B. 2011. Palaeo-environmental and dietary analysis of intestinal contents of a mammoth calf (Yamal Peninsula, northwest Siberia). Quaternary Science Reviews 30, 3935-3946. https://doi.org/10.1016/j.quascirev.2011.10.009. Wells, P., Stewart, J., and Johnson, W. 1987. Spruce charcoal, conifer macrofossils, and landsnail and small-vertebrate faunas in Wisconsinan sediments on the High Plains of Kansas. Kansas Geological Survey Guidebook 5, 129-140. 73 West, R.G., Lambert, C.A., and Sparks, B.W. 1964. Interglacial deposits at Ilford, Essex. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences 247, 185-212. https:// doi.org/10.1098/rstb.1964.0002 Woida, K. and Thompson, M.L. 1993. Polygenesis of a Pleistocene paleosol in southern Iowa. Geological Society of America Bulletin 105, 1445-1461. http://dx.doi.org/10.1130/0016- 7606(1993)105%3C1445:POAPPI%3E2.3.CO;2 Zimov, S.A., Zimov, N.S., Tikhonov, A.N., and Chapin, F.S. 2012. Mammoth steppe: a high-productivity phenomenon. Quaternary Science Reviews 57, 26-45. https://doi.org/10.1016/j.quascirev.2012.10.005. 74 Appendices Table 3. Total data used in study organized by population. NHMUK=National History Museum of the United Kingdom, HLMD=Hessichen Landesmuseum Darmstadt, NMM=Naturhistorisches Museum Mainz, Luomus=Finnish Museum of Natural History, Kierikkikeskus=Kierikki Stone Age Centre, EMNH= Estonian Museum of Natural History, UNSM=University of Nebraska State Museum, AMNH=American Museum of Natural History, UINHM= University of Iowa Natural History Museum, GCPM=George C Page Museum. * indicates data provided by Juha Saarinen. Specimen Information Tooth Species Population Age Mean Mesowear Angle Source NHMUK-PV (Robert Mutch collection, 386) right m2/3 M. meridionalis UK Early Pleistocene 105 NHMUK* NHMUK-PV M 18988 left M1 M. meridionalis UK Early Pleistocene 107.7 NHMUK* NHMUK-PV M 6804 left M2 M. meridionalis UK Early - Middle Pleistocene 108.6 NHMUK* NHMUK-PV M 18493 left M2 M. meridionalis UK Early Pleistocene 109 NHMUK* NHMUK-PV M 18982 right M2 M. meridionalis UK Middle Pleistocene? 111.3 NHMUK* NHMUK-PV M 6800 left m3 M. meridionalis UK Early - Middle Pleistocene 111.7 NHMUK* NHMUK-PV M 18016 left m3 M. meridionalis UK Middle Pleistocene 111.8 NHMUK* NHMUK-PV M 18102 left M3 M. meridionalis UK Early Pleistocene 113.1 NHMUK* NHMUK-PV M 6854 left M3 M. meridionalis UK Early Pleistocene 113.2 NHMUK* NHMUK-PV M 18495 left m2 M. meridionalis UK Early Pleistocene 113.4 NHMUK* NHMUK-PV M 6788 left m1 M. meridionalis UK Middle Pleistocene 113.8 NHMUK* NHMUK-PV M 19040 left m2/3 M. meridionalis UK Middle Pleistocene 114.2333333 NHMUK* NHMUK-PV M 18103 right M3 M. meridionalis UK Early Pleistocene 114.325 NHMUK* NHMUK-PV M 6796 left M3 M. meridionalis UK Middle Pleistocene 114.5 NHMUK* NHMUK-PV M 6802 right m3 M. meridionalis UK Early - Middle Pleistocene 114.6 NHMUK* NHMUK-PV M 6858 right M1 M. meridionalis UK Early Pleistocene 114.6333333 NHMUK* NHMUK-PV M 6837 right m1 M. meridionalis UK Middle Pleistocene 115.2 NHMUK* NHMUK-PV M 6798 right M2 M. meridionalis UK Early - Middle Pleistocene 115.5 NHMUK* NHMUK-PV (Robert Mutch collection, 339) fragment M. meridionalis UK Early Pleistocene 116.7 NHMUK* NHMUK-PV M 18099 left m3 M. meridionalis UK Early Pleistocene 117.0333333 NHMUK* NHMUK-PV M 18099 left m3 M. meridionalis UK Early Pleistocene 117.0333333 NHMUK* NHMUK-PV M 18113 left m2 M. meridionalis UK Early - Middle Pleistocene 117.3333333 NHMUK* NHMUK-PV M 19008 left m1? M. meridionalis UK Early Pleistocene 118.2 NHMUK* NHMUK-PV M 18127 right m1? M. meridionalis UK Early - Middle Pleistocene 119.1 NHMUK* NHMUK-PV M 25206 (cast) fragment M. meridionalis UK Early Pleistocene 119.9 NHMUK* NHMUK-PV M 6816 ? m3 M. meridionalis UK Early Pleistocene 119.9333333 NHMUK* 75 NHMUK-PV M 6812 fragment M. meridionalis UK Middle Pleistocene 120.2 NHMUK* NHMUK-PV M 6801 left m2 M. meridionalis UK Middle Pleistocene 121.7333333 NHMUK* NHMUK-PV M 18979 right M2 M. meridionalis UK Early Pleistocene 121.8 NHMUK* NHMUK-PV M 18980 left m1/2 M. meridionalis UK Early - Middle Pleistocene 121.9 NHMUK* NHMUK-PV M 18101 left m2 M. meridionalis UK Early - Middle Pleistocene 124.1333333 NHMUK* NHMUK-PV M 18999 right m1 M. meridionalis UK Early - Middle Pleistocene 127.6666667 NHMUK* NHMUK-PV (Robert Mutch collection, S34) fragment M. meridionalis UK Early Pleistocene 128.4 NHMUK* NHMUK-PV (Robert Mutch collection, 576) right M2/3 M. meridionalis UK Early Pleistocene 130.2 NHMUK* NHMUK-PV (Robert Mutch collection, S79) right m2/3 M. meridionalis UK Early Pleistocene 130.5 NHMUK* NHMUK-PV (Robert Mutch collection, 44) fragment M. meridionalis UK Early Pleistocene 132 NHMUK* NHMUK-PV 16710 left m2 M. primigenius UK Late Pleistocene 103.7 NHMUK* NHMUK-PV (no number) right M2/3 M. primigenius UK Late Pleistocene 104.7 NHMUK* NHMUK-PV 16133 right m1 M. primigenius UK Late Pleistocene 105.2666667 NHMUK* NHMUK-PV (C.R. 62/886.072) left m3 M. primigenius UK Late Pleistocene 113.3 NHMUK* NHMUK-PV M 25041 right M3 M. primigenius UK Late Pleistocene 114.8 NHMUK* NHMUK-PV 16168 left m1 M. primigenius UK Late Pleistocene 117.95 NHMUK* NHMUK-PV 16711 right m2 M. primigenius UK Late Pleistocene 120.8666667 NHMUK* NHMUK-PV 16169 left m1 M. primigenius UK Late Pleistocene 123.3333333 NHMUK* NHMUK-PV 15007 right m1 M. primigenius UK MIS3 125.3 NHMUK* NHMUK-PV (no number) left m3 M. primigenius UK Late Pleistocene 126.3666667 NHMUK* NHMUK-PV 16182 left M1 M. primigenius UK Late Pleistocene 126.6 NHMUK* NHMUK-PV M 788 left m1 M. primigenius UK MIS3 129.225 NHMUK* NHMUK-PV M 20441 left M1 M. primigenius UK Late Pleistocene 129.5 NHMUK* NHMUK-PV M 21796 left m3 M. primigenius UK Late Pleistocene 130.2 NHMUK* NHMUK-PV (no number) left m3 M. primigenius UK Late Pleistocene 135.2 NHMUK* NHMUK-PV M 20444 right m3 M. primigenius UK Late Pleistocene 135.3 NHMUK* NHMUK-PV M 44677 right m3 M. primigenius UK Late Pleistocene 135.8 NHMUK* NHMUK-PV M 21799 right m2 M. primigenius UK Late Pleistocene 141.1 NHMUK* NHMUK-PV (no number, Brx.?) right M2 M. primigenius UK Late Pleistocene 141.8 NHMUK* NHMUK-PV (no number) right M1 M. primigenius UK Late Pleistocene 143.3 NHMUK* NHMUK-PV 16712 left m2 M. primigenius UK Late Pleistocene 146.5333333 NHMUK* NHMUK-PV M 6852 right M3 M. trogontherii UK Middle Pleistocene 102.4 NHMUK* NHMUK-PV M 18107 right M3 M. trogontherii UK Middle Pleistocene 102.6666667 NHMUK* NHMUK-PV M 6856 right M1 M. trogontherii UK Early - Middle Pleistocene 104.2 NHMUK* NHMUK-PV M 6883 left M2/3 M. trogontherii UK Middle Pleistocene 108.2 NHMUK* 76 NHMUK-PV M 6884 left m2 M. trogontherii UK Early - Middle Pleistocene 109.2666667 NHMUK* NHMUK-PV M 18987 right M2? M. trogontherii UK Early - Middle Pleistocene 111.2 NHMUK* NHMUK-PV M 18110 left M3 M. trogontherii UK Early Pleistocene 111.38 NHMUK* NHMUK-PV M 18141 right m2 M. trogontherii UK Early Pleistocene 112.0333333 NHMUK* NHMUK-PV 21315 a right m1 M. trogontherii UK MIS 7 113.1 NHMUK* NHMUK-PV M 44459 right M2 M. trogontherii UK MIS 7 114.1 NHMUK* NHMUK-PV M 6819 right m1 M. trogontherii UK Early - Middle Pleistocene 114.3333333 NHMUK* NHMUK-PV M 18093 right M3 M. trogontherii UK Middle Pleistocene 115.2 NHMUK* NHMUK-PV M 6121 left M2 M. trogontherii UK Middle Pleistocene 115.8 NHMUK* NHMUK-PV 40704 right M3 M. trogontherii UK MIS 7 116 NHMUK* NHMUK-PV M 18994 left M2 M. trogontherii UK Middle Pleistocene 117.5 NHMUK* NHMUK-PV 44996 right M1 M. trogontherii UK MIS 7 117.6666667 NHMUK* NHMUK-PV M 18094 right m3 M. trogontherii UK Middle Pleistocene 119.4 NHMUK* NHMUK-PV M 6124 left M1 M. trogontherii UK Early - Middle Pleistocene 119.4333333 NHMUK* NHMUK-PV M 18097 right m3 M. trogontherii UK Middle Pleistocene 120.2 NHMUK* NHMUK-PV M 6818 right M3 M. trogontherii UK Middle Pleistocene 120.6 NHMUK* NHMUK-PV M 6874 right M1 M. trogontherii UK Early - Middle Pleistocene 120.6 NHMUK* NHMUK-PV M 6823 left m3 M. trogontherii UK Middle Pleistocene 121.1 NHMUK* NHMUK-PV M 6851 left m2? M. trogontherii UK Early - Middle Pleistocene 121.7 NHMUK* NHMUK-PV 44944 left M1 M. trogontherii UK MIS 7 122.6 NHMUK* NHMUK-PV 44933 right M3 M. trogontherii UK MIS 7 123.2333333 NHMUK* NHMUK-PV 44996 left M1 M. trogontherii UK MIS 7 123.6333333 NHMUK* NHMUK-PV M 6841 right M1/M2 M. trogontherii UK Early - Middle Pleistocene 123.7 NHMUK* NHMUK-PV M 6862 right m1 M. trogontherii UK Middle Pleistocene 124.3 NHMUK* NHMUK-PV 44934 right M3 M. trogontherii UK MIS 7 125.6 NHMUK* NHMUK-PV M 6820 left M1 M. trogontherii UK Early - Middle Pleistocene 125.6333333 NHMUK* NHMUK-PV M 18092 right M3 M. trogontherii UK Middle Pleistocene 125.8 NHMUK* NHMUK-PV 21315 a right m1/2 M. trogontherii UK MIS 7 126.1333333 NHMUK* NHMUK-PV M 6815 right m2 M. trogontherii UK Middle Pleistocene 127 NHMUK* NHMUK-PV M 6815 left m2 M. trogontherii UK Middle Pleistocene 127.5333333 NHMUK* NHMUK-PV M 6892 left M1 M. trogontherii UK Early - Middle Pleistocene 127.8 NHMUK* NHMUK-PV M 19002 right m1? M. trogontherii UK Early - Middle Pleistocene 127.9333333 NHMUK* NHMUK-PV (no number) right m1 M. trogontherii UK MIS 7 129.1 NHMUK* NHMUK-PV M 6827 left m2 M. trogontherii UK Early - Middle Pleistocene 129.8666667 NHMUK* NHMUK-PV M 19032 right M1 M. trogontherii UK Early - Middle Pleistocene 129.9 NHMUK* NHMUK-PV 21315 left m1 M. trogontherii UK MIS 7 130.2666667 NHMUK* NHMUK-PV 44932 right M3 M. trogontherii UK MIS 7 130.8333333 NHMUK* NHMUK-PV 1604 left m3 M. trogontherii UK MIS 7 131.1333333 NHMUK* NHMUK-PV M 18112 left M1 M. trogontherii UK Early Pleistocene 131.4666667 NHMUK* 77 NHMUK-PV 44994 left m3 M. trogontherii UK MIS 7 132.2 NHMUK* NHMUK-PV 44935 left M3 M. trogontherii UK MIS 7 132.5 NHMUK* NHMUK-PV 1604 right m3 M. trogontherii UK MIS 7 133.2 NHMUK* NHMUK-PV 44991 right m1 M. trogontherii UK MIS 7 133.5 NHMUK* NHMUK-PV M 18998 left m1 M. trogontherii UK Early - Middle Pleistocene 134.7 NHMUK* NHMUK-PV 21315 right M1/2 M. trogontherii UK MIS 7 134.7 NHMUK* NHMUK-PV 44937 left M3 M. trogontherii UK MIS 7 136.2666667 NHMUK* NHMUK-PV M 19001 right m2/3 M. trogontherii UK Middle Pleistocene 136.4 NHMUK* NHMUK-PV 44990 right m1 M. trogontherii UK MIS 7 139.2 NHMUK* NHMUK-PV 44991 left m1/2 M. trogontherii UK MIS 7 143.9 NHMUK* NHMUK-PV M 6846 left M2 M. trogontherii UK Middle Pleistocene 144.3333333 NHMUK* HLMD-Mb.60 left m3 M. trogontherii GER Middle Pleistocene 99.36 HLMD* NMM 1961/300 right M2 M. trogontherii GER Middle Pleistocene 100.1 NMM* HLMD-Mb.958 left m2? M. trogontherii GER Middle Pleistocene 105.125 HLMD* HLMD-Mb.1006 left m2 M. trogontherii GER Middle Pleistocene 107.2 HLMD* NMM 1954/540 right M1? M. trogontherii GER Middle Pleistocene 107.28 NMM* HLMD-Mb.1006 right m2 M. trogontherii GER Middle Pleistocene 111.4 HLMD* NMM 1960/117 right m3 M. trogontherii GER Middle Pleistocene 116.125 NMM* HLMD-Mb.973 fragment M. trogontherii GER Middle Pleistocene 117.1 HLMD* NMM 1968/208a right M3 M. trogontherii GER Middle Pleistocene 117.6666667 NMM* HLMD-Mb.59a fragment M. trogontherii GER Middle Pleistocene 118.9 HLMD* NMM 1914/285 left m3 M. trogontherii GER Middle Pleistocene 121.8 NMM* NHMUK-PV 34337 (cast) right m2 M. meridionalis ITA Early Pleistocene 101.1 NHMUK* NHMUK-PV 28820 left M3 M. meridionalis ITA Early Pleistocene 103 NHMUK* NHMUK-PV (label badly worn) left m2-m3 M. meridionalis ITA Early Pleistocene 109.7 NHMUK* NHMUK-PV (cast, no number) left M1/2 M. meridionalis ITA Early Pleistocene 111.875 NHMUK* NHMUK-PV (cast, label badly worn) left m2-m3 M. meridionalis ITA Early Pleistocene 118.5 NHMUK* NHMUK-PV (cast, label badly worn) right m2-m3 M. meridionalis ITA Early Pleistocene 119.1333333 NHMUK* NHMUK-PV 3412 (syntype) right m1 M. meridionalis ITA Early Pleistocene 120.4 NHMUK* HLMD-WT.428 left M3 M. meridionalis HUN Early Pleistocene 99.63333333 NHMUK* HLMD-WT.603 right M3 M. meridionalis HUN Early Pleistocene 102.3333333 NHMUK* HLMD-WT.605 left M3 M. meridionalis HUN Early Pleistocene 106.425 NHMUK* HLMD-WT.606 left M3 M. meridionalis HUN Early Pleistocene 111.02 NHMUK* NRM P left M3 M. primigenius FIN Late Pleistocene 130.7666667 Kierikkikeskus KN 39971 Right M3 M. primigenius FIN Late Pleistocene 115.1 Luomus KN 39975 Left m2 M. primigenius FIN 112.95 Luomus G21:4 right M2/3 M. primigenius EST 117.25 EMNH G319:1 left m2 M. primigenius EST 119.4 EMNH G139:1 left M2/3 M. primigenius EST 120.02 EMNH 78 G319:2 right M2 M. primigenius EST 121.4 EMNH NHMUK-PV M 12357 right M3 M. primigenius RUS Late Pleistocene 92.2 NHMUK* NHMUK-PV M 4008 right m2-m3 M. primigenius RUS Late Pleistocene 116.8 NHMUK* NHMUK-PV M 82405 right m3 M. primigenius RUS Late Pleistocene 120.1 NHMUK* NHMUK-PV M 400 left m2-m3 M. primigenius RUS Late Pleistocene 120.5666667 NHMUK* NHMUK-PV M 3411 right m3 M. primigenius RUS Late Pleistocene 121.1 NHMUK* NHMUK-PV (no number) left m3 M. primigenius RUS Late Pleistocene 124.7 NHMUK* NHMUK-PV M 10965 left m3 M. primigenius RUS Late Pleistocene 137.3 NHMUK* NHMUK-PV M 10965 right m3 M. primigenius RUS Late Pleistocene 147 NHMUK* P1003 left M? M. primigenius RUS Quaternary 106.8666667 Luomus G21:29 left M2 M. primigenius RUS 122 EMNH TAM G441:73 left m2 M. primigenius RUS 131.6 EMNH G21:1 right M3 M. primigenius RUS 120.3 EMNH P26922 right M3 M. sp. RUS Quaternary 109.8 Luomus G285:559 right M3 M. sp. RUS 121.5 EMNH G134:1 right m1/2 M. sp. RUS 117.66 EMNH P26915 left M2 M. sp. RUS Quaternary 115.6 Luomus 93 left m? M. sp. RUS 127 Luomus 98 left M? M. sp. RUS 105.2666667 Luomus P26013 right m2? M. sp. RUS Quaternary 121.6333333 Luomus P26510 left M? M. sp. RUS Quaternary 114.4666667 Luomus UNSM 2041 right M3 M. primigenius AK 120.5 UNSM UNSM 1331 left M3 M. primigenius AK 129.4 UNSM UNSM 1937-1 left m3 M. primigenius AK 128.04 UNSM UNSM 1937-2 right m3 M. primigenius AK 125.48 UNSM AMNH 206 (A-Bx 46-U.CL-1934) right m3 M. primigenius AK Late Pleistocene 112.1 AMNH* NHMUK-PV 04077 left M3 M. primigenius AK Late Pleistocene 123.7 NHMUK* NHMUK-PV M 13240 left M2 M. primigenius AK Late Pleistocene 125.2 NHMUK* AMNH 211 (A-Fox 1933) left m3 M. primigenius AK Late Pleistocene 127.6666667 NHMUK* NHMUK-PV M 10500 left M3 M. primigenius AK Late Pleistocene 130.2 NHMUK* AMNH 216 (A-Bx 60-BAN.CR.-1934) left m3 M. primigenius AK Late Pleistocene 134.2 AMNH* NHMUK-PV M 13717 left m2 M. primigenius AK Late Pleistocene 141.06 NHMUK* 567 right M3 M. primigenius CAN Pleistocene 132 UIMNH PM 3805-10-L left m2 M. columbi CA Late Pleistocene 119 GCPM* PM 3804-4-R right m3 M. columbi CA Late Pleistocene 121.25 GCPM* PM 3805-15-L left m3 M. columbi CA Late Pleistocene 121.8333333 GCPM* PM 3805-8-L left m3 M. columbi CA Late Pleistocene 123.2 GCPM* PM 3805-1-L left m3 M. columbi CA Late Pleistocene 125.6 GCPM* PM Z7004 right m3 M. columbi CA Late Pleistocene 125.75 GCPM* 79 PM 3805-6-R right m3 M. columbi CA Late Pleistocene 128 GCPM* PM 3805-4-R right m2 M. columbi CA Late Pleistocene 131.3333333 GCPM* PM 3895-4-L left m3 M. columbi CA Late Pleistocene 131.8 GCPM* LACMHZ7006 right m3 M. columbi CA Late Pleistocene 132.6 GCPM* LACMHG68179 left m3 M. columbi CA Late Pleistocene 132.75 GCPM* PM 3804-2-R right m2 M. columbi CA Late Pleistocene 133.25 GCPM* LACMHG68177 right m3 M. columbi CA Late Pleistocene 135 GCPM* PM 3804-8-R right m2 M. columbi CA Late Pleistocene 135.25 GCPM* PM 3802-7-L left m2 M. columbi CA Late Pleistocene 135.5 GCPM* PM 3804-5-R right m3 M. columbi CA Late Pleistocene 136 GCPM* PM 3804-11-R right m3 M. columbi CA Late Pleistocene 136.2 GCPM* PM 3805-16-L left m2 M. columbi CA Late Pleistocene 137.3333333 GCPM* PM 3805-11-L left m3 M. columbi CA Late Pleistocene 137.6 GCPM* PM 3804-12-R right m3 M. columbi CA Late Pleistocene 137.8 GCPM* LACMH68194 left m2 M. columbi CA Late Pleistocene 138 GCPM* PM 3804-9-R right m2 M. columbi CA Late Pleistocene 138.1666667 GCPM* PM 3805-5-L left m3 M. columbi CA Late Pleistocene 138.5714286 GCPM* PM 3805-L-13 left m3 M. columbi CA Late Pleistocene 140.5 GCPM* LACMHC68176 right m3 M. columbi CA Late Pleistocene 141 GCPM* UNSM 53-68-1 right M3 M. sp. MT 132.1 UNSM UNSM 53-68-2 left m3 M. sp. MT 134.3 UNSM UNSM 53-68-3 right m3 M. sp. MT 136.325 UNSM UNSM 2088 left M/m? M. sp. CO 110.0666667 UNSM UNSM 144258 left m3 M. columbi NM Late Rancholabrean 97.2 UNSM UNSM 15-53-1 left m? M. sp. NM 116.66 UNSM UNSM 15-53-2 right m? M. sp. NM 118 UNSM UNSM 2193 fragment M. sp. SD 124 UNSM UNSM 2131 left m1 M. columbi NE Pleistocene 119.6 UNSM UNSM 2197 right M3 M. columbi NE Pleistocene 123.8 UNSM UNSM 1922 right M3 M. columbi NE 124.9 UNSM UNSM 2038 left m3 M. columbi NE Pleistocene 120.36 UNSM UNSM 2119 right M3 M. columbi NE 114.1 UNSM UNSM 1895 left m3 M. columbi NE Pleistocene 127 UNSM UNSM 5458 right m2 M. columbi NE 120.2 UNSM UNSM 5442 left m3 M. columbi NE 123.3 UNSM UNSM 1365-78 left m1 M. columbi NE 120.7 UNSM UNSM 2051 left M3 M. columbi NE 122.74 UNSM UNSM 5449-1 left m3 M. columbi NE 107.18 UNSM UNSM 5449-2 right m3 M. columbi NE 107.28 UNSM 80 UNSM 1004-023 right m3 M. columbi NE 115.2 UNSM UNSM 2071 right M2 M. columbi NE 110.14 UNSM UNSM 2056 left M3 M. columbi NE 119.625 UNSM UNSM 2249 right M2 M. columbi NE 111.2 UNSM UNSM 2179 right m2 M. columbi NE 116.9 UNSM UNSM 1785 left m3 M. columbi NE 120.18 UNSM UNSM 2124 left m1 M. columbi NE 114.22 UNSM UNSM 2022 right M3 M. columbi NE 106.42 UNSM UNSM 1787 right m3 M. columbi NE 119.35 UNSM UNSM 2545 left M3 M. columbi NE 121.38 UNSM UNSM 1768 right m3 M. columbi NE 120.4 UNSM UNSM 2158 left M? M. columbi NE 115.44 UNSM UNSM 1320 right m3 M. columbi NE 111.8 UNSM UNSM 1229-014 left m3 M. columbi NE Early Irvingtonian 107.4 UNSM UNSM 2063 right M3 M. columbi NE 104.15 UNSM UNSM 2230 right m1 M. columbi NE 120.46 UNSM UNSM 2060 left m3 M. columbi NE 111.38 UNSM UNSM 2102 left m3 M. columbi NE 127.96 UNSM UNSM 2133 right M3 M. columbi NE 112 UNSM UNSM 1818 right M3 M. columbi NE Pleistocene 119.9 UNSM AMNH AINS 134 (1933) left m3 M. columbi NE Late Pleistocene 121.3 AMNH* UNSM 43-48 M3 M. ?imperator NE Pleistocene 112.6 UNSM UNSM 2059 m2 M. ?imperator NE Pleistocene 100.95 UNSM UNSM 2150 left m1 M. ?imperator NE Pleistocene 108.9 UNSM UNSM 2074 left m3 M. ?imperator NE Late Irvingtonian 122.7 UNSM UNSM 1319 right M3 M. ?imperator NE Late Irvingtonian 119.7 UNSM UNSM 2176 right m3 M. ?imperator NE Late Irvingtonian 115.7 UNSM UNSM 380-59 right M3 M. ?imperator NE Pleistocene 117.7333333 UNSM UNSM 1643-60 right m3 M. ?imperator NE Pleistocene 117.625 UNSM UNSM 2136 left m3 M. ?imperator NE 122.4 UNSM UNSM 125565 right M3 M. ?imperator NE Rancholabrean 119.7 UNSM UNSM 2132 right m2 M. ?imperator NE 99.3 UNSM UNSM 2000 right M2 M. ?imperator NE 110.8 UNSM UNSM 1299 left m2 M. ?imperator NE 123.2 UNSM UNSM 1313-1 left m3 M. ?imperator NE 126.2 UNSM UNSM 1313-2 right m3 M. ?imperator NE 126.2 UNSM UNSM 1262-1 left m3 M. ?imperator NE 116.48 UNSM UNSM 1262-2 right m3 M. ?imperator NE 124.32 UNSM UNSM 1271 left m3 M. ?imperator NE 115.7 UNSM 81 UNSM 1507-1 left m2 M. ?imperator NE 126.3 UNSM UNSM 1507-2 right m2 M. ?imperator NE 126.74 UNSM UNSM 1305 right m3 M. ?imperator NE 108.8 UNSM UNSM 2020 left M3 M. ?imperator NE Pleistocene 103 UNSM UNSM 2116 fragment M. ?imperator NE 120.0666667 UNSM UNSM 2483 left m3 M. ?imperator NE 110.76 UNSM UNSM 2237 right M1/2 M. ?imperator NE 115.72 UNSM UNSM 1284 right M3 M. ?imperator NE 117.3 UNSM UNSM 1938 right m2 M. ?imperator NE 116.9 UNSM UNSM 2033 right M2 M. ?imperator NE 121.5666667 UNSM UNSM 45931 right m2 M. ?imperator NE 113.58 UNSM UNSM 2077 right m3 M. ?imperator NE 110.14 UNSM UNSM 1257 right M3 M. ?imperator NE 111.04 UNSM UNSM 2045 right M3 M. ?imperator NE 105.88 UNSM UNSM 2156 left M3 M. ?imperator NE 108.42 UNSM UNSM 1707 right m3 M. ?imperator NE 115.6 UNSM UNSM 2107 left m3 M. ?jeffersonii NE 119.6 UNSM UNSM 1328 left m3 M. ?jeffersonii NE 114.3 UNSM UNSM 2454 right m3 M. ?jeffersonii NE 111.04 UNSM UNSM 1261 right M3 M. primigenius NE 118.4 UNSM UNSM 1901 left m3 M. primigenius NE Pleistocene 118.075 UNSM UNSM 1304 right M3 M. primigenius NE 119 UNSM UNSM 2014 fragment M. primigenius NE 131.6 UNSM UNSM 2127 right M1 M. primigenius NE 124.52 UNSM UNSM 2134 right M3 M. primigenius NE 113.6 UNSM UNSM 2220 fragment M. primigenius NE 134.5 UNSM UNSM 2157 right M3 M. primigenius NE 116.175 UNSM UNSM 2188 right M3 M. primigenius NE 126.16 UNSM UNSM 2460 right m3 M. primigenius NE 121.98 UNSM UNSM 2097 left M2 M. primigenius NE 111.8 UNSM UNSM 2076 right m2 M. primigenius NE 130.6 UNSM UNSM 2103 right m3 M. primigenius NE Pleistocene 106.04 UNSM UNSM 1861 right m1 M. primigenius NE 120.08 UNSM UNSM 2142 right m3 M. sp. NE Pleistocene 124.46 UNSM UNSM 1916 left M2 M. sp. NE 109.85 UNSM UNSM 2453 right M3 M. sp. NE 106.52 UNSM UNSM 5444 right m2 M. sp. NE 118.86 UNSM No number right m1 M. sp. NE 125.5 UNSM UNSM 1852 fragment M. sp. NE 111.94 UNSM 82 UNSM 5456 right M3 M. sp. NE 112.62 UNSM UNSM 1860 right M3 M. sp. NE 117.04 UNSM UNSM 1868 fragment M. sp. NE 117.7 UNSM UNSM 2113 left m3 M. sp. NE 111.72 UNSM UNSM 1935-1 left M3 M. sp. NE 110 UNSM UNSM 1935-2 right M3 M. sp. NE 110 UNSM UNSM 63-56 left M2 M. sp. NE 117.16 UNSM UNSM 1735 right M3 M. sp. NE 128.7333333 UNSM UNSM 1725 fragment M. sp. NE 113.88 UNSM UNSM 1719 right M3 M. sp. NE 109.7 UNSM UNSM 1720 left M3 M. sp. NE 109.34 UNSM UNSM 1717 left m2 M. sp. NE 111.5 UNSM UNSM 1716 right M3 M. sp. NE 106.6 UNSM UNSM 1722 left M3 M. sp. NE 98.2 UNSM UNSM 48549 fragment M. sp. NE 110.7 UNSM UNSM 1779 fragment M. sp. NE 129.6333333 UNSM UNSM 1763 fragment M. sp. NE 116.68 UNSM UNSM 1762 fragment M. sp. NE 120.9 UNSM UNSM 1756 left M3 M. sp. NE 106.14 UNSM UNSM 1848 right m3 M. sp. NE 119.1 UNSM UNSM 1274 right M3 M. sp. NE 105.74 UNSM UNSM 2487 right M3 M. sp. NE 111.8 UNSM UNSM 1806 left m1 M. sp. NE 111.6 UNSM UNSM 1807 left m1 M. sp. NE 114.1 UNSM UNSM 133367 right M3 M. sp. NE Early Irvingtonian 107.05 UNSM UNSM 133363 left M2 M. sp. NE Early Irvingtonian 106.22 UNSM UNSM 133359 fragment M. sp. NE Early Irvingtonian 122.8 UNSM UNSM 1743 left M3 M. sp. NE 101.2 UNSM UNSM 2114 fragment M. sp. NE 113.76 UNSM UNSM 1826 left m2 M. sp. NE 123.5 UNSM UNSM 1821 right M3 M. sp. NE 110.1 UNSM UNSM 2200 right m2 M. sp. NE 123 UNSM UNSM 5454 right M3 M. sp. NE 121.5 UNSM UNSM 1819 right m3 M. sp. NE 119.2 UNSM UNSM 2092 left M3 M. sp. NE 117.4 UNSM UNSM 1812 left m1 M. sp. NE 121.6 UNSM UNSM 2110 right M? M. columbi KS 109.3 UNSM UNSM 2086 right M2/3 M. columbi KS 120.18 UNSM UNSM 2091 left M3 M. ?imperator KS 118.1 UNSM 83 UNSM 1275 right m3 M. sp. KS 116.1 UNSM UNSM 2057 fragment M. sp. KS 114.7 UNSM 420 left M1 M. columbi TX Pleistocene 128.78 UINHM NHMUK-PV M 10211 right m3 M. columbi TX 112.6 NHMUK* NHMUK-PV 39218 left m3 M. columbi TX 115.6 NHMUK* NHMUK-PV 20702 right m3 M. columbi TX 117.5666667 NHMUK* NHMUK-PV M 10211 fragment M. columbi TX 123 NHMUK* NHMUK-PV 20702 fragment M. columbi TX 123.38 NHMUK* UNSM 2148 left M? M. columbi MEX 118.2 UNSM UNSM 2137 right M3 M. sp. MEX 112.82 UNSM UNSM 144241 M/m3 frag M. columbi IA Late Rancholabrean 117.4 UNSM UNSM 145140 right M3 M. columbi IA Late Rancholabrean 104.2333333 UNSM UNSM 145268 right m3 M. columbi IA Late Rancholabrean 112.85 UNSM UNSM 144405 right m1 M. columbi IA 114.76 UNSM 14 right M M. columbi IA Pleistocene 108.9 UINHM 167 right M2 M. columbi IA Pleistocene 116.5 UINHM 309 right M1 M. columbi IA Pleistocene 131.98 UINHM 325 right m3 M. columbi IA Pleistocene 124.08 UINHM 411 right M3 M. columbi IA Pleistocene 117.56 UINHM 497 right M2 M. columbi IA Pleistocene 117.72 UINHM 553 right m2 M. columbi IA Pleistocene 125.3 UINHM 406 right M? M. columbi IA Quaternary 137.1666667 UINHM 47232 right m2 M. columbi IA Quaternary 126.88 UINHM UNSM 144245 left m2 M. primigenius IA Late Rancholabrean 101.3 UNSM UNSM 144257 right m3? M. primigenius IA Late Rancholabrean 105.8 UNSM UNSM 145134 right M3 M. primigenius IA Late Rancholabrean 116.2 UNSM UNSM 144276 right m3 M. primigenius IA Late Rancholabrean 113.25 UNSM UNSM 145342 right M3 M. primigenius IA Late Rancholabrean 117 UNSM UNSM 144163 left m3 M. primigenius IA Late Rancholabrean 114.9 UNSM UNSM 145945 left m3 M. primigenius IA Late Rancholabrean 117.04 UNSM UNSM 144265 left m2 M. primigenius IA Late Rancholabrean 110.5 UNSM UNSM 145265 left m3 M. primigenius IA Late Rancholabrean 113.3 UNSM UNSM 145350 left m3 M. primigenius IA Late Rancholabrean 89.55 UNSM USNM 145285 left M/m3 M. primigenius IA Late Rancholabrean 115.34 UNSM UNSM 144275 left M3 M. primigenius IA Late Rancholabrean 108.94 UNSM UNSM 145267 right m3 M. primigenius IA Late Rancholabrean 98.9 UNSM UNSM 144376 right m3 M. primigenius IA Late Rancholabrean 119.05 UNSM UNSM 145084 right m3 M. primigenius IA Late Rancholabrean 111.2 UNSM UNSM 144351 right m3 M. primigenius IA Late Rancholabrean 121.9666667 UNSM 84 UNSM 145284 left m2 M. primigenius IA Late Rancholabrean 120.5 UNSM UNSM 144681 left m2 M. primigenius IA Late Rancholabrean 108.6 UNSM UNSM 145175 left m2 M. primigenius IA Late Rancholabrean 106.22 UNSM UNSM 145244 right m2 M. primigenius IA Late Rancholabrean 117.5 UNSM UNSM 145345 right m2 M. primigenius IA Late Rancholabrean 107.75 UNSM UNSM 144266 right M1/2 M. primigenius IA Late Rancholabrean 118.34 UNSM UNSM 144271 right m3 M. primigenius IA Late Rancholabrean 95.84 UNSM UNSM 144261 right m3 M. primigenius IA Late Rancholabrean 125 UNSM UNSM 144247 left m2 M. primigenius IA Late Rancholabrean 111.5 UNSM 381 right M3 M. primigenius IA Pleistocene 108 UINHM 384 left M? M. primigenius IA Pleistocene 117.7 UINHM 21 left M2 M. primigenius IA Pleistocene 128.4 UINHM 61 right m2 M. primigenius IA Pleistocene 120.9 UINHM 324 right m1 M. primigenius IA Pleistocene 113.5 UINHM 357 left m3 M. primigenius IA Pleistocene 129.42 UINHM 492 left ? M. primigenius IA Pleistocene 120.3 UINHM 499 right m? M. primigenius IA Pleistocene 117.96 UINHM 517 Left M3 M. primigenius IA Quaternary 125.6 UINHM 578 right m3 M. primigenius IA Pleistocene 117.9 UINHM 32771 left M ?? M. primigenius IA Pleistocene 121.08 UINHM 33547 right M2 M. primigenius IA Pleistocene 113.4 UINHM 33551 left m1 M. primigenius IA Pleistocene 116.48 UINHM UNSM 144683 left M3 M. sp. IA Late Rancholabrean 123.1 UNSM UNSM 146668 right M2/3 M. sp. IA Late Rancholabrean 111.6333333 UNSM 43810 right m1 M. sp. IA Pleistocene 130.1 UINHM 32133 right m3 M. sp. IA Pleistocene 125.94 UINHM 53189 right M3 M. sp. IA Pleistocene 123 UINHM UNSM 145958 left m2 M. columbi MO Late Rancholabrean 100.5 UNSM UNSM 145947 left? m2 frag M. columbi MO Late Rancholabrean 116.3 UNSM UNSM 145944 left M? M. columbi MO Late Rancholabrean 117.9 UNSM UNSM 244250 left M3 M. columbi MO Late Rancholabrean 106.8 UNSM UNSM 144251 left M2 M. columbi MO Late Rancholabrean 117.34 UNSM UNSM 145339 right M2 M. primigenius MO Late Rancholabrean 114.34 UNSM UNSM 144682 left M2 M. primigenius MO Late Rancholabrean 113.05 UNSM UNSM 145956 left M?2 M. primigenius MO Late Rancholabrean 97.425 UNSM UNSM 144368 left m2 M. primigenius MO Late Rancholabrean 126.5 UNSM UNSM 144406 right M/m3 M. primigenius MO Late Rancholabrean 105.95 UNSM UNSM 144687 right M3 M. primigenius MO 103.7666667 UNSM UNSM 144256 right M3 M. primigenius MO Late Rancholabrean 131.45 UNSM 85 UNSM 144255 left m3 M. primigenius MO Late Rancholabrean 114.75 UNSM UNSM 145348 left M3 M. primigenius MO Late Rancholabrean 96.8 UNSM UNSM 145962 right m3 M. primigenius MO Late Rancholabrean 114.42 UNSM UNSM 144688 right m3 M. primigenius MO Late Rancholabrean 116.4 UNSM UNSM 144029 right m3 M. primigenius MO Late Rancholabrean 122.96 UNSM UNSM 144263 left m2 M. primigenius MO Late Rancholabrean 124.2 UNSM UNSM 144253 left m2 M. primigenius MO Late Rancholabrean 117.2 UNSM UNSM 145644 right m2 M. primigenius MO Late Rancholabrean 126.25 UNSM UNSM 145341 right m2 M. primigenius MO Late Rancholabrean 129 UNSM UNSM 144242 right m2 M. primigenius MO Late Rancholabrean 130.7 UNSM UNSM 144596 right m1 M. primigenius MO Late Rancholabrean 122.5 UNSM UNSM 144375 left M3 M. primigenius MO Late Rancholabrean 119.3 UNSM UNSM 1317 left m3 M. columbi IL 106.92 UNSM UNSM 133083 left m3 M. columbi KY Late Rancholabrean 119.96 UNSM USNM 133084 right M2/3 M. columbi KY Late Rancholabrean 119.4 UNSM UNSM - left M/m? M. columbi KY Late Rancholabrean 100.6 UNSM