Origins of Horse Herding and Transport in the Eastern Steppe

William T T Taylor

In the dry steppes of eastern Eurasia, domestic horses (E. caballus) provide the economic and cultural foundations of nomadic life. With no written records and sparse archaeological data from early nomadic societies, however, the ecological context of the first horse herding and transport, and its role in the formation of herding societies is poorly understood. Some of the earliest evidence for domestic horses in the region come from small ritual horse burials at sites belonging to the Deer Stone-Khirigsuur (DSK) culture, a late Bronze Age cultural often linked with the first mobile pastoral societies in Mongolia. This dissertation employs archaeological and archaeozoological techniques to assess how DSK people used domestic horses, and evaluates the role of horse herding and transport in the emergence of mobile herding in eastern Eurasia. I present results in five discrete published studies. The first study identifies evidence for selective culling of young and old animals as part of maintaining a breeding herd, with the selective burial of adult male transport horses in prominent ritual mounds along the eastern perimeter of DSK monument sites. A second set of three closely related studies investigates the skulls of contemporary wild and domestic horses, identifying anthropogenic changes to the equine skull caused by exertion, bridling, and pressures related to horseback riding. Applying these criteria to the late Bronze Age DSK archaeological record indicates that DSK people bridled and used horses for transport, and may have engaged in early mounted horseback riding. Finally, a precision radiocarbon model suggests a rapid expansion of DSK horse use around ca. 1200 BCE – during a period of climate amelioration and increased rainfall, and concurrent with major changes in ritual practice and the spread of horses to new parts of the continent. These results provide compelling links between the adoption of horseback riding, new ecological opportunities, and the development of mobile pastoralism in eastern Eurasia.

William Taylor Candidate Anthropology Department This dissertation is approved, and it is acceptable in quality and form for publication: Approved by the Dissertation Committee: Emily Lena Jones, Chairperson E. James Dixon Bruce Huckell William Fitzhugh Melinda Zeder Sandra Olsen i THE ORIGINS OF HORSE HERDING AND TRANSPORT IN THE EASTERN STEPPE by WILLIAM TIMOTHY TREAL TAYLOR B.A., Carleton College, May 2011 M.S., University of New Mexico, May 2013 DISSERTATION Submitted in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy Anthropology The University of New Mexico Albuquerque, New Mexico May, 2017 ii DEDICATION Dedicated with love to Papa, Uncle Will, Nana, Grandma Ethel, and Cousin Dick. iii ACKNOWLEDGEMENTS There are so many people who deserve credit for helping me through this dissertation that I will ultimately fall far short of thanking all those who deserve to be thanked. That being said, I am especially grateful to my wonderful parents, Barbara Morrison and James Park Taylor, who since I decided to be an archaeologist at the age of five or six, never wavered in their love and support. Many people put their blood, sweat, and tears into this document – especially Dr. Emily Lena Jones and Dr. E. James Dixon, who are umatched in their patience and tenacity, and Dr. William Fitzhugh who inspired me to study in Mongolia. Many others helped me shape these crazy ideas into something useful, especially Jacqueline Marie Kocer, Dr. Bruce Huckell, Dr. Melinda Zeder, Dr. Lawrence Guy Straus, Dr. Sandra Olsen, Dr. Jeffrey Long, Dr. William Honeychurch, Dr. Joshua Wright, Dr. Robin Bendrey, Dr. Michael Dee, Dr. Julia Clark, Dr. JeanLuc Houle, my uncle Dr. Rodney Flint Taylor, Dr. Scott Bender, Dr. Dave Fly, Mi hael Gag oose Reese, and my talented sister Dr. Jocelyn Whitworth. My list of Mongolian mentors, who taught me most of what I have learned and guided the ideas formulated here, includes my older brothers Dr. J. Bayarsaikhan, Dr. Z.Batsaikhan, Dr. Ts. Turbat, Dr. Ts. Egiimaa, my dearest friend and colleague T. Tuvshinjargal, B. Jargalan, and many others. The research would not have been possible without the generous financial and logistical support of several funding agencies. The project was supported by the US. Embassy in Mongolia and the A assado s Lo al G a ts P og a , National Science Foundation Doctoral Dissertation Improvement Grant #15222024, National Geographic Young Explorer Grant #9713-15, and Fulbright U.S. Student Award #34154234. The first, and most challenging several years were funded almost exclusively by the American Center for Mongolian Studies, through their Summer Research and Li a Fello ship p og a s. I d like to e te d a spe ial tha ks to Ma Tasse a d Baigalmaa Begszuren for their help in solving some of the many challenges that arose over the years. Additional monetary support from the Frison Institute for Archaeological Research, the International Council for Archaeozoology (ICAZ), the Alexandria Archive, the Society for Archaeological Sciences, the University of New Mexico s Department of Anthropology, and the UNM Graduate and Professional Student Association was also invaluable. The list of people who provided material support for the research is also long and star-studded. For several years, the National Museum of Mongolia has provided research space and graciously granted access to their collections. Special thanks to Dr. Greg Hodgins, his wife Gretchen, and his staff at the Accelerator Mass Spectrometry Lab at the University of Arizona for providing training and material support for radiocarbon analysis, as well as a beautiful home with a loving pair of dogs to stay with. Sukhtulga Tserennadmid helped to facilitate the study of Przewalski horses at Khustai Nuruu National Park, while Mary Thompson helped with analysis of collections from the Department of Interior, Bureau of Reclamation, Bureau of Land Management, and Idaho Museum of Natural History (Pleistocene and Pliocene equids). Dr. Bill Cooke and Amy Biesel at the International Museum of the Horse in Lexington were immensely supportive during my research visit, as were the staff and faculty at the Smithsonian National Museum of Natural History and the Museum Support Center. Dr. Joseph Cook and Dr. Jon Dunnum provided professional guidance and helped access the natural history collections at the Museum of Southwestern Biology. Photographs and insights into Mongolian and Altai horses were provided generously by Nils Larsen, Mark Jenkins, and Orsoo Bayarsaikhan. Friendship and moral support iv came from many wonderful people, but chief among them were Gandhi Yetish, Brian Geer, Avery Carroll, Matthew Park Taylor, Cecily Whitworth, and my sweet monster Logan. 3D scanning equipment for this research was provided by Dr. Heather Edgar at the Osteology Laboratory, as well as Dr. Emily Lena Jones and the Zooarchaeology Laboratory at the University of New Mexico. Mike Rendina helped to design support structures for 3D data collection, and Dr. Heather Edgar and Katelyn Rusk provided valuable wisdom and methodological advice on 3D analysis. The detailed comments of several anonymous reviewers greatly improved the organization and clarity of the chapters contained within. Finally, I owe a deep debt of thanks to Cliffe Arrand and Enkhzaya Burged for keeping me well fed and caffeinated, and the talented B. Khorloo provided a source of creative inspiration at a crucial moment during the long and arduous writing journey. NOTE TO THE READER This disse tatio is su itted follo i g guideli es fo U i e sit of Ne Me i o s o t aditio al o h id disse tatio , hi h o sists of a i t odu tio , a se ies of elated, independently published manuscripts, and a terminal synthesis/conclusion. Because each chapter is presented as published, separate sections may repeat information already mentioned elsewhere, and transitions may be abrupt. v THE ORIGINS OF HORSE HERDING AND TRANSPORT IN THE EASTERN STEPPE By William Timothy Treal Taylor B.A., Carleton College, 2011 M.S., University of New Mexico, 2013 Ph.D., Anthropology, University of New Mexico, 2017 ABSTRACT In the dry steppes of eastern Eurasia, domestic horses (E. caballus) provide the economic and cultural foundations of nomadic life. With no written records and sparse archaeological data from early nomadic societies, however, the ecological context of the first horse herding and transport, and its role in the formation of herding societies is poorly understood. Some of the earliest evidence for domestic horses in the region come from small ritual horse burials at sites belonging to the Deer Stone-Khirigsuur (DSK) culture, a late Bronze Age cultural often linked with the first mobile pastoral societies in Mongolia. This dissertation employs archaeological and archaeozoological techniques to assess how DSK people used domestic horses, and evaluates the role of horse herding and transport in the emergence of mobile herding in eastern Eurasia. I present results in five discrete published studies. The first study identifies evidence for selective culling of young and old animals as part of maintaining a breeding herd, with the selective burial of adult male transport horses in prominent ritual mounds along the eastern perimeter of DSK monument sites. A second set of three closely related studies investigates the skulls of contemporary wild and domestic horses, identifying anthropogenic changes to the equine skull caused by exertion, bridling, and pressures related to horseback riding. Applying these criteria to the late Bronze Age DSK archaeological record indicates that DSK people bridled and used horses for transport, and may have engaged in early mounted horseback riding. Finally, a precision radiocarbon model suggests a rapid expansion of DSK horse use around ca. 1200 BCE – during a period of climate amelioration and increased rainfall, and concurrent with major changes in ritual practice and the spread of horses to new parts of the continent. These results provide compelling links between the adoption of horseback riding, new ecological opportunities, and the development of mobile pastoralism in eastern Eurasia. vi TABLE OF CONTENTS Dedication .................................................................................................................................................... iii Acknowledgments........................................................................................................................................ iv Note to the reader ........................................................................................................................................ v Abstract ........................................................................................................................................................ vi Table of contents ........................................................................................................................................ vii List of figures ................................................................................................................................................. x List of tables .................................................................................................................................................xv Chapter: Introduction ................................................................................................................................... 1 1.1 Defining nomadic pastoralism ................................................................................................................ 1 1.2 Origins of horse transport in Central Asia............................................................................................... 2 1.3 Explanations for the development of nomadic pastoralism................................................................... 3 1.3.1 Sedentary agricultural societies ....................................................................................................... 3 1.3.2 Climate deterioration or resource scarcity ...................................................................................... 4 1.3.3 Human behavioral ecology and niche construction ........................................................................ 5 1.4 Late Bronze Age mobile pastoralism .................................................................................................. 5 1.5 Research design ...................................................................................................................................... 6 1.6 Summary ................................................................................................................................................. 9 Chapter 2: Horse demography and use in Bronze Age Mongolia ................................................................. 9 2.1. Introduction ......................................................................................................................................... 10 2.2 Regional setting .................................................................................................................................... 10 2.3. Materials and methods ........................................................................................................................ 16 2.4. Results .................................................................................................................................................. 21 2.4.1. Mortality profiles .......................................................................................................................... 21 2.4.2. Spatial patterning.......................................................................................................................... 23 2.4.3. Nasal pathologies .......................................................................................................................... 23 2.5. Discussion............................................................................................................................................. 23 2.5.1. Mortality profiles .......................................................................................................................... 23 2.5.2. Spatial patterns ............................................................................................................................. 28 2.5.3. Osteological data .......................................................................................................................... 28 2.5.4 Implications of DSK horse transport .............................................................................................. 29 2.6. Conclusion ............................................................................................................................................ 30 vii Chapter 3: Equine cranial morphology and the archaeological identification of riding and chariotry in Bronze Age Mongolia .............................................................................................................. 31 3.1 Introduction .......................................................................................................................................... 31 3.1.1 The horse in ancient Mongolia and beyond .................................................................................. 32 3.1.2 Archaeozoological identification of riding and chariotry............................................................... 33 3.2 Materials and methods ......................................................................................................................... 36 3.3 Results ................................................................................................................................................... 38 3.3.1 Nuchal ossification ......................................................................................................................... 38 3.3.2 Nasal remodelling .......................................................................................................................... 41 3.4 Archaeological applications .................................................................................................................. 44 3.5 Deer Stone-Khirigsuur results ............................................................................................................... 45 Chapter 4: Reconstructing equine bridles in the Mongolian Bronze Age................................................... 48 4.1 Introduction .......................................................................................................................................... 48 4.1.1 Late Bronze Age Archaeology and early horse use in Mongolia.................................................... 49 4.1.2 Ancient East and Central Asian Bridles .......................................................................................... 50 4.2 Reconstructing ancient bridles through equine osteology ................................................................... 52 4.2.1 Noseband use and nasal remodeling ............................................................................................. 53 . . Bit/ outhpie e use a d it ea .............................................................................................. 54 4.2.3 Cheekpieces and premaxillary remodeling .................................................................................... 54 4.3 Methods ................................................................................................................................................ 58 4.4 Results ................................................................................................................................................... 59 4.5 Discussion.............................................................................................................................................. 60 4.6 Conclusion ............................................................................................................................................. 62 Chapter 5: Horseback riding, asymmetry, and anthropogenic changes to the equine skull: e ide e fo ou ted idi g i Mo golia s late Bronze Age ..................................................................... 63 5.1. Introduction ......................................................................................................................................... 63 5.1.1 Identifying Mongolian riding: an ethnoarchaeological approach.................................................. 64 5.2. Materials and methods ........................................................................................................................ 69 5.2.1 Contemporary Mongolian horses .................................................................................................. 70 5.2.2 Iron and Middle Age Mongolian horses......................................................................................... 70 5.2.3 Contemporary American and Przewalski horses ........................................................................... 70 5.2.4 Bronze Age Mongolian horses ....................................................................................................... 70 5.2.5 Data collection protocols ............................................................................................................... 71 5.2.5.1 Nasal and premaxillary remodeling ........................................................................................ 71 viii 5.2.5.2 Dental pathologies linked to bitting ....................................................................................... 71 5.3 Results ................................................................................................................................................... 74 5.3.1 Nasal remodeling ........................................................................................................................... 74 5.3.2 Premaxillary remodeling ................................................................................................................ 74 5.3.3 Oral and bitting damage ................................................................................................................ 75 5.4. Discussion............................................................................................................................................. 78 5.4.1 Horse monument at Arvaikheer .................................................................................................... 79 5.4.2 Assessing late Bronze Age horse use ............................................................................................. 80 5.5 Conclusions ........................................................................................................................................... 81 Chapter 6: A Bayesian chronology for early domestic horse use in the Eurasian Steppe .......................... 83 6.1 Introduction .......................................................................................................................................... 83 6.1.1 DSK horse use in chronological context ......................................................................................... 84 6.2 Materials and methods ......................................................................................................................... 85 6.2.1 Aggregating published 14C dates .................................................................................................... 85 6.2.2 New 14C analysis ............................................................................................................................. 86 6.2.3 Modeling DSK horse use ................................................................................................................ 86 6.2.4 Modeling deer stone and khirigsuur construction ........................................................................ 86 6.3 Results ................................................................................................................................................... 87 6.4 Discussion.............................................................................................................................................. 91 Chapter 7: Conclusions ............................................................................................................................... 94 7.1 Overview ............................................................................................................................................... 94 7.2 Implications for the origins of horse riding and nomadic societies in East Asia ................................... 95 7.2.1 Connectivity, communication, and trade ...................................................................................... 97 7.2.2 Social legacy ................................................................................................................................... 97 7.3 Conclusion ............................................................................................................................................. 99 Appendix I. Data used in dental estimates of age and sex, along with provenience data for analyzed DSK sample (Ch.2). ..................................................................................................................... 105 Appendix II. Age and sex data utilized in this study (Ch. 2) ...................................................................... 105 Appendix III. Age and pathology measurements/scores for horses included in the study (Ch. 3)........... 107 Appendix IV. Dental and cranial osteological data for horses used in this study (Ch. 5) ......................... 105 Appendix V. Radiocarbon dates from deer stone and khirigsuur archaeological sites (Ch.6).................. 110 References ................................................................................................................................................ 115 ix LIST OF FIGURES Figure 2.1. Khirigsuur and deer stone sites included in the study (filled dots), as they relate to modern political boundaries and other locations mentioned in the text. ................................................. 12 Figure 2.2. A. Plan view schematic of an idealized khirigsuur monument, based on site maps from localities studied in this analysis.. ...................................................................................................... 14 Figure 2.2B. Plan view schematic of a model deer stone site, based on site maps from localities studied in this analysis. Diagram shows the burial of horse heads on all sides of the stone.. ................... 15 Figure 2.3. Illustration depicting horse head mounds at a deer stone (background). Stones are usually mounded 20-50 cm above the original ground surface, and buried heads usually face east/southeast (Allard et al., 2007). Illustration by Barbara Morrison....................................................... 16 Figure 2.4. Schematic showing procedure for rotating deer stone sites to share a common axis, and mapping relative position between horse burial mound and nearest deer stone. Daagan Del map by J. Bayarsaikhan and T. Tuvshinjargal.............................................................................................. 20 Figure 2.5. Mortality profile histogram for DSK sample, showing high frequency (proportion) of young horse remains, and second peak of adult animals between 6 and 15 years. .................................. 22 Figure 2.6. A. Schematic representation of demographic patterns at seven khirigsuur sites (Urt Bulagyn 1, Urt Bulagyn 2, Ushigiin Ovor, On Khad, Nukhtiin Am, AD40, Zeerdegchingiin Khoshuu). .................................................................................................................................................... 24 Figure 2.7. A. Schematic representation of nasal grooving at three khirigsuur sites (On Khad, Zeerdegchingiin Khoshuu, and Urt Bulagyn 1)............................................................................................ 25 Figure 2.8. Idealized mortality profiles for a managed pastoral horse herd (solid line, after Levine, 1999:31), and a transport assemblage (dotted line, after Levine, 1999:30), as compared to DSK assemblage. ..................................................................................................................................... 26 Figure 2.9. Mortality profile from Arzhan-2, from Benecke (2007). All horses are male, and classified according to median estimated age. Diagram shows emphasis on adult male horses for transport burials. ........................................................................................................................................ 27 Figure 3.1. Depictions of small horses alongside weapons such as daggers, bows and quivers (left–centre) on deer stones in Mongolia; also depicted are chariots (second from right) and x ha iot ei hooks o o -shaped o je ts fa ight efe e ed i the te t; odified f o Volkov (2002). ............................................................................................................................................. 33 Figure 3.2. A) Nuchal ossification on a ridden horse (left); B) ossification scores for horses used fo idi g, t a tio o d i i g , as o pa ed to u o ked E. przewalskii from European zoos (right); modified from Bendrey (2008). ...................................................................................................... 35 Figure 3.3 Medial (A) and lateral (B) groove formation on the nasal process of the premaxilla i isi e o e of a idde ho se, U“ Ge e al Joh J. Pe shi g s a ho se Kid o left , a d the same region on a feral Chincoteague pony (right); specimens from the Smithsonian National Museum of Natural History. ....................................................................................................................... 35 Figure 3.4. Medial groove depth, measured perpendicular to the intersection of groove walls and the dorsal surface of the premaxilla (incisive bone), shown here on an archaeological specimen. .................................................................................................................................................... 37 Figure 3.5. Nuchal ossification/musculoskeletal stress marker (MSM) scores (1–6) for museum sample specimens from ridden horses (Bendrey 2008 and data from this study); driven horses (Bendrey 2008); E. przewalskii f o p o a le zoo p o e a e Be d e ; P z. ; ho ses of k o zoo p o e a e P z. , this stud a d fe al a i als this stud ......................................... 37 Figure 3.6. Nuchal ossification/musculoskeletal stress marker (MSM) scores for ridden horses Be d e , d i e ho ses Be d e , P ze alski s ho ses f o Eu opea zoos Be d e 2008), and feral horses analysed in this study............................................................................................ 41 Figure 3.7. A) Medial nasal groove depth across ridden, feral and zoo samples (top); B) the same data normalised to premaxilla (incisive bone) width (bottom left); C) corrected for age using OLS residuals (bottom right). ............................................................................................................................. 42 Figure 3.8. Plot of medial groove depth and nuchal ossification score showing co-occurrence of high values in ridden specimens. ................................................................................................................ 43 Figure 3.9. A) Lateral groove depth by group (left); B) after normalisation to bone width and correcting for age (right). ............................................................................................................................ 43 Figure 3.10. Schematic diagram of horse cranium, indicating position of simple rope halter relative to nasal remodelling, and the path of the infraorbital nerve (arrow). .......................................... 44 Figure 3.11. A) Nuchal ossification score for DSK sample (top), as compared with known groups, d i i g a d idi g data f o Be d e ; B o alised a d age-corrected medial groove xi depth for DSK sample (lower left), as compared to known groups; C) normalised and agecorrected lateral groove depth for DSK and comparative horses (lower right). ........................................ 46 Figure 3.12. Lateral vs medial groove depth across groups, showing co-occurrence of high values in ridden and DSK samples.......................................................................................................................... 47 Figure 4.1. Khirigsuur and deer stone sites included in the study (filled dots), as they relate to contemporary political boundaries, the Minusinsk Basin in South Siberia, and Anyang, China (Shang Dynasty capital)............................................................................................................................... 51 Figure 4.2. Interior and exterior views of antler tine bridle cheekpieces from the site of B-007 in Egiin Gol Valley, Mongolia, dated to circa 940–800 BC (Honeychurch 2015:129). Drawing by Dr. Joshua Wright, reprinted with permission.. ............................................................................................... 52 Figure 4.3. Left: adult horse with facial bones badly deformed around undersized halter, found in Wyoming. Photo courtesy of Dr. Danny Walker. Center: nasal deformation in a contemporary Mongolian riding horse. Right: diagram showing direct connection between reins and noseband in Mongolian bridles. .................................................................................................................................. 53 Figure 4.4. Top: nasal depressions on working horses in the Altai of Xinjiang, China.Bottom: Altai horses in traction work, showing long reins looped through a body harness and attached to bridle headstall. Photos courtesy of Nils Larsen, Altai Skis. ....................................................................... 55 Figure 4.5. Diagram showing remodeling to the medial (A) and lateral (B) aspects of the equine premaxilla. Illustration by Rebecca Tuccillo................................................................................................ 56 Figure 4.6. Left: equipment of horses analyzed for premaxillary remodeling. A) simple loose ring snaffle, B) Turkic-era snaffle bit with S-shaped iron cheekpiece, C) Pazyryk snaffle with wooden cheekpiece, and D) Weymouth bridle under rein pressure. Right: US Cavalry curb bit similar to equipment used on Kidron s We outh idle, ith la k li e i di ati g the path of the infraorbital nerve in area of lateral remodeling. ........................................................................................ 56 Figure 4.7. Top: lateral vs. medial remodeling depth across a sample of wild extant and fossil equids, feral domestic horses, captive E. przewalskii, and ridden horses with documented equipment (A simple loose ring snaffle, B and C- archaeological snaffle with rigid cheekpiece, DWeymouth or double bridle). Bottom: lateral vs. medial remodeling depth for DSK horses (black) as compared to a sample of wild extant and fossil equids, feral domestic horses, captive E. przewalskii, and ridden horses with documented equipment (A- simple loose ring snaffle, B and C- archaeological snaffle with rigid cheekpiece, D- Weymouth or double bridle). .................................... 57 xii Figure 4.8. 3D model showing facial deformation from noseband use in a specimen from the site of Khushuutiin Gol (left, center) radiocarbon dated to 2910 +/- 40 14C BP (1224–980 cal yrs BC), and possible deformation in a young horse from Tsatstain Khushuu (right, 2920 +/- 40 14C BP). ........... 60 Figure 4.9. Young foal haltered and tied to a rope line with other foals during summer milking season, Bayankhongor province, Mongolia. ............................................................................................... 61 Figure 5.1. A group of Mongolian riders watch a horse race in Khuvsgul province, northern Mongolia. Image shows the ubiquity of the left-handed riding posture.................................................... 64 Figure 5.2. Mongolian herder riding left-handed, leaning to one side and stabilizing himself with the reins, with visible pressure the left nasal area. Herder using lasso pole visible in background. Photo by Orsoo Bayarsaikhan photography. .............................................................................................. 66 Figure 5.3. Asymmetric lateral remodelling to the premaxilla caused by remodeling of the bone in the area of the infraorbital nerve, shown on an archaeological specimen from Mongolia. .................. 67 Figure 5.4. Statue depicting a warrior from the Great Mongol Empire, 13th-14th centuries CE, outside the Parliament building in the capital city of Ulaanbaatar. ........................................................... 68 Figure 5.5. Petroglyphs from Tsagaan Gol in western Mongolia, showing driver holding two sets of reins, and reins running through a terret affixed to the pole (right). Photographs: Gary Tepfer. Copyright: Mongolian Altai Inventory Collection, University of Oregon. Reprinted with permission................................................................................................................................................... 69 Figure 5.6. The nasal bones of a horse from Uvurkhangai province in central Mongolia, showing p o ou ed tapho o i eathe i g to the a i al s left side i the a ea of asal defo ation. ............ 72 Figure 5.7. Asymmetric deformation to the nasal bones on a mummified horse dating to the Middle Ages from Ulaan-Uneet (left), and similar feature on a late Bronze Age horse from the site of Khushuutiin Gol in northern Mongolia (right). ................................................................................ 72 Figure 5.8A (top), showing measured left vs. right maximum premaxilla groove depths for feral American horses (n = 6), Przewalski horses (n = 7), contemporary American horses (n =11), contemporary Mongolian horses (n=13), post-Bronze Age archaeological horses (n = 7), and those from deer stones and khirigsuurs (DSK, n = 12). B (bottom), showing left minus right maximum premaxilla groove depths for feral American horses, Przewalski horses, contemporary American horses, contemporary Mongolian horses, post-Bronze Age archaeological horses, and those from deer stones and khirigsuurs (DSK). .......................................................................................... 73 xiii Figure 5.9A (top), showing concave wear to the upper P2 occlusal surface and flat beveling of the lower P2 in a Pazyryk horse from western Mongolia. B (center), flat beveling of both lower and upper premolars in a Xiongnu period horse from western Mongolia. C (bottom), bone formation on the left mandibular exterior on a horse dating to the Turkic period, likely caused by a bit. ............................................................................................................................................................ 77 Figure 5.10. Identical enamel chips on the anterior surface of the lower second premolars of a horse from the site of Zeerdegchingiin Khoshuu in northern Mongolia, which may have been caused by a hard bar bit.............................................................................................................................. 77 Figure 5.11. Racehorse skulls at Arvaikheer displaying premaxillary remodeling (top), and nasal thinning (bottom). ...................................................................................................................................... 80 Figure 6.1. Posterior calibrated probability ranges for 14C dates from horse remains at deer stones a d khi igsuu s. P io dist i utio i di ated i light g a . Dist i utio la eled D“K Ho se ep ese ts the output of O Cal s “u fu tio , a d su a izes the ge e al h o ologi al spread of the data. ...................................................................................................................................... 88 Figure 6.2. Modeled start and end dates for DSK horses, Khirigsuurs, and deer stones. Dashed line indicates median modeled start date for DSK horse ritual, falling within the 1-sigma range for deer stones but outside the modeled probability distribution for khirigsuurs. ................................... 89 Figure 6.3. Spatial distribution of DSK horse radiocarbon dates with available geographic provenience. For each date, the diameter of each circle corresponds to the percentage of the date s poste io p o a ilit dist i utio hi h falls ithi the ti e-slice................................................. 90 Figure 6.4. Modeled cultural phase start dates, as compared to large-scale climate data from Wang et al. 2011 (yellow), and important regional events in horse use. ................................................... 93 xiv LIST OF TABLES Table 2.1. Number of sample DSK horse specimens identified in each age category. ............................... 21 Table 2.2. Nasal groove index scores for analyzed horses from deer stones and khirigsuurs, along with demographic estimates and spatial provenience. .............................................................................. 22 Table 4.1. Osteological features of the skull and their potential significance for equine harness equipment................................................................................................................................................... 59 Table 5.1. Samples used in this study, along with number of specimens analyzed for cranial deformations and oral bitting damage. ...................................................................................................... 71 Table 5.2 Possible Bit-related oral damage among adult horses from post-Bronze Age archaeological contexts. ............................................................................................................................. 76 Ta le . . P o a ilit that t left olu p e edes t top o usi g OXCAL s O de fu tio . ........... 89 Table 6.2. Radiocarbon dates from Ulaanzuukh/Tevsh/Shorgooljin Bulsh features containing horse remains. ............................................................................................................................................ 92 xv CHAPTER 1: INTRODUCTION Domestication of the horse (E. caballus) prompted dramatic changes to human subsistence, interaction, and social organization. In locations as diverse as the plains of North America, the Pampas of South America, and the bushlands of Africa, the introduction of mounted horse transport prompted major changes to the scale and structure of social interaction, leading to new forms of hunting, warfare, and wealth (Mitchell 2015) and facilitating the long-distance travel of people and objects (Anthony et al. 1991). In the steppes of Central Asia, horses became the economic center of nomadic pastoral life. Horses are a valuable source of meat, dairy, and secondary products; in addition, as transport horses are used by nomads to control and move animals over long distances. And of course, beyond their pastoral significance, horse cavalry provided the military foundation for the cyclical emergence of pan-Eurasian nomadic empires (Rogers 2012). For these reasons, the development of horseback riding is often linked by scholars with the first emergence of nomadic life in eastern Eurasia (e.g. Beardsley 1953). Nonetheless, the question of when horseback riding was first developed, and its relationship to the first highly mobile herding societies of Central Asia remains unclear. Competing ideas for when and how horses were first used for transport– whether they were hitched to vehicles and chariots, or in true mounted horseback riding – complicate the issue further, and make it difficult to align ancient horse use with other environmental, technological, or social factors implicated in the emergence of nomadic pastoralism. In this dissertation, I combine new and established criteria for assessing ancient horse use with archaeofaunal data, using this information to shed light on the emergence of mobile herding life in Mongolia and the eastern Eurasian steppes. 1.1 DEFINING NOMADIC PASTORALISM The term pastoralism generally refers to a form of subsistence involving economic dependence on domestic herd animals. This category subsumes people engaged in a broad range of subsistence practices, varying widely in terms of both the degree of their dependence on animals and their residential mobility (Chang and Koster 1986:98-99). Migratory pastoralism, in its many forms, is practiced across the globe – including areas of Africa, the Near East, the Arabian peninsula, Scandinavia, Russia, and the Central Asian steppes (Khazanov 1984:19). Herding life in contemporary Mongolia is characterized by a highly mobile type of pastoralism, with residence in ephemeral felt structures (known as gers), regularized seasonal movements of up to several hundred kilometers, and a central reliance on horses as both transport and livestock animals (Bold 2012:130; Honeychurch 2015:92-94). Horses provide an important source of milk during the summer months and a source of meat during the fall and winter (Bold 2012). As transport, horses are used by herders to manage animals and move them over long seasonal migrations. Herders on horseback can move 2-3 times as far per day as those moving on foot alone, enabling more animals to be tended over larger pastures (Anthony et al. 1991). This manuscript focuses on understanding this particular kind of horse-based nomadic pastoralism –with the key components of fixed seasonal migrations, a high degree of residential mobility, and an economic strategy centered on the pasturing of livestock. 1 1.2 ORIGINS OF HORSE TRANSPORT IN CENTRAL ASIA The practice of managing horses as livestock has great antiquity in western Central Asia, but the story of when horses were first ridden is much less clear. The oldest reliable evidence for horse domestication can be found at sites of the Botai culture in the steppes of Kazakhstan and southern Siberia, dating to ca. 3500 BCE, where both dietary assemblages and ritual burials are dominated by horse remains (Olsen 2006). Although many of these horses were probably hunted or accumulated through mass harvesting, Botai culture sites have also produced corrals, ritual burials, and residues suggestive of the equine meat and dairy products (Olsen 2006; Outram et al. 2009). Botai s ho se he de s li ed i pe a e t settle e ts, a d p o a l practiced a sedentary way of life (Olsen 2006:107). Dental wear linked with the use of bridle bits is present on some domestic horse specimens at Botai (Outram et al. 2009). However, because similar equipment – a bridle and bit – would have been used to control both ridden mounts and those hitched to a vehicle (Dietz 2003:191) extant evidence cannot easily distinguish whether these animals were used for riding or pulling carts. By ca. 3100 BCE, other incipient pastoral economies – known as Yamnaya and Afanaseivo – had sprouted in southwestern Siberia and northern Kazakhstan. The bones of cattle, sheep, and horse are found inside Yamnaya and Afanasievo burials, while the discovery of wheeled carts points to a slightly more mobile lifestyle (Frachetti 2008:44). These groups probably still relied heavily on resources from hunting, fishing, and gathering activities for a large portion of their diet (Frachetti 2008:20). Just as in Botai, it is unclear how (or even if) these groups used horses for transport. Some influential works have raised the idea that horses were ridden since their initial domestication in the 4th millennium BCE (e.g. Anthony 2007). Others argue compellingly that at an earlier stage of domestication, chariots may have been a much more reliable form of transport than riding horseback (Dietz 2003:190) – which was done bareback, with strange posture, the use of a nose ring, and seems to have been undertaken for occasional athletic displays (Drews 2004:31-64). Yamnaya burials sometimes contained horses and wagons (Baumer 2012:96), but may have even preferred oxen to horses as draft animals (Drews 2004:29-30). Specialized pastoral groups – sometimes referred to under the broad umbrella of Andronovo – spread across a large swath of interior Central Asia by the end of the 3rd millennium BCE (Baumer 2012:141-151). These groups were marked by an increased dependence on livestock (specialized pastoralism) and a reduction in wild animal consumption (Frachetti 2008:47). The earliest, unmistakable evidence for horse transport also emerged during this time, in the form of chariot burials of the Sintashta culture (ca. 2100-1800 BCE). These features contain horses paired with chariots and bridle equipment (Outram et al. 2011:119), but provide little indication of mounted riding. Sintashta and other middle Bronze Age pastoral cultures practiced seasonal migration, but either lived in permanent fortified structures, or permanent but unfortified seasonal settlements (Baumer 2012141-151). At least some of these groups appear to have practiced altitudinal transhumance, moving to higher mountain pastures in summer months with a dietary focus on cattle, and to a lesser extent, sheep and goat (Frachetti 2008:56,131). Chariot petroglyphs are abundant in some areas associated with middle Bronze Age herding settlements (Frachetti 2008:139-140). Chariot burials notwithstanding, horses made a modest contribution to subsistence during this interval, and continued to leave a decreasingly visible signature in the archaeological record across the 2nd millennia BCE (Frachetti 2008:160; Outram et al. 2011:126). 2 One interpretation of this pattern is that these pastoralists lacked horse-riding technology – instead relying on horse-drawn chariots and ox carts for transportation need, and in geographically-limited seasonal migrations (Khazanov 1984:92-93). Due to the low frequency of dated or carefully analyzed faunal material from Early or Middle Bronze Age sites in the region, the chronology of early livestock use in Mongolia is less well characterized. However, evidence suggests that beginning in the 3rd millennium BCE, early pastoral groups also occupied some regions of western and central Mongolia. Dated Afanasievo sites have been found in some mountainous regions of the Altai and Khangai mountains, along with wooden vehicle parts, and faunal remains of sheep (Eregzen 2016; Kovalev and Erdenebaatar 2010:150-153; Houle 2010:4-5). A large body of horse chariot petroglyphs is also found on Mongolian rock art panels, variously attributed to the 3rd through the 1st millennium BCE (Erdene-Ochir and Khodyakov 2016: 23-30). Curiously, these images are spatially concentrated in the mountains of western and central Mongolia, reaching a stark easternmost boundary in eastern central Mongolia (Honeychurch 2015:193). Widespread evidence for mounted riding first appeared at the beginning of the first millennium BCE, when equestrian peoples are mentioned in historical records, and riding tack found at archaeological sites across much of Central Asia (Argent 2011:31; Honeychurch et al. 2009). Archaeological evidence links this period with the appearance of highly mobile herding societies in Mongolia and the eastern Eurasian steppes. In particular, domestic campsites and associated faunal assemblages in northern and central Mongolia suggest that by the end of the second millennium BCE, people had begun living in portable dwellings and consumed a diet heavy in sheep and goat meat – similar to the lifestyle of many nomads in contemporary Mongolia (Clark 2014; Houle 2010:180-1). In summary, the Bronze Age saw the development of specialized pastoral societies in western Central Asia, and their subsequent spread across the continental interior between the 3rd and 1st millennium BCE. Key changes in domestic animal use also occurred during this interval–the first domestication of the horse as a herd animal, a general increase in the reliance of many cultures on domestic fauna, and a decrease in the permanence of residential structures. The abundance of chariot petroglyphs and the occurrence of chariot burials in early pastoral cultures imply that in early pastoral groups, horses were commonly used to pull light chariots, rather than in mounted riding. However, it remains difficult to say with precision when horses were first ridden, or what role equine transport played in these pastoral transformations. 1.3 EXPLANATIONS FOR THE DEVELOPMENT OF NOMADIC PASTORALISM Intertwined with the chronology of horse riding and chariot use is the question of why mobile pastoralist lifeways developed in eastern Eurasia. Researchers have proposed a wide variety of potential causes, including social, economic, and technological, and environmental processes, to explain the formation of horse-riding, nomadic societies. 1.3.1 Sedentary agricultural societies Classic formulations of the origins of pastoral societies often attribute a causative role to sedentary agricultural states (Houle 2010:9; Rogers 2012:215-216). The spread of nomadic pastoralism in Eurasia was generally coeval with the formation of the first cities, leading some to 3 suggest that urban development itself led to horse transport (Sheratt 2003). Some such models argue that pastoral nomads fo ed f o a sloughi g off of e ess populatio f o sede ta agricultural groups (Lees and Bates 1974), or the domination of regional exchange systems by sedentary polities (Chang and Koster 1986:105; Cavalli Sforza 1996). In his classic work, Lattimore (1940:58-61) argued that mobile pastoralism in Northeast Asia developed in peripheral groups along the steppe frontier of northern China, as a result of increased agricultural specialization in Chinese polities during the 3rd and 4th centuries BCE. In recent years, however, several detailed regional studies indicate that across most of Central Asia hunting and gathering directly preceded the first herding groups, with little influence from sedentary states (e.g. Clark 2014:26; Frachetti 2008:20-21; Janz 2012:185; Wright 2006:285). In much of the continental interior, pastoral exploitation of domestic livestock appears to predate the local evidence for agriculture, suggesting a direct transition from hunting and gathering to pastoralism (Frachetti 2008:20). Thus, although agricultural societies may have played a key role in the initial domestication of some animal species in other parts of Eurasia, alternative frameworks may be necessary to characterize the development of mobile pastoralism in the eastern steppes. 1.3.2 Climate deterioration or resource scarcity Some influential models for nomadic origins prioritize the negative pressures of climate deterioration or resource scarcity (Kradin 2003:75). For example, Khazanov (1984:93) recognized that semi-sedentary herders had occupied many areas of Central Asia before the first millennium BCE, but that these societies differed in key ways (particularly in terms of their mobility) from horse-riding nomads. He linked the development of horse riding and highly mobile nomadic societies in eastern Eurasia with a prolonged period of drought and climate deterioration – which would have prompted herders to seek new ways to subsist in a more challenging environment. Crisis-oriented explanations for nomadic societies su h as Khazo o s reflect a recognition of the flexible and effective solutions provided by horse husbandry to the complex environmental obstacles of the Mongolian steppe. The short Mongolian summer features mild temperatures, and is wet enough to sustain relatively productive grasslands (Goulden et al. 2011:91). However, due to the seasonality of precipitation, plant cover regenerates slowly, and is particularly susceptible to damage from grazing. The Himalayas block monsoons as they traverse the continental interior, causing westerly winds to lose much of their moisture as they meet the high Altai mountain range. This leads to an arid climate, with long, cold, and dry winters. Despite the low precipitation, melting and immediate refreezing of snow cover can create a persistent crust over buried grasses which cannot be penetrated by sheep and cattle, causing them to starve during the winter. Because horses are able to access these buried pastures by using their hooves to dig, the inclusion of horses in a pastoral herd can be enough to sustain a mixed group through the winter season (Anthony and Brown 2011). By increasing mobility, horse transport might have reduced the risk of overgrazing seasonally rain-dependent pastures (Goulden et al. 2011). Both the incorporation of horses into domestic livestock herds, as well as the innovation of horseback riding would have improved the viability of mobile herding in Mongolia, perhaps in response to negative climate or environmental pressures. 4 1.3.3 Human behavioral ecology and niche construction Although climate deterioration is a compelling candidate for explaining nomadic pastoral origins in Mongolia, recent research has pointed to a link between the expansion and success of nomadic polities in the eastern steppes, and periods of climate amelioration or increased precipitation (Houle in preparation; Pederson et al. 2014; Putnam et al. 2016). This pattern hints at a fundamentally different relationship between mobile pastoralism and climate which is difficult to reconcile with drought-based explanatory frameworks for nomadic origins in eastern Eurasia. An alternative theoretical perspective that may help understand this pattern comes from human behavioral ecology (HBE) and niche construction theory (NCT). Considering the context of plant and animal domestication events around the globe, Smith (2012) and Zeder (2016) suggest that rather than an adaptation to environmental challenges, the process of domestication was a human enhancement of our own ecological niche. Because of the need for cooperation, intergenerational transmission of ecological knowledge, and sustained genetic changes to new generations of domestic organisms, NCT predicts that new niche constructing behaviors should emerge in conditions of stable, abundant resources, and relaxed selective pressures (Zeder 2016:336). As grasslands improved and expanded in the dry steppes of Mongolia, the new resources they offered would likely have remained largely inaccessible to those without the means of riding horseback. Even while pastures expanded in formerly unproductive regions, these areas would still have been characterized by extreme seasonality, with a high danger of overgrazing. Horseback riding gave herders the necessary tools to circumvent these issues– providing increased mobility, the ability to maintain larger herds with higher yield, and easier control of resources and territory over large distances. By altering the ecological constraints of the dry eastern Steppe in these ways, the innovation (or local adoption) of horseback riding can be understood as a niche-constructing behavior. From this perspective, the initial development of equestrian nomadic groups in Mongolia and eastern Eurasia might have been stimulated by experimentation under a favorable climate regime, as herders sought to capitalize on the new resources proffered by grassland expansion. 1.4 Late Bronze Age mobile pastoralism The clearest evidence for highly mobile pastoral herding in the eastern Steppe comes from the Deer Stone-Khirigsuur (DSK) complex, an archaeological culture named for large decorated sta di g sto es dee sto es a d u ial ou ds khirigsuurs) constructed across Mongolia, southern Tuva, eastern Kazakhstan, and some areas of northern China during the late second and early first millennium BCE. Analysis of settlement patterns indicates that DSK people practiced residential mobility, while faunal remains reveal a diet of sheep, cattle, and horse (Broderick et al. 2014; Houle 2010). A defining cultural trait of the DSK complex is the presence of small stone burial mounds found around the perimeter of deer stones and khirigsuurs – typically containing a single horse head, two or four hooves, and the atlas, axis, and other neck vertebrae (Fitzhugh 2009). At some sites, the number of these features can number into the hundreds or thousands (Allard and Erdenebaatar 2005). 5 Though the widespread practice of equine burials at DSK sites implies an important role for domestic horses, very few artifacts are recovered from either kind of site with which to directly evaluate how horses were used in DSK culture. Khiriguurs typically contain human skeletal remains, but no burial goods (Frohlich et al. 2009:108), while deer excavations at deer stones rarely yield cultural material beyond stelae and faunal remains from stone satellite features (Fitzhugh 2009:185). In particular, no bridle remains or tack have been recovered from DSK archaeological sites, leaving the question of how DSK horses were used for transport open to debate (Honeychurch et al. 2009). Consequently, the relationship between DSK culture, horseback riding, and nomadic pastoralism cannot be easily assessed. Chronological ambiguity further impairs evaluation of the role of climate and environmental processes in DSK social developments. Paleoclimate research has produced an increasingly highresolution body of ancient environmental data in recent years, suggesting the end of a prolonged dry period across much of the region circa 1200 BCE (e.g. Fukumoto et al. 2012; Propokenko et al. 2007; Wang et al 2011). This finding is consistent with arguments by Houle (2010:185-6) that DSK herders enjoyed relatively high population densities, and did not suffer from recognizable ecological stress. However, estimates for the onset of deer stone and khirigsuur construction vary by several centuries, between ca. 1500 BCE (Honeychurch 2015) and 1200 BCE (Fitzhugh 2009). Consequently, evaluating the role of ecological stressors or climate amelioration in early horse herding and riding in Mongolia requires detailed chronological study of DSK horse remains, as well a means of identifying horse transport practices using only archaeofaunal materials. 1.5 RESEARCH DESIGN Ritual burials at DSK archaeological sites provide a large, if variable, body of equine skeletal material useful for investigating horse use in late Bronze Age Mongolia. Within each satellite mound, horse remains were typically interred at a relatively shallow depth – around a half meter from the contemporary ground surface. Consequently, preservation of recovered specimens is occasionally poor. Many horse crania were fractured or partially crushed at the time of burial, due to the weight of the overlying stones used in mound construction. At some localities, remains have been further damaged from rodent disturbance, livestock trampling, or exposure to the elements. Although some mounds thus contain little more than scattered bone and tooth fragments, others produce complete or mostly intact specimens which can be used for detailed osteological study. These same horse specimens also provide organic material suitable for radiocarbon dating of DSK horse use. Using archaeological and archaeozoological data from DSK horse burials, this dissertation project evaluates the timing and nature of late Bronze Age horse use in Mongolia. The dissertation is organized into three complementary lines of archaeological inquiry, and five separate papers. Together, these manuscripts investigate how DSK people managed and used horses, producing a precision chronology for these processes to evaluate the ecological context of the emergence of nomadic pastoralism in Mongolia. Chapter 2: Horse Demography and Use in Bronze Age Mongolia (in press, Quaternary International) 6 By analyzing the dentition of archaeological horses excavated from deer stones and khirigsuurs, this manuscript evaluates the hypothesis that DSK people bred and managed horses. Based on ethnographic observations by previous researchers, the management of an equine livestock herd typically involves the culling of young males before they reach breeding age, as well as older horses beyond the age of reproductive viability (especially females). By comparing published age and sex estimates with a sample of newly analyzed horse specimens from DSK sites across Mongolia, this study reveals a consistent pattern of juvenile horse and elderly mare selection, consistent with expectations for the mortality profile of a managed domestic herd. Curiously, the sample also contained a meaningful p opo tio of p i e age adult ale ho ses which cannot easily be explained by herd management practices. Many natural and cultural selection processes could yield a high proportion of adult male animals, but these male horses were all concentrated in specific locations along the eastern periphery of monuments, often considered to have special ritual significance in DSK culture. Moreover, osteological features associated with transport activity support the idea that these adult male horses were used for transport. These results indicate that in addition to managing horses as a pastoral animal, transport horses had already assumed an important cultural role in Mongolian societies by the late Bronze Age. Chapter 3: Equine Cranial Morphology and the Archaeological Identification of Riding and Chariotry in the Mongolian Bronze Age (Antiquity 89(346): 854-871) This manuscript compares osteological changes to the equine skull between a sample of wild and domestic equids to evaluate their relationship to horse transport, bridling, and human activity. In previous veterinary studies, remodeling to the equine premaxilla has been associated with heavy training or chronic exertion. Based on comparisons between museum collections of wild and domestic horses, results of this study indicate that the depth of grooves to the interior margin of the premaxilla are more pronounced among horses used for riding or traction. A second type of groove to the lateral aspect of the premaxilla, associated with the infraorbital nerve, also appears more frequent and severe among ridden horses. Previous research linked new bone formation to the rear of the skull with the use of horses for riding. Our data suggest that captive wild animals may also develop this pathology from stress or chronic posture issues, whereas feral animals may develop new bone in this area at a far lower frequency. Investigating these osteological features in a sample of late Bronze Age horses from DSK sites provides compelling evidence that at least some of these animals were used for riding or chariotry. Chapter 4: Reconstructing Equine Bridles in Bronze Age Mongolia (in press, Journal of Ethnobiology) This paper develops expectations for the reconstruction of archaeological tack using equine cranial remains. As used he e, the te igid heekpie e efe s to a elo gated, i fle i le bridle component running perpendicular to the mouth of the horse, which functions to keep the mouthpiece in place and serves as a turning aid. Based on archaeological finds from other areas of central and eastern Asia, it can be inferred that late Bronze Age bridle equipment probably incorporated a noseband, rigid cheekpieces, and an organic bit. However, the late Bronze Age Mongolian archaeological record itself has yielded few bridle artifacts or other evidence for how early domestic horses might have been bridled. Researchers have connected metal bit use with several kinds of oral pathology and changes to equine dentition, including occlusal beveling, wear to the anterior margin of the lower second premolar, and bone formation to the diastema. 7 A noseband with direct rein attachment may also cause recognizable deformation to the bridge of the nose, and some contemporary horse specimens raise the possibility that lateral premaxilla deformation is exacerbated by exterior pressure or irritation from bridle cheekpieces. Using these features, osteological study of horse skulls can help to identify the use of particular bridle components in the archaeological record, even in the absence of other artifacts. Remodeling to the skulls of Bronze Age horses from DSK sites suggests that some of these animals were bridled or haltered with a noseband, and likely ridden with an organic bit. If future studies validate this link, the presence of severe remodeling to the premaxilla exterior observed on DSK horses may also indicate use of a rigid cheekpiece. This inferred bridle type, constructed of organic materials, might have played an important role in early nomadic horse control in Mongolia and the eastern Steppe. Chapter 5: Horseback Riding, Asymmetry, and Anthropogenic Changes to the Equine Skull: E ide e fo Mou ted ‘idi g i Mo golia s Late B o ze Age i e ie , O o Books: Proceedings of the 6th Animal Paleopathology Working Group, Budapest, Hungary) A central challenge in the study of early Mongolian horse transport is distinguishing horses used for riding from those used to pull carts or chariots. Because both types of transport use similar bridle equipment and involve chronic heavy exertion, the cranial features linked to horseback riding observable on archaeological samples may also characterize horses used to pull chariots. As a result, archaeozoological analyses are typically unable to provide reliable data on how a particular horse was used. Combining ethnographic observations and zoological data, this manuscript presents new evidence that the left-handed riding style used in contemporary Mongolian horseback riding may produce asymmetric effects to the equine skull, an observation which can help identify equestrian activity in the archaeological record. Consistent with the predicted effects of a cheekpiece and noseband under leftward rein tension, contemporary Mongolian horse skulls show evidence of deeper remodeling to the right premaxillary margin, and greater remodeling of the left side of the bridge of the nose. Similar patterns of asymmetry were also observed in a small sample of archaeological riding horses from Mongolian burials. In contrast, domestic horse skulls from American museum collections displayed generally symmetric patterns of deformation linked to riding and bridling. Finally, a sample of late Bronze Age skulls from deer stones and khirigsuurs exhibit deeper grooves on the right premaxilla, and a single well-preserved horse has pronounced remodeling of the left nasal margin similar to that observed in ridden horses. Future study will be necessary to assess whether pulling chariots could also have produced similarly asymmetric deformations, but these data provide preliminary evidence that DSK horses were used for mounted riding. Chapter 6: A Bayesian Chronology for Early Domestic Horse Use in the Eastern Steppe (in review, Proceedings of the National Academy of Sciences) Assessing the relationship between horses, key environmental and cultural processes, and the origin of nomadic societies in eastern Eurasia requires a precise chronology for DSK horse use. This paper uses Bayesian statistical modeling to produce a high-resolution model for domestic horse ritual from DSK archaeological sites. Results indicate that ritual horse burials spread rapidly across Mongolia by circa 1200 BCE, several centuries after the oldest khirigsuur burial sites, but concurrent with the construction of the first deer stones. Comparison of this modeled date with available paleoecological data indicates the expansion of horse use was associated with climate amelioration, rather than drought. The expansion of domestic horse use by DSK 8 people, perhaps prompted by the spread of mounted horseback riding and nomadic pastoralism, may have prompted the spread of horses into new areas of the continent such as Shang China. 1.6 SUMMARY While the innovation and spread of horse transport is often recognized as a catalyst in human social changes, the role of horseback riding in the emergence of nomadic herding societies is poorly characterized – in part due to the lack of direct evidence for human use of horses in the archaeological record. Circumstantial evidence points to a close link between the widespread adoption of horse transport and the emergence of nomadic societies in the Eastern Steppe during the late Bronze Age, but this premise has proven challenging to evaluate using extant archaeological data. Drawing from both new and established archaeozoological techniques for identifying human activity using equine skeletal remains, faunal remains from ritual burials at deer stones and khirigsuurs indicate that late Bronze Age people herded and bred horses and used them for transport. Chronological modeling suggests that this process took place in the context of climate amelioration, rather than drought or ecological crisis. These results align with the predictions of niche construction theory, which suggests that human modifications of ecological niches through domestication should occur in the context of resource abundance and stability. This framework may help explain the delayed emergence of sophisticated horseback riding and the timing of the initial spread of horses into other areas of the Eurasian continent. CHAPTER 2: HORSE DEMOGRAPHY AND USE IN BRONZE AGE MONGOLIA In press, Quaternary International (2016), http://dx.doi.org/10.1016/j.quaint.2015.09.085 William Taylor1 This paper presents new archaeozoological evidence for horse pastoralism and transport in Mongolia's Deer Stone-Khirigsuur (DSK) Complex (circa 1300-700 BCE). As both livestock and transport, the domestic horse fundamentally altered life in the dry steppe of eastern Eurasia. However, the timing and process of mobile pastoralism's adoption in Mongolia and Northeast Asia remains poorly understood. To evaluate previous suggestions of late Bronze Age horse herding in the DSK complex, I produced age and sex estimates for archaeological horse crania from DSK sites across Mongolia. This sample yielded a high proportion of juvenile animals and an elderly female specimen, consistent with the culling practices of contemporary equine pastoralists. However, the sample also contained a significant proportion of prime age adult male animals. This finding is seemingly inconsistent with the practical requirements of pastoral herd management, but comparable with other archaeological assemblages of ritually-sacrificed transport horses. Spatial comparison suggests that these adult males were buried in specific ritual contexts, along the eastern edge of stone mounds known as khirigsuurs, while osteological features of the premaxilla point to harnessing or heavy exertion. Together, these data provide compelling evidence that adult male DSK horses were used for chariotry or mounted riding. Results support 9 the interpretation of DSK people as early mobile pastoralists, and suggest an important role for horse transport in late Bronze Age social dynamics and the development of herding societies in Northeast Asia. Keywords: Pastoralism, Horse transport, Mongolia, Late Bronze Age, Deer stone, Khirigsuur 1 Department of Anthropology, University of New Mexico, MSC01-1040, Albuquerque, NM 87131, USA 2.1. INTRODUCTION This paper investigates the role of the horse in late Bronze Age Mongolia and its implications for the origins of pastoral nomadism in the eastern Steppe of Asia. As defined here, pastoralism refers to the tending of domestic animal herds (Chang and Koster 1986:99), while mobile or nomadic pastoralism refers to those types of herding which rely on coordinated movement, and lack permanent settlements (Salzman 2004:3-6). Horse transport will refer to the use of horses for mounted riding, as well as to pull chariots or other vehicles. In modern Northeast Asia, mobile pastoralism is characterized by reliance on the horse as both livestock and transport (Barfield 2011:109). During the last three millennia, nomadic people from the dry steppes of eastern Asia developed new forms of social organization, forming complex societies and empires that shaped much of modern Eurasia (Rogers 2012; Honeychurch2015). Although horses were domesticated by circa 3500 BCE in western areas of the Eurasian steppes (Olsen 2003; Outram et al., 2009), a systematic understanding of when and how horse-using pastoral societies first emerged in Northeast Asia remains elusive. Here, I present demographic and paleopathological analysis of 25 horse crania from late Bronze Age archaeological sites of Mongolia's Deer Stone Khirigsuur (DSK) cultural complex. Using data from dental eruption and wear patterns, I estimated the age and sex of all specimens, evaluating each for cranial pathologies and osteological features related to equine transport using a NextEngine3D scanner. Results provide support for DSK horse herding, with a high proportion of juveniles and senescent mares indicative of management and breeding. Unlike previous studies, this sample also contained a significant proportion of adult male animals. Spatial patterns suggest that these horses were buried in prominent ritual locations along the eastern axis of stone monuments, while osteological features of the skull provide preliminary evidence of their use in transport. Taken together, these data support characterizations of the DSK complex as an early mobile pastoralist society, utilizing the domestic horse for subsistence, ceremony, and transport. Results suggest a sophisticated knowledge of equine ecology, and raise new questions about the environmental and economic conditions under which early nomadic pastoral societies formed in Northeast Asia. 2.2 REGIONAL SETTING The political boundaries of contemporary Mongolia correspond largely with the ecological boundaries of the eastern Steppe (Figure 2.1), which extends across a vast plateau extending from Kazakhstan in the west to the mountains bordering Manchuria in the southeast (Barfield 2011; Goulden et al. 2011). The plateau's location inland of the Himalayas prevents rain-heavy 10 monsoon from reaching the interior, and causes westerly winds to lose much of their moisture as they meet the high Altai mountain range. The result is an arid climate, at both high latitude and high elevation. Winters are long, cold, and dry in Mongolia, although the short summer boasts mild temperatures and is comparatively wet (Goulden et al. 2011:91). Due to extreme seasonality of precipitation, Mongolian herders must move often to prevent overgrazing (Goulden et al. 2011:99), and the horse is the most important form of transport (Bold, 2012:9192). The horse is also an important livestock animal, providing, meat, dairy, leather, dung, and other important products. As a result, horses remain critically important to subsistence in the modern Mongolian Steppe (Bold 2012:130). The development of specialized, nomadic pastoral societies has often been linked to growth of agricultural societies (Lees and Bates 1974; Khazanov 1984), and framed as a response to resource scarcity, climate deterioration, and other political and economic consequences of sedentary state formation (Khazanov 1984:95; Chang and Koster 1986:105; Cribb 2004:12-15). In his classic work, Lattimore (1940:58-61) argued that mobile pastoralism in Northeast Asia developed in peripheral groups along the steppe frontier of northern China, as a result of increased agricultural specialization in Chinese polities during the 3rd and 4th centuries BCE. As access to the Mongolian archaeological record has improved, however, it has become clear that hunting and gathering directly preceded pastoralism as a subsistence strategy in many areas (e.g. Wright 2006:285; Janz 2012:185; Clark 2014:26). Some scholars have suggested that equestrian herding in Mongolia may date as far back as the late Bronze Age, to the late second millennium BCE's Deer Stone-Khirigsuur (DSK) complex (Houle 2009; Fitzhugh 2009a). This interval might have seen the onset of a comparatively wet and productive climate regime (Wang et al., 2011; Fukumoto et al., 2012:88), and steppe people might have controlled important trade routes across the continental interior at this time (Christian 2000). If DSK people were indeed mobile pastoralists, this context might warrant reevaluation of the chronology and causes of the formation of nomadic herding societies. Material remnants from the DSK period with which to evaluate horse use are scarce, consisting primarily of stone monuments (deer stones and khirigsuurs), and associated sacrificial animal deposits. Khirigsuur is the Mongolian term for large stone mounds dating to the late Bronze Age, which were built across a wide geographic area, from Baikal to the northern Gobi, starting in the late second millennium BCE. These monuments are typically encircled by a rectangular or circular stone fence (Figure 2.2A, Fitzhugh 2009a). Deer stones are anthropomorphic standing stones frequently associated with khirigsuurs. These stelae are regularly decorated with elaborate deer carvings, from which their name derives, as well as weapons and other images interpreted as warrior motifs (Fitzhugh 2009a). Some deer stones depict bow-shaped objects which could be associated with chariotry (Fitzhugh and Bayarsaikhan 2011:178; Wu 2013:40). However, neither type of monument is associated with grave goods which definitively indicate horse transport (Frohlich et al. 2009). Archaeozoological materials thus hold the key to assessing DSK horse use and its potential role in early nomadic lifeways. 11 Figure 2.1. Khirigsuur and deer stone sites included in the study (filled dots), as they relate to modern political boundaries and other locations mentioned in the text. Horse sacrifice was a key focus of ceremonialism at both deer stones and khirigsuurs (Allard and Erdenebaatar 2005; Fitzhugh 2009a). The heads of sacrificed horses, which might have been oriented towards the rising sun (Allard et al. 2007), were buried in small stone mounds along with cervical vertebrae and hoof bones around the perimeter of many DSK ritual sites (Fitzhugh 2009a; Figs. 2.2 and 2.3). At khirigsuurs, the E/SE fence of the monument is a common location for these satellite horse mounds, with additional rows also found along the north, east, and southern perimeters at some larger monuments (Figure 2.2A). Deer stones themselves are typically oriented to the east or southeast, as indicated by the monument's face (Fitzhugh 2009:189). At isolated stones, a ring of horse burial mounds may encircle the monument (Figure 2.2B), while larger stelae clusters are surrounded by carefully or haphazardly arranged groups of mounds (see Figure 2.4 for an example). At both monument types, an exterior ring of stone circles often brackets the area of horse burials (Figure 2.2B). Equine sacrifice appears to have served a memorial purpose at both kinds of monument. Khirigsuur mounds often contain human burials (Littleton et al. 2012), although they may not have been solely mortuary in function (Wright 2014). Although deer stone sites are typically devoid of human remains, the unique set of carvings on each stone suggests that each may depict a particular individual (Fitzhugh 2009a:188). Scholars have noted that the structure and orientation of horse mounds themselves appears remarkably consistent across DSK sites (Allard et al. 2007; Fitzhugh, 2009a; Wright 2014). This consistency may reflect functional similarity; with horse sacrifice at khirigsuurs memorializing the physical burial of a fallen leader or relative (Fitzhugh 2009a:191), and deer stones serving as cenotaphs (Jacobson, 1993:144, 153-157; Fitzhugh and Bayarsaikhan 2011:182-3). However, while the horse mounds at some DSK monuments were purposefully arranged (perhaps in a single construction event), others are laid out in a more haphazard 12 fashion (Wright 2014: 153- 4). This scenario may suggest variation in when and why horses were sacrificed at deer stones and khirigsuurs. Previous archaeozoological research (Allard and Erdenebaatar 2005; Houle 2010; Broderick et al. 2014) indicates that DSK people managed domestic livestock, including sheep, cattle, and horses. Habitation and refuse sites are exceedingly rare in the late Bronze Age archaeological record, but one midden deposit contained evidence of horses slaughtered for meat, and a diet heavy in sheep/goat and cattle (Houle 2010:123, 127-129). Faunal evidence of livestock consumption is also known from DSK ritual features. For example, recent research by Broderick et al. (2014) identified highly calcined caprine and bovine remains in the stone circles found at many DSK monuments (see schematic in Figure 2.2A). Together, these finds provide compelling reason to conclude that DSK people bred and herded a variety of domestic livestock. Archaeological remains may be useful in identifying ancient equine herd management practices. On the steppes, pastoral horse herds mimic the family structure of wild horses. Natural herds typically consist of a single stallion, a harem of around six females, and a contingent of juveniles (Levine 1999:23). Among modern horse pastoralists in Central Asia, domestic herds include more mares (8 to 15 according to Bold 2012:145) and an approximately equal number of foals (Levine 1999:22). Ethnographic accounts suggest that young males are usually gelded or culled before reaching sexual maturity, when they will compete for mates and potentially disrupt herd hierarchy (Levine 1999; Olsen 2006b:87).In one study, contemporary pastoral horse herds consisted of up to one-third gelded males, the rest mares and foals (Levine 1999:22). Mares that have not produced young for several breeding seasons will usually be culled for meat, while healthy transport animals reach old ages of 15-20 years or more before death or slaughter (Levine 1999; Bold 2012:92). In groups that focus on mare's milk production, it may also be advantageous to cull extraneous foals at the age of 6 months, so as to increase winter herd survival and improve birth spacing for the coming spring (Olsen 2006b:88-89).Under ideal conditions, these processes should produce a mortality profile with a high frequency of subadult and elderly animals, and few prime-age mature individuals. Ethnographic pastoral practices may differ from late Bronze Age herd management strategies, but some archaeological evidence suggests that a similar scenario characterized DSK horse management. Based on epiphyseal fusion patterns, Houle (2010:127-129) estimated that all of the horse remains from a late Bronze Age midden in central Mongolia were between 2 and 3 years of age at death, and were probably culled for meat. A similar argument was made by Allard et al. (2007), based on a sample of 15 ritual horse inhumations from the large khirigsuur site of Urt Bulagyn in central Mongolia. Although excavated from a ritual context and a variety of mound locations, their sample contained mostly young horses, and several were buried adjacent to elderly mare specimens. This pattern is consistent with the sacrifice of young males and senescent females to maintain a breeding herd (Levine 1999), and has been interpreted as evidence that DSK horse ritual was closely linked to the practical reality of herd management (Allard et al. 2007). 13 Figure 2.2. A. Plan view schematic of an idealized khirigsuur monument, based on site maps from localities studied in this analysis. Map shows a central burial mound ringed with a rectangular fence, and surrounded by satellite burials of horse crania and stone circles with calcined caprine and bovine remains. The sides of the rectangle may appear irregular, and in other cases the fence might be circular (Frohlich et al. 2009). Although usually concentrated on the E/SE side of khirigsuurs, at larger monuments, horse mounds are also found along the N and S axes (shown as transparent grey circles). 14 Figure 2.2B. Plan view schematic of a model deer stone site, based on site maps from localities studied in this analysis. Diagram shows the burial of horse heads on all sides of the stone. At some larger deer stone sites, multiple stelae may be aligned in an approximate N-S arrangement, and a pavement of mounds may surround individual stelae or larger groupings (see Figure 2.4 for a useful example). 15 Figure 2.3. Illustration depicting horse head mounds at a deer stone (background). Stones are usually mounded 20-50 cm above the original ground surface, and buried heads usually face east/southeast (Allard et al., 2007). Illustration by Barbara Morrison. The predominance of juvenile and adult female horse remains in Allard et al.'s sample, however, differs considerably from assemblages linked to equine transport (e.g., Rudenko 1970; Levine 1999:30; Benecke 2007; Outram et al. 2012). In such contexts, mortality profiles show mostly prime-aged, adult male horses (see Figs. 2.8 and 2.9, this paper). Nonetheless, DSK researchers have occasionally noted the occurrence of adult male horse heads in prominent satellite mounds along the E/SE edge of DSK sites (Takahama et al. 2006; Allard et al. 2007). This raises the possibility of undocumented spatial patterns in equine demography, which may be instructive regarding late Bronze Age horse use in Mongolia. To clarify this issue and characterize DSK horse use, a synthesis of demographic and spatial analysis is necessary. 2.3. MATERIALS AND METHODS To systematically test the hypotheses of DSK horse herding and transport, I conducted a demographic study of 25 horse crania excavated from satellite mounds associated with deer stones and khirigsuurs. Specimens originated from 12 deer stone sites and five khirigsuurs at a total of 13 unique localities across central and western Mongolia (Figure 2.1, see Appendix I for site names). Most study sites are represented by only one or two excavated horse mounds. The 18 study specimens with associated radiocarbon dates span a wide temporal range (1337-769 cal. BCE, limits of 2-sigma confidence interval for 18 dates, see Fitzhugh and Bayarsaikhan, 2009: Appendix II). Subsequently, chronological or geographic differences are likely obscured, and sample generalizability must be tested with additional research rather than assumed. Nonetheless, this provides a useful starting point to investigate equine demography for the DSK complex. 16 Although horses are not highly sexually dimorphic, adult male horses develop large canine teeth at the age of 5 years. Most mares lack these teeth entirely, and those that develop them often have fewer and smaller canines (Evans et al. 2006; Olsen 2006b:86). I assigned a designation of male to specimens with four large canines, and female to adult specimens with sufficient preservation to identify that no canines were present. Specimens with incomplete preservation, or with estimated ages of less than five years were considered to be of indeterminate sex. For heuristic purposes, I placed specimens with at least one large canine in the category of possibly male . This protocol inevitably underestimates the absolute frequency of females (and elderly male horses that have lost teeth), and leaves juvenile animals unsexed. However, this conservative approach also reduces error in sex assignment, and enables a more robust exploration of spatial patterns in horse demography for those with sufficient preservation. To estimate age for sample horses, dental eruption and wear patterns were compared to schedules derived from modern horses. Subadult and juvenile horses younger than five years can be identified with some precision based on the regular eruption schedule of deciduous teeth, followed by permanent incisors, cheek teeth, and canines (Evans et al. 2006). Because dental eruption through bone precedes that of the gums, values derived from archaeological remains may produce a slight overestimate (Olsen 2006b:86). The age of adult horses can be approximated using a measurement of crown height, the distance between the occlusal surface and the intersection of the permanent tooth roots (Levine 1982). These measurements are likely to vary among individual horses, and are not useful beyond 20 years of age (Levine 1982). However, the efficacy of crown height in estimating age may be increased by measurement of multiple cheek teeth from the same individual (Enloe and Turner 2002). Finally, morphological attributes of the incisors are helpful in assigning age estimates to adult horses. Between five and 12 years, sequential changes to the cups of the upper and lower incisors provide a useful estimator of age (Evans et al. 2006). References used for this task are derived largely from contemporary domesticates, and their reliability is influenced by issues such as malocclusion, diet, genetics, and behavior (Allen 2008). Although the capacity for error is even greater after 12 years, the shape and orientation of the incisor grinding table may provide age estimates up to the age of 30 (Pasquini et al. 2003:246-253;Academy of Equine Dentistry 2013). It is improbable that these methods accurately reproduce the age of death for freeranging animals raised on rough steppe forage, particularly for older specimens. However, until a more appropriate system can be developed, this technique at least enables intra-sample comparisons and facilitates the assignment of archaeological specimens to approximate age categories. For specimens younger than five years, I estimated age using eruption schedules (Evans et al. 2006). For adult horse skulls, I measured crown height from loose cheek teeth if available (Levine 1982), and results were corroborated by incisor wear tables (Evans et al. 2006; Academy of Equine Dentistry 2013). Specific age estimates for animals older than 20 years, based on incisor morphology, are provided in Appendix I. However, due to the greater error associated with this method these animals are only reported as 18+ in mortality profile analysis. As different estimation techniques often yield slightly different results, the median value of the total estimated age range was used to sort specimens into age classes for comparison. Finally, I checked these provisional estimates against professional assessments by an equine veterinarian. 17 As a result, these data may differ slightly from values published in Taylor et al. (2015). All values are reported in Appendix I. After characterizing age and sex of the DSK sample, I compared it with known archaeological mortality profiles for both transport and pastoralist assemblages (Rudenko 1970; Levine 1999:30; Benecke 2007; Lepetz 2013). To assess spatial patterns in equine demography, I plotted estimated age, sex, and provenience in relation to an approximate and idealized monument layout. To do this, I combined site report maps of 25 horse burial mounds, recording the relative location of each horse burial with demographic data on a composite diagram. Although horse mounds are typically less than 4 m in diameter, khirigsuur mounds range widely in size from several meters to several hundred meters across (Frohlich et al. 2009:99). Nonetheless, khirigsuurs of different scales seem to express many of the same fundamental components (Wright 2014:148). As a result, the relative position of horse burial features may still be a comparable attribute between sites. To facilitate comparison of horse head mound position at khirigsuurs of different sizes, I scaled and translated original archaeological site maps so that the distance between the center of the khirigsuur mound and the first row of satellite horse burial features was equivalent for all features on the combined diagram. At deer stones, the structure of horse burial features was often difficult to distinguish from surface maps. In addition, major architectural variability at deer stone sites (ranging from single stones and associated horse mounds, to enormous stelae/mound complexes) prevented any consistent rescaling efforts. As a result, deer stone site plans were not rescaled in this fashion. Variability in site orientation also complicates the evaluation of demographic patterns in DSK horse placement. The alignment of DSK horse heads may correspond to the local direction of sunrise at the time of initial slaughter (Allard et al. 2007). If so, differences in horse head orientation and site layout could reflect factors such as latitude and season of sacrifice. Ethnographic study in contemporary Mongolia suggests that horse slaughter usually takes place in the late autumn, when it will improve winter herd survival (Allard et al. 2007). Similar practices may also have characterized Bronze Age horse sacrifice. Although horse head position was not considered in this study, the orientation of each DSK site was assessed based on longitudinal alignments of deer stone stelae, khirigsuur fences, or horse mound rows. With the possible exception of Ulaan Tolgoi, whose deer stones are aligned slightly SE to NW but face east (see Fitzhugh 2006:181), and several sites which face roughly due east (Nukhtiin Am KS, Ushigiin Ovor DS, and Ulaan Tolgoi KS) all study sites are apparently oriented to the southeast. This arrangement could be consistent with an autumn or winter construction (Olsen 2006:103; Allard et al. 2007). Whatever the cause, this inter-site variation makes it difficult to synthesize and compare the relative position of horse mounds from different DSK localities. To correct this issue, I rotated each map so that it shared a common axis. For example, if the longitudinal axis of a site ran southwest to northeast, such that the first row of satellite mounds faced southeast, the site's map was repositioned to run parallel with a perfect N-S alignment (Figure 2.4). The position of each satellite feature relative to the center of the khirigsuur or nearest deer stone was then mapped on a common plot (Figure 2.4, bottom). This translation process obscures differences in absolute directionality, but allows the relative position of horses within each site to be compared between monuments of different orientation. At deer stone sites, mounds which yielded horse remains, but were not conclusively associated with a particular deer stone were excluded from this portion of the analysis. 18 After plotting relative position of each horse burial mound, I combined these data with dental estimates of age and sex patterns to visualize spatial trends in equine demography. To maximize sample size, new data produced in this study were combined with published site maps and age/sex estimates from the fully-excavated site of Ushigiin Ovor (Takahama et al. 2006), as well as a partial sample of horse mounds from Urt Bulagyn KYR-2 (Allard et al. 2007). In some cases, previously published data provided only general age estimates (e.g., adult , juvenile ). To accommodate visualization, an arbitrary age class of 9-12 years for adult, and 0-3 years for juvenile horses was assigned to these specimens. Horses from Ushigiin Ovor and Urt Bulagyn KYR-2 were not included in mortality profile analysis, nor were these specimens analyzed for osteological changes. 19 Figure 2.4. Schematic showing procedure for rotating deer stone sites to share a common axis, and mapping relative position between horse burial mound and nearest deer stone. Daagan Del map by J. Bayarsaikhan and T. Tuvshinjargal. Finally, I synthesized these spatial and demographic data with osteological study of skeletal changes to the premaxilla linked to horse transport. Previous analyses (Taylor et al. 2015; Taylor et al. 2016) indicated that remodeling to the medial aspect of the equine premaxilla is linked to 20 heavy exertion or stress, while changes to the lateral aspect of the same bone may also be caused by chronic exertion, or perhaps bridle equipment. Although the development of these features probably increases with use, age estimates do not appear to explain either lateral or medial groove depth among ridden horses (Taylor et al. 2015). In analyzed samples of wild or unworked equids, both kinds of premaxillary remodeling appear to be reduced or absent (Taylor et al. 2015; Taylor et al. 2016). Thus, while individual specimens cannot be usefully assigned to a transport or nontransport category by cranial morphology alone, the severity of premaxillary remodeling across spatial and demographic categories provides another line of evidence to test ideas about DSK horse use. For each specimen with premaxillary preservation (n = 9), I combined estimates of medial and groove depth, corrected for differences in bone width and age (data from Taylor et al. 2015) into a Nasal Groove Index (size and age corrected lateral groove depth + size and age-corrected medial groove depth). This simplifies a complex anatomical issue, but produces a unidimensional measure which can be easily compared spatially. Resultant data were superimposed on the composite site map in order to visualize pathology patterns within the DSK sample. 2.4. RESULTS 2.4.1. Mortality profiles The sample of 25 horse skulls from satellite burials surrounding deer stones and khirigsuurs yielded a high proportion of subadult animals (n = 9, or 36%, Table 2.1). Due to poor preservation or age-related tooth loss, many DSK horses could not be assigned a sex, but the single female identified was estimated to be over 20 years of age. The sample also contained an unusually high frequency of prime age adult horse remains (n = 12, or 48% of the sample, Figure 2.5). Among those horses between six and 15 years that could be assigned a sex (n = 7), all specimens were male. Age Range 0-3 3-6 6-9 9-12 12-15 15-18 18+ Frequency 9 0 4 3 5 0 4 Table 2.1. Number of sample DSK horse specimens identified in each age category. 21 Figure 2.5. Mortality profile histogram for DSK sample, showing high frequency (proportion) of young horse remains, and second peak of adult animals between 6 and 15 years. ID Median est. age Sex Medial groove (mm) Lateral groove (mm) Deer stone/ khirigsuur Location NGI NMM 8 7 Male 1.488 0.871 khirigsuur E 2.471 NMM 30 13.5 Male 1.662 0.702 khirigsuur SE 2.180 NMM 28 10.5 Male 0.813 0 khirigsuur SE 0.895 NMM 3 24 1.821 1.879 deer stone E of DS 3.155 NMM 9 NMM 27 24 7.5 Poss. Male Male Male 1.576 0.954 0 0.613 deer stone deer stone 1.975 1.34 NMM 1 24 Indet. 0.61 0.551 deer stone NMM 10 NMM 18 2.25 0.875 Indet. Indet. 0.505 0.696 0 0 deer stone deer stone E of DS SE of DS E of DS? (stone missing) E of DS NW of DS 1.004 0.775 0.628 Table 2.2 Nasal groove index scores for analyzed horses from deer stones and khirigsuurs, along with demographic estimates and spatial provenience. 22 2.4.2. Spatial patterning While juvenile horse remains tended to be concentrated on the periphery of khirigsuurs, all adult male horse heads with provenience came from mound contexts along the east/southeast axis of the monuments (Figure 2.6A). The study sample was cobbled together from a variety of discrete khirigsuur sites, but a complete feature excavation at Ushigiin Ovor by Takahama et al. (2006) corroborates this result. At this site, three male horses were recovered in mounds 9, 11, and 15, adjacent to the eastern fence (these specimens are included in Figs. 2.6A/2.7A). The newly analyzed sample did not contain female horses from khirigsuur contexts, but Allard et al. (2007) reported adult females along the southern axis of Urt Bulagyn KYR-2, paired with juvenile burials to the immediate southwest (Figure 2.6A, lower left). At deer stone sites, two male horse crania were recovered from the east and southeast edge of monuments at the sites of Ulaan Tolgoi and Tsatstain Khoshuu, while a single female horse head was identified near the northeastern perimeter of the site of Daagan Del (Figs. 2.4 and 2.6B). 2.4.3. Nasal pathologies Of the 25 horse heads analyzed, nine had a sufficient level of preservation to measure premaxillary remodeling, including two juveniles, two adults of indeterminate sex, and five adult males. Results indicate high levels of pathology formation in four horse skulls from east or southeast mounds (Figure 2.7A and 2.7B, Table 2.2). These specimens, including three male animals and one possible male, produced NGI values of greater than 1.5, and two had extremely pronounced lateral grooves of greater than 0.8 mm. In contrast, two juveniles in the sample showed little in the way of nasal pathologies (Table 2.2). 2.5. DISCUSSION 2.5.1. Mortality profiles Demographic data provide additional support for the hypothesis of DSK horse breeding and herd management. A high frequency of juvenile horse remains is consistent with previous demographic data from the site of Urt Bulagyn (Allard et al. 2007), and aligns with expectations for pastoral culling of surplus male animals before breeding age (Levine 1999). The single female identified was near the age of reproductive senescence, corroborating the examples of elderly mare sacrifice reported by Allard et al. (2007). Although a similar demographic profile can be produced by other cultural formation processes, such as scavenging or group slaughter (Levine 1999:29-33), these results are generally consistent with the practical requirements of pastoral culling for meat or dairy (Allard et al. 2007). In contrast to expectations for herd management, however, results also indicated a significant proportion of prime-age male horse remains between the ages of six and 15 years (Figure 2.5). In ethnographic horse herds, stallions or geldings can reach ages of up to 30 years before they are slaughtered (Levine 1999), although breeding stallions may be culled at around 12-15 years in contemporary Mongolia (Enkhtuvshin and Tumurjav 2011:169). 23 Figure 2.6. A. Schematic representation of demographic patterns at seven khirigsuur sites (Urt Bulagyn 1, Urt Bulagyn 2, Ushigiin Ovor, On Khad, Nukhtiin Am, AD40, Zeerdegchingiin Khoshuu). Male horses are shown as diamonds, female horses as triangles. In cases where only general life history category was available (e.g., juvenile, adult), adult specimens are depicted as nine years a d ju e iles as t o ea s of age. Maps ha e ee otated less tha ◦ to o e t fo differences in E/SE orientation, and scaled such that the relative distance between mound center and the first row of horse burial mounds is equal for all features. Khirigsuur map by J. Bayarsaikhan and T. Tuvshinjargal. Data from Taylor et al. (in press), Allard et al. (2007), and Takahama et al.(2006). B. Schematic representation of demographic patterns at 13 deer stones from nine localities (Khushuutiin Gol, Bor Hujiriin Gol, Ulaan Tolgoi, Tsatstain Khoshuu, Jargalantyn Am, Uguumor, Daagan Del, Nukhtiin Am, and Ushigiin Ovor). Male horses are shown as diamonds, female horses as triangles. In cases where only general life history category was available (e.g., juvenile, adult), adult specimens are depicted as nine years and juveniles as two 24 years of age. Maps have been rotated (less than 90◦) to correct for differences in E/SE orientation, but retain equivalent scales. Figure 2.7. A. Schematic representation of nasal grooving at three khirigsuur sites (On Khad, Zeerdegchingiin Khoshuu, and Urt Bulagyn 1). Male horses are shown as diamonds, female horses as triangles. Specimens with measurable nasal grooves and mapped provenience are shown in bold (n ¼ 2). Maps have been rotated (less than 90◦) to correct for differences in E/SE orientation, and scaled such that the distance between mound center and the first row of horse burial mounds is equal for all features. Khirigsuur map by J. Bayarsaikhan and T. Tuvshinjargal. B. Schematic representation of nasal grooving at three deer stone sites (Khushuutiin Gol, Tsatstain Khoshuu, and Ulaan Tolgoi). Male horses are shown as diamonds, female horses as 25 triangles. Specimens with measurable nasal grooves and mapped provenience are shown in bold (n = 5). Maps have been rotated (less than 90◦) to correct for differences in E/SE orientation, but retain equivalent scales. Most of the prime-age specimens recovered from the DSK sample, particularly those between six and 12 years of age, are thus unlikely to have accumulated through typical herd management techniques. Other processes can produce assemblages of adult male horses, including selective hunting of larger wild animals for meat (Levine 1999:33). Members of the genus Equus first arrived in Asia during the late Pliocene (Webb and Hemmings 2006:14), and wild equids were apparently still exploited by hunters in Mongolia during the mid-Holocene (e.g., Janz 2012:6364). As a result, the possibility of wild horse hunting by DSK people cannot be outwardly dismissed. Nonetheless, several specimens exhibit cranial deformation which must be caused by a halter or noseband (Taylor et al. 2016). As a result, the hunting of wild animals is an unlikely explanation for the observed demographic profile. Figure 2.8. Idealized mortality profiles for a managed pastoral horse herd (solid line, after Levine, 1999:31), and a transport assemblage (dotted line, after Levine, 1999:30), as compared to DSK assemblage. 26 Figure 2.9. Mortality profile from Arzhan-2, from Benecke (2007). All horses are male, and classified according to median estimated age. Diagram shows emphasis on adult male horses for transport burials. Archaeological comparisons suggest that the ritual sacrifice of transport animals may be a viable explanation for adult male horse remains at DSK sites. In the early Iron Age, equestrian peoples built funerary barrows in many parts of Central Asia, including western Mongolia, eastern Kazakhstan, and south Siberia. Although important cultural differences distinguish early Iron Age cultures from the preceding DSK complex, mortuary features of this period provide a useful window into the demography of transport horse assemblages. At the barrows of Arzhan-2 and Pazyryk in South Siberia, and Berel in far eastern Kazakhstan, fully-equipped and outfitted riding horses were included alongside human burials (Rudenko 1970; Benecke 2007; Lepetz 2013). The animals selected for these inhumations were male, and in cases of soft tissue preservation, many of these early Iron Age horses appear to have been gelded (Rudenko 1970; Littauer 1971; Benecke 2007). This context has led to the interpretation that burial horses chosen for mortuary inclusion were riding mounts (e.g., Benecke 2007:120). Such assemblages of transport horses occasionally include subadults, as well as old animals of 20 years or more (Rudenko 1970:119; Levine 1999:30). However, most appear to fall between six and 15 years of age (Rudenko 1970; Levine 1999; Benecke 2007:119; see Figure 2.8). Analogical comparison of Arzhan horses with those from DSK satellite mounds illustrates similarities between the ages of male specimens at these sites (Figure 2.9). If the DSK male horses were similarly used for transport, these animals might exhibit evidence of special ritual treatment, and show osteological features related to use in transport. 27 2.5.2. Spatial patterns Relative position maps (Figure 2.6) reveal age and sex patterning across horse burials, pointing to special ritual treatment for adult male animals at khirigsuurs, and perhaps deer stones. All six definitively adult male horse specimens from the khirigsuur sample came from mounds along the east/southeast axis of their respective sites. Several of these mounds were noted by the investigator as the most prominent ritual features at their respective sites, owing to their larger size and central location (e.g., Fitzhugh 2006:37; Fitzhugh and Bayarsaikhan 2008:21). This concentration of adult males is identifiable in both the newly analyzed sample, as well as previously published ones. For example, the only adult male horse from Urt Bulagyn KYR-2 (Allard et al. 2007) came from such a context, as did the three male horses at Ushigiin Ovor mentioned above (Takahama et al. 2006). Due to the peculiarities of the study sample, the generalizability of this pattern cannot yet be usefully evaluated, and the possibility remains that monuments of differing scales and layouts may yield different demographic patterns. Complete intra-site excavation like that of Takahama et al. (2006) will be necessary to evaluate these issues more effectively. However, these initial results point to preferential ritual treatment for adult male horses at khirigsuurs, and hint at a role for such animals in transport. Spatial patterning in horse age and sex is less obvious at deer stone sites (Figure 2.6B), but may also reflect eastern positioning of male horse features. Two definitively male specimens identified in the deer stone site sample came from mounds along the east/southeast edge of their sites, although one (Tsatstain Khoshuu Feature 2) was not buried in the innermost mound (Fitzhugh 2005). A single elderly horse cranium from the site of Khushuutiin Gol found adjacent to the central deer stone was likely male, as excavation notes indicate that a canine was removed for radiocarbon dating (Fitzhugh and Bayarsaikhan 2008:43). However, the mandibular canines may have worn away from extreme age. A third male horse from the site of Jargalantyn Am appears near the periphery of Figure 2.6B (top right). However, this apparent location is actually caused by a scaling issues: examination of the original site maps shows that this horse originated from the eastern interior of an enormous horse burial complex (Bayarsaikhan 2011). Similarly, although a female horse appears in the center-right of Figure 2.6B, this specimen came from the outermost northeast mound of a large stelae cluster at Daagan Del (Bayarsaikhan and Tuvshinjargal 2013, see Figure 2.4 this paper).The youngest horse specimens from the deer stone sample (<1 year) were recovered from smaller features along the northern edge of Deer Stone 5 at the site of Ulaan Tolgoi (Fitzhugh 2005). Few horse heads in the analyzed sample were unearthed from the west or southwest perimeter of deer stone sites, so demographic patterns in these locations cannot be evaluated. Thus, while preliminary results from deer stone sites could indicate the preferential burial of male horses in eastern locations, a more spatially diverse sample drawn from both isolated deer stone sites as well as larger stelae complexes will be necessary to test this hypothesis. 2.5.3. Osteological data Cranial pathology analysis supports inferences from demographic and spatial patterning, suggesting that adult male horses were used for transport. Groove indices in four sample specimens (Table 2.2) are markedly higher than those documented in unworked horses. The DSK specimens with the most extreme cases of medial and lateral premaxillary remodeling appear to be located nearest to the monument along the east or southeast axis (Figure 2.7A and B), and 28 most come from definitively male horses (Table 2.2). In a sample of 31 wild equids and feral domestic horses from ancient and modern contexts, none exhibited a combined lateral and medial groove score of more than 1.5, or a lateral groove of greater than 0.8 mm. In contrast, several DSK adult males had marked medial and lateral grooves, exceeding the range yet observed in wild or captive unworked equids. In addition, the elderly horse (of indeterminate sex) from the site of Khushuutiin Gol also showed a marked depression to the bridge of the nose, which was probably caused by a bridle noseband (Taylor et al. 2016). Although sample size is small, these data suggest that many of the adult male horses buried at DSK sites were used for transport. Because of their occurrence in smaller, exterior mounds, juvenile and female horse remains in the studied sample were apparently subjected to a higher degree of weathering and taphonomic degradation. As a result, premaxillary bones were rarely preserved for osteological study. However, the two juvenile horse heads with suitable preservation showed very little premaxillary remodeling. Contemporary Mongolian herders begin breaking and training horses at the age of one year (Enkhtuvshin and Tumurjav 2011:173-4). Without a better understanding of the process of groove formation, few conclusions can be drawn regarding whether or not these young animals were used in transport. Although animals younger than 3-4 years are not typically ridden heavily among modern herders (Bold 2012), a few juvenile horses were buried in the central eastern periphery of khirigsuurs (Figure 2.6A). Rudenko (1970:119) noted the regular inclusion of 2-3 year old horses in Pazyryk riding horse burials, and a similar pattern could characterize some DSK assemblages. Additional research will be necessary to explore these questions, and to characterize pathology levels in adult female horses from DSK archaeological contexts. Nonetheless, these preliminary results support the inference that adult male horses found in prominent mound locations were used for riding and/or traction. 2.5.4 Implications of DSK horse transport Results from demographic estimates, spatial comparisons, and nasal pathology measurements suggest that male horses were used for chariotry or riding in DSK society, and were buried in positions of ritual significance. When combined with other archaeozoological data for domestic livestock use (Allard et al. 2007; Houle 2010; Broderick et al. 2014), results of this study reveal the DSK complex as a pastoralist society, combining equine transport with a diverse livestock economy. This inference is consistent with other evidence for an expansion of horse transport out of interior Asia during the second millennium BCE (Hanks 2010:475-6), and supports the idea that mobile pastoralism's spread in East Asia was closely linked to developments in horse control (e.g., Beardsley 1953). Horse transport in the DSK complex warrants its consideration in models for late Bronze Age social dynamics and the spread of the domestic horse to other parts of East Asia (Taylor et al. 2015). The late Bronze and early Iron Age saw important technological changes which facilitated mounted horseback riding (Renfrew 1998; Drews 2004). Towards the end of the second millennium BCE, chariots arrived in the city of Yinxu and Shang Dynasty China (Linduff 2003; Kelekna 2009a). The geographic origin of early Chinese domestic horses has not been established (Yuan and Flad 2006), but many kinds of burial goods and material culture point to Bronze Age interaction between Chinese groups and the northern steppes (Shelach 2009). It is likely that one or more routes of exchange passed through Mongolia (Honeychurch 2015:208210). Additional osteological research may help to shed light on the type of equine transport 29 used by DSK people, and to evaluate the role of DSK horse use in broader technological and cultural developments in riding and chariotry. Future study should evaluate the patterns identified in this study through detailed intra-site analysis, so as to mitigate issues caused by compiling many individual specimens from a wide variety of DSK monuments. If the patterns identified here are indeed representative, complete assemblages of horse burials from within a single khirigsuur locality should continue to reveal adult males with transport pathologies concentrated in eastern satellite features. Comprehensive excavation should seek to clarify the poorly characterized demographic patterns at deer stone sites, which will also enable improved comparison between sacrifice practices at different monument types. For example, consistency between deer stones and khirigsuurs would corroborate the idea of a shared ritual function, such as memorializing fallen leaders or relatives (Fitzhugh 2009a:191). On the other hand, major differences in equine demography, provenience, or cranial osteology between horses from different monumental contexts might reveal new information about the dynamics of late Bronze Age ritual. It may be particularly instructive to compare the demographic patterns at monuments with a carefully planned arrangement of horse mounds, with those which seem to have an accretionary or haphazard arrangement (e.g., Wright 2014:153). Finally, ongoing study should address the possibility of equine castration at DSK sites. Some question remains as to whether DSK monuments were constructed in single events, or over extended periods (Wright 2012), and caution must be used in assuming that ritual assemblages reflect actual death assemblages (Levine 1983). However, if the high proportion of adult male horse remains in DSK mounds reflects the original herd composition, it is likely that many of these animals were gelded. This pattern would align with castration observed in riding horses in barrows at Pazyryk and Arzhan (Rudenko 1970; Benecke 2007). Among modern Mongolian herders, male horses are castrated in the spring before their second year unless they are to be used as breeding stallions (Enktuvshin and Tumurjav:170). Gelding of young males indicates a strong familiarity with equine ecology, and early castration induces developmental changes which may be advantageous in transport horses (Littauer 1971). Qualitative differences in cranial osteology between complete and gelded males, such as facial elongation and convexity (Littauer 1971; Olsen 2006b) may be useful in developing systematic archaeological techniques for identifying DSK castrates. 2.6. CONCLUSION Demographic profiles suggest that people of the Deer Stone-Khirigsuur complex herded horses and used them for transport, an inference that is corroborated by both spatial patterning and cranial osteology. Consistent with later mortuary traditions in Central Asia, adult male DSK horse heads seem to have been preferentially buried in important ritual contexts, and many show skeletal evidence of exertion or bridling implicating their use as transport animals. Additional research may provide further clues to the role of DSK horse use in larger social and technological processes, such as the spread of the domestic horse to East Asia, late Bronze Age innovations in horse control, and the development of complex pastoral societies. Results of this research support a growing body of archaeological evidence for DSK mobile pastoralism, and could point to an important role for horse transport in the origins of specialized herding in Mongolia. 30 CHAPTER 3: EQUINE CRANIAL MORPHOLOGY AND THE ARCHAEOLOGICAL IDENTIFICATION OF RIDING AND CHARIOTRY IN BRONZE AGE MONGOLIA Antiquity 89(346): 854–871 doi:10.15184/aqy.2015.76 William Timothy Treal Taylor1, Jamsranjav Bayarsaikhan2 & Tumurbaatar Tuvshinjargal2 The adoption of the horse for chariots, wagons and riding had a major impact on human societies, but it has proved difficult to reliably identify early domesticated horses in the archaeological record. This comparative study of equine palaeopathology addresses the problem by analysing wild and domestic horses used for traction or riding. Osteological changes to the skull appear to be the result of mechanical and physiological stress from the use of horses for transport. The results are applied to archaeological examples from the Deer Stone-Khirigsuur Complex of Bronze Age Mongolia (1300–700 BC) and show that those horses were probably bridled and used for transport. Keywords: Mongolia, Bronze Age, zooarchaeology, nasal remodeling Deer Stone-Khirigsuur Complex, equestrianism, 1 Maxwell Museum of Anthropology, MSC01 1050, 1 University of New Mexico, Albuquerque, NM 87131–000, USA 2 National Museum of Mongolia, Juulchin Street-1, Ulaanbaatar 210646, Mongolia 3.1 INTRODUCTION Horseback riding and the use of chariots has been linked with dramatic changes to the form and scale of social organisation among prehistoric peoples, but the chronology of their adoption in East-Central Asia remains poorly understood. Towards the end of the Bronze Age (c. 1300– BC , dee sto e stelae, a o pa ied stone khirigsuur mounds and ritual horsesacrifice features, appear to be the archaeological signature of the first horse-riding nomadic pasto alists i the Easte “teppe of Eu asia. Ho se t a spo t i this Dee “to e-Khirigsuur Co ple ould i pl i po ta t ha ges i o ilit , so ial st atifi atio a d structure (Anthony et al. 1991; Honeychurch 2015: 212–15), and suggest an expanded role in regional patterns of social interaction during the late Bronze Age. Where other forms of direct evidence are lacking, faunal pathology is a promising means of identifying equine transport through archaeological data (Dietz 2003). Dental morphology (Anthony & Brown 1998) and other osseous changes to the appendicularand axial skeleton have been linked to equestrian activity (Levine 1999; Olsen 2006b:93; Bendrey 2007). Archaeologists have, however, struggled to pinpoint pathological markers of riding or traction that pre-date the use of the metal bit and can be regularly assessed in archaeological samples. This paper develops expectations for anatomical changes to the equine skull that should accompany the use of do esti ho ses fo t a spo t. Ho se t a spo t he e efe s to oth ou ted ho se a k riding and the use of horses as draught animals to pull chariots or other vehicles. Using precision measurement of 3D models, these predictions are tested using a sample of feral, captive and domestic horse remains from museum collections. Finally, the results are applied to an 31 assemblage of horse skulls from archaeological sites belonging to the Deer Stone-Khirigsuur (DSK) Complex to evaluate the possibility of horse transport in the Eastern Steppe during the late Bronze Age. 3.1.1 The horse in ancient Mongolia and beyond Between 1300 and 700 BC, and perhaps earlier, deer stones and khirigsuur monuments spread across north-west and central Mongolia. Khirigsuurs are large, fenced stone burial mounds. They regularly contain human remains (Littleton et al. 2012), although they may have also served non-mortuary functions (Wright 2014). Deer stones may have been memorials for revered warriors. These standing stones typically depict stylised earrings, belts and representations of the face, along with belts, weapons, deer and other animal carvings (Honeychurch et al. 2013: 80). Together, these monument types appear to be archaeological manifestations of a single cultural complex (Fitzhugh 2009a). Many have suggested that DSK people were also pastoralists using the horse for transport (Allard et al. 2007; Houle 2010; Honeychurch et al. 2013), a contention that is supported by developments elsewhere in Eurasia. From the late third and early second millennium BC, there was an expansion in the use of chariots across the continent, and by 1200 BC, cultures using the horse for transport had already spread eastwards out of Central Asia (Hanks 2010:475–76). Iconographic depictions of horse riders, and bone cheekpieces from mortuary contexts, show that horse-control technology was present in the forested regions bordering northern Mongolia c. 1400–1000 BC, coeval with the early DSK period (Legrand 2006). Due to an absence of actual tack or riding artefacts in the DSK archaeological record, there remains some question as to how horses were used by DSK people (Honeychurch et al. 2009). Horse sacrifice was the most important aspect of ceremonial activity at deer stones and khirigsuurs, and DSK sites have produced circumstantial evidence for horseback riding and chariotry. Small stone mounds around the perimeter of both monument types contain equine skulls, often accompanied by cervical vertebrae and hoof bones (Fitzhugh 2009a). Previous faunal analyses suggest that horses were culled and eaten by DSK people (Allard et al. 2007; Houle 2010: 126–29). Most compellingly, deer stone carvings show artefacts and other features that suggest the ho se as used fo t a spo t. Fo e a ple, the elt of weapons and tools carved into many of these deer stones commonly includes a small horse representation, alongside weapons and other important equipment (Volkov 2002) (Figure 3.1, left). Deer stone carvings show bow-shaped objects, often found in Chinese chariot burials (Wu 2013: 40), which might have been used as chariot rein hooks (Fitzhugh 2009b; Fitzhugh & Bayarsaikhan 2011: 178) (Figure 3.1, right). Although only one deer stone depicts a chariot (Volkov 2002), such vehicles are a regular feature of Mongolian rock art panels attributed to the late Bronze Age (Honeychurch 2015: 192–93). Given these considerations, it is possible that people of the DSK Complex were nomadic horsepastoralists (Honeychurch et al. 2013). 32 Figure 3.1. Depictions of small horses alongside weapons such as daggers, bows and quivers (left–centre) on deer stones in Mongolia; also depicted are chariots (second from right) and ha iot ei hooks o o -shaped o je ts fa ight efe e ed in the text; modified from Volkov (2002). If DSK people did indeed use horses for riding or chariotry, they may have played an underappreciated role in the spread of the horse into other parts of East Asia. Centuries before the “ilk ‘oad t ade outes e e fo alised, the g assla ds of the “teppe acted as an informal “teppe ‘oad , fa ilitati g ultu al a d e o o i e ha ge a oss the Eu asia o ti e t (Christian 2000). Domestic horses first reached China, along with chariots, during the late Shang Dynasty (1600–1050 BC), the earliest specimens dating to c. 1300–1200 BC (Yuan & Flad 2006; Kelekna 2009b), much later than elsewhere in Central Asia. The geographic source of these first animals is an open question (Yuan & Flad 2006: 258–59), but severa lines of evidence implicate the steppe cultures of Mo golia. Alo g Chi a s o the f o tie , steppe artefacts in late Bronze Age burials, particularly those of elites, suggest an acceleration of Sino-Mongolian interaction (Shelach 2009: 128–29). Perhaps most interestingly, recent genetic research indicates a close phylogenetic link between ancient Chinese and modern Mongolian horses (Cai et al. 2009). Clarifying the role of the horse in DSK society is thus crucial to an understanding of East Asian social dynamics in the first and second millennia BC (Honeychurch 2015: 205–10). 3.1.2 Archaeozoological identification of riding and chariotry Although human management of domestic horses dates to the Eneolithic, c. 3500 BC (Olsen 2006a; Outram et al. 2009), archaeologists disagree as to when domestic horses were first used 33 for riding and human transport (e.g. Renfrew 1998; Anthony 2007), and how to identify the signature of such transport in the archaeological record (Levine 1999). Diagnostic horse-control devices, such as metal bits, are rarely recovered in East Asia from before the first millennium BC (Mair 2003: 170). Leather harnesses and other organic methods of control used by early equestrians are unlikely to have been preserved in most archaeological contexts (Olsen 2006b). In the absence of texts or exceptional preservation, palaeopathology provides the most direct dataset for the evaluation of ancient horse transport (Dietz 2003). Osteological techniques for identifying equestrianism in horse remains have been debated at length (e.g.Anthony&Brown 1998, 2003; Levine 1999; Olsen 2006a,b; Bendrey 2007). The bestk o zooa haeologi al i de of e uest ia is is it ea : lo alised e elli g of the anterior surface of the second premolar caused by grinding or chewing a bit. Experimental and comparative studies have suggested that bevels greater than 3mm in magnitude are evidence of equestrianism (Anthony & Brown 1998, 2003; Anthony et al. 2006). In recent years, additional research has bolstered the argument that metal bits can produce archaeologically recognisable ha ges to the ho se s se o d p e ola , a d ause e o e formation to the diastema of the lower jaw (Bendrey 2007; Outramet al. 2009). The impact of an organic bit or halter on premolar form is less clear. Although Anthony et al. (2006) argue that organic bits should also cause measurable bevelling, natural malocclusion can produce similar changes to the teeth of unworked horses (Levine 1999; Olsen 2006b:100–101).More importantly, many forms of early horse control appear to have relied on pressure from a noseband, without the use of a bit at all (Littauer 1969). Difficulty in resolving the bit wear debate increases the importance of seeking alternative archaeozoological criteria for identifying horse transport. The high frequency of equine crania in the archaeological record of many parts of Central Asia (Kuzmina 2006) makes them particularly useful for the palaeopathological study of horse transport. Bendrey (2008) compared dozens of the skulls of horses used in riding and traction with those of E. przewalskii, which has never been domesticated. He identified that new bone formation at the site of nuchal ligament attachment (enthesopathy) occurred in high frequency among highly trained horses. Although similar features can be caused by other factors such as bacterial infection (e.g. Bendrey et al. 2011), nuchal enthesopathy appears to be commonly caused by habitual activity such as horseback riding (Figure 3.2a). Despite this interesting pattern, Bendrey concluded that nuchal pathologies had limited utility for archaeological identification of horse use, as age dependency was a major concern (Bendrey 2008: 30). In addition, the E. przewalskii specimens showed a wide range of ossification levels, overlapping sig ifi a tl ith o ked ho ses used for traction (Figure 3.2B). The e thesopathi patte s of P ze alski s ho ses e o ded Be d e , ho e e , may not accurately characterise those of predomesticate horses. Most or all of the E. przewalskii spe i e s i Be d e s stud a e f o zoo collections (Robin Bendrey, pers.comm.). Although these captive equids were never driven or ridden, the impact of zoo related stress on the equine skeleton may mimic the effects of horseback riding. Horsesrespond to negative stimuli through avoidance o pa i ked flight Dietz : . I apti e a i als, f ust atio of these atu al avoidance mechanisms can induce neurological problems that cause nuchal stress (Hosey et al. 2013: 231). For example, up to 40% of wild e uids i zoos de elop ste eot pies , he e the animal engages in repetitive headshaking or similar behaviour (McDonnell 1988). Other aspects of zoo life, such as chronic posture changes associated with feeding, may also increase neck 34 st ai o e a ho se s lifeti e. As a result, undomesticated but captive equids are not an ideal comparative sample for the identification of cranial pathologies related to transport. Figure 3.2. A) Nuchal ossification on a ridden horse (left); B) ossification scores for horses used for riding, traction o d i i g , as o pa ed to u o ked E. przewalskii from European zoos (right); modified from Bendrey (2008). Figure 3.3 Medial (A) and lateral (B) groove formation on the nasal process of the premaxilla (incisive bone) of a ridden horse, US General Joh J. Pe shi g s a ho se Kid o left , a d the same region on a feral Chincoteague pony (right); specimens from the Smithsonian National Museum of Natural History. In addition to nuchal bone formation, osseous changes to the nasal portion of the skull are a new and potentially useful marker of ancient equine transport. Siberian people used pointed bone cheekpieces with tightened halters to control domestic reindeer during the Iron Age, and this strategy may have a more ancient history in the region (Fedorova 2003a & b). Studded nosebands (Dietz 2003: 191) or burred cheekpieces were especially common methods of control in early horse headgear (Littauer 1969: 290–92). Such devices, which rely on stimulation of pressure-sensitive areas of the face, may have preceded the first use of the bit (Littauer 1969: 293), and were used in bridles for chariot horses on the northern steppes as early as the third 35 millennium BC (Kuzmina 2000). Pressure on the nose and cheek from nosebands or burred cheekpieces will irritate sensitive facial nerves, and nosebands may hinder the breathing of animals to the point of tissue damage (Littauer 1969: 293; Brownrigg 2006: 170). Despite the great antiquity of these devices, most modern bridles still rely to some degree on pressure and stimulation of the nasal region of the horse for control (Dietz 2003: 191). In the collections we examined, we discovered that many modern, ridden horse skulls have a pronounced groove along the dorso-medial border of the premaxilla/incisive bone (hereafter efe ed to as the edial g oo e Figu e .3). This feature has been investigated in several veterinary studies (Perez & Martin 2001; Vanderwegen & Simoens 2002). It is sometimes accompanied by a second, exterior groove, located rostrally along the o e s late al a gi late al g oo e . T o elated a ato i al e ha is s a e p o a l i ol ed i su h g oo e formation. Medial grooves are associated with activity of the lateralis nasi muscle and its accessory cartilage. It is hypothesised that sustained nostril dilation under conditions of chronic heavy breathing causes this muscle to hypertrophy, which in turn causes bone remodelling (Perez & Martin 2001). The second, lateral groove may be developmentally related. This lateral grooving appears, however, to be related to an internal nasal branch of the infraorbital nerve (Perez &Martin 2001). Preliminary study suggests that both kinds of remodelling are generally absent from wild equids such as zebra (Vanderwegen & Simoens 2002: 200). Insofar as the use of horses for transport causes increased respiration, grooves may track meaningful differences in human horse use. Domestic ridden animals should demonstrate extensive nasal remodelling caused by chronic exertion.Given that heavy breathing and nasal dilation are common stress responses in captive animals (Casey 2002; Harris et al. 2006; Hosey et al. 2013: 237), grooving should be more severe in zoo populations than in feral animals. In contrast, neither exertion nor other anthropogenic stressors should affect feral horses, where little grooving is expected. Finally, amongst animals where human transport drives osteological changes, high levels of nuchal ossification should also be matched by deeper nasal grooves. 3.2 MATERIALS AND METHODS We studied a sample of 31 feral, ridden and zoo horse crania from museum collections at the National Museum of Mongolia, Khustai Nuruu National Park, the Museum of Southwestern Biology, the Smithsonian National Museum of Natural History and educational collections at the Navajo Nation Veterinary Clinic in Chinle, Arizona (see Appendix III). Using a NextEngine3D scanner, the nasal and nuchal portions of the skull were scanned at a resolution of 2000 dots per inch (DPI). For each horse, the maximum extent of new bone formation at the nuchal crest was measured from a 3D model using open-source measurement software. Following criteria outlined in Bendrey (2008), we assigned a qualitative score of 1–6 to each specimen based on coverage and depth of new bone formation. In cases of a split score between the upper and lower portion of the occipital (e.g. 3/2), the average of the two values (in this case 2.5) was used in analysis. To compare these score distributions across groups, we followed this with a nonparametric Kruskal-Wallis analysis of variance by rank and pairwise Wilcoxon sign-rank tests. 36 Figure 3.4. Medial groove depth, measured perpendicular to the intersection of groove walls and the dorsal surface of the premaxilla (incisive bone), shown here on an archaeological specimen. Figure 3.5. Nuchal ossification/musculoskeletal stress marker (MSM) scores (1–6) for museum sample specimens from ridden horses (Bendrey 2008 and data from this study); driven horses (Bendrey 2008); E. przewalskii from probable zoo provena e Be d e ; P z. ; ho ses of k o zoo p o e a e P z. , this stud a d fe al a i als this stud . 37 Next, the maximum depth of medial and lateral nasal grooves was measured digitally for specimens in each group. From the point of deepest groove formation, a straight line along the plane of the nasal process of the premaxilla was drawn at the point of intersection between the g oo e all a d the a of the o e s do sal su fa e Figu e 3.4). We measured groove depth perpendicular to this first line. In cases of asymmetry, we recorded the deeper of the two measurements. To account for variation in size between breeds, we subsequently normalised eachmeasurement to the diameter of the bone in the area of groove formation. As environmentally stimulated bone changes may be age-dependent, demographic differences between samples might affect the observed patterns. To correct for this, we calculated an age estimate for each specimen using dental eruption and wear guides (Evans et al. 2007). When possible, we constrained estimates based on incisor morphology with crown-height measurements of cheek teeth (after Levine 1982). In cases where a precise estimate was not possible, we employed the median of the estimated age range in analysis. Whether or not a historically documented age was available, we recorded dental estimates and used these in analysis, in the hopes of maximising data comparability, and ensuring that bias was at least consistent across specimens. Ossification scores are ordinal data, so we assessed the relationship between estimated age and nuchal ossification using a correlation test “pea a s ho . A O di a Least “ ua es (OLS) linear regression between age and groove depth helped to estimate the effects of age on nasal remodelling. Using this regression, we then calculated an age-predicted lateral and medial groove depth for each specimen. Finally, we tabulated the residual between each spe i e s o se ed depth a d its age-predicted value, and compared this age-corrected metric across groups using a one-way ANOVA and pairwise ttests. Results for all tests are given below. 3.3 RESULTS 3.3.1 Nuchal ossification Nuchal ossification scores differ markedly between feral and zoo specimens (Figure 3.5). Corroborating findings by Bendrey (2008), ridden horses in museum collections had a unimodal, left-ske ed dist i utio of ossifi atio s o es, e t ed o alues of a o , hypertrophic projection between 7.5– i le gth ; Be d e : . E. przewalskii from zoos shared this pattern, supporting the hypothesis that nuchal ossification in captive animals may be exacerbated by differences in posture, stereotypy or other captivity-related stress to the neck area. Feral horses, in contrast, tended to have dramatically lower ossification levels. A few feral horses developed extreme enthesopathy, but scores for this group tended to a ds alues of (no ossification). A Kruskal-Wallis analysis of variance by ranks test suggests that these differences are significant across groups (K-W χ2 = 32.96, df = 4, p = <0.001), and pairwise Wilcoxon rank-sum tests indicate that this significance is driven by higher ossification levels in ridden horses than in feral (p<0.001), driven (p<0.01), or zoo Przewalski (p<0.001) horses. Nuchal ossification scores across all groups are also significantly age-dependent “pea a s correlation test p<0.01). Feral horses in this study had a younger mean age than the ridden sample, so differences in nuchal ossification could be influenced by systematic differences in sample age. Nonetheless, a pattern of reduced ossification in feral horses appears to persist 38 39 40 Figure 3.6. Nuchal ossification/musculoskeletal stress marker (MSM) scores for ridden horses (Bendrey 2008), driven horses (Bendrey 2008), Przewalski s ho ses f o Eu opea zoos Be d e 2008), and feral horses analysed in this study. across all adult age classes except the very youngest, where all horses display similarly low scores (Figure 3.6). When compared with feral specimens, nuchal enthesopathy in ridden horses thus seems to be much more compelling evidence for horse transport than previously recognised. Additional study will be necessary to assess the impact of other factors, such as body size, on nuchal bone formation. 3.3.2 Nasal remodelling As hypothesised, nasal remodelling also differs significantly across horses with different work histories. A one-way ANOVA (F = 9.74, p<0.001, with 2 and 28 degrees of freedom) provides strong evidence against the null hypothesis of equality in medial groove depth (Figure 3.7). Feral horses have a lower mean depth than ridden specimens (p<0.01), even after normalising the data to bone size and correcting for age (p<0.01). Zoo specimens show intermediate levels of medial remodelling, with deeper medial grooves than feral horses (p<0.01), and cannot be statistically distinguished fromridden specimens. Given the sedentary nature of zoo life, this nasal remodelling is probably not due to physical exertion, although it might be linked to captivity stress (e.g. heavy breathing). If so, captive equids should exhibit deeper lateral grooves than their wild counterparts. Groove measurements from one free- a ge P ze alski s ho se a e indeed markedly shallower than observations from E. przewalskii residing in zoos, providing preliminary support for this hypothesis (Figure 3.7). Akin to nuchal ossification, an OLS linear regression model indicates that medial groove depth is also nominally age-dependent (p<0.01). The correlation coefficient, however, is very small (0.028mm/yr), and age explains very little of the observed variance in medial groove depth (adjusted R2 = 0.18). Most importantly, even after calculating residuals between observed and age-predicted OLS residuals, ridden horses can still be clearly distinguished from feral specimens (p<0.01, Figure 3.7c). Ridden animals also show an association between groove depth and nuchal ossification. In feral and captive horses with either high nuchal scores or groove depths, extreme values of the corresponding pathology do not commonly co-occur (Figure 3.8). This supports the contention that a common mechanism, use in transport, drives the development of both features. In contrast, variation in pathology levels among feral and captive zoo horses is probably driven by a wider range of context-specific causes. Among all horses, lateral grooves formed somewhat inconsistently. Nonetheless, ridden horses appear to develop them at much higher frequency (Figure 3.9), and a one-way ANOVA followed by pairwise t-tests indicate that ridden horses have significantly deeper grooves than feral specimens (F = 5.166, p<0.05). Similar to medial groove depth, age and lateral groove depth are loosely related (coefficient=0.016, p<0.10, adjusted R2 =0.07).Due to the high frequency of specimens without lateral grooves, including all sub-adult and most feral/zoo horses, regression may be a flawed technique for age-correction. Regardless, comparing OLS residuals did not remove differences in lateral groove depth between wild and ridden horses (ANOVA p = 0.07, Figure 3.9b). Moreover, when lateral grooving did occur, it was often matched with extreme medial depths. Among captive horses even pronounced medial grooves were not accompanied by significant lateral remodelling. As a result, it appears unlikely that inter-sample age differences are driving lateral groove patterns. One mechanism that might increase the 41 Figure 3.7. A) Medial nasal groove depth across ridden, feral and zoo samples (top); B) the same data normalised to premaxilla (incisive bone) width (bottom left); C) corrected for age using OLS residuals (bottom right). 42 Figure 3.8. Plot of medial groove depth and nuchal ossification score showing co-occurrence of high values in ridden specimens. Figure 3.9. A) Lateral groove depth by group (left); B) after normalisation to bone width and correcting for age (right). 43 Figure 3.10. Schematic diagram of horse cranium, indicating position of simple rope halter relative to nasal remodelling, and the path of the infraorbital nerve (arrow). frequency of lateral remodelling is pressure on the nasal region and infraorbital nerve from bridle components, such as a low noseband or bridle cheekpiece (Figure 3.10). Variation in horse use and bridle morphology could explain the high frequency of lateral grooves among ridden horses, as well as their occasional asymmetry. The most extreme case of lateral remodelling o se ed i this stud elo gs to Ge e al Joh Pe shi g s a ho se Kidron, which saw active duty in the Spanish-American war. Extreme lateral remodelling is also present on horses from ancient Pazyryk (c. 600–300 BC) and Turkic (c. AD 600–800) burials in collections at the National Museum of Mongolia. Further research is needed to substantiate this hypothesis, but if lateral grooving is related to riding equipment, it may ultimately prove to be an especially valuable index of human horse use. These results indicate that nasal remodelling tracks both exertion and other forms of stress in horses. Further study of wild equids under high predation pressure may thus be needed before nasal pathologies can conclusively distinguish ridden from wild animals. Modern feral horses experience little in the way of predation, and might exhibit less nasal remodelling than heavily hunted populations. In contrast, horses experiencing high predation pressure in antiquity, such as pre-domesticate equids of Central Asia, would be expected to demonstrate deeper and more frequent medial grooves. These same horses, however, should also have infrequent lateral remodelling and limited nuchal ossification.Consideration of all three measures should enable the separation of equine transport from other mechanisms that might cause medial grooving or other cranial pathology in isolation. 3.4 ARCHAEOLOGICAL APPLICATIONS 44 Given the marked differences between feral and ridden horses reported here, cranial pathologies can be used to evaluate horse use in antiquity. Nasal groove depth and nuchal ossification scores provide an independent dataset for testing the hypothesis that DSK horses were used for riding or chariotry. Following the methods described earlier, we analysed a sample of 25 DSK horse crania from sites in central Mongolia, scanning these at high resolution. Eighteen of these specimens had associated radiocarbon dates, falling between 1337–769 cal BC (at 2-sigma confidence interval, see Fitzhugh 2009c: 219–20). For those specimens with sufficient preservation for nasal (n = 9) or nuchal (n = 6) analysis, we measured lateral and medial groove depth, assigned a nuchal ossification score and estimated age for each specimen. We compared the resultant palaeopathological data from the DSK specimens with the feral and domestic samples to test the hypothesis of DSK equestrianism. 3.5 DEER STONE-KHIRIGSUUR RESULTS Nu hal ossifi atio s o es fo the D“K sa ple ha e a ode of a o h pe t ophi p oje tio less tha . i le gth ; Be d e : . Although the sa ple size is s all, this result is statistically distinguishable from scores of the modern feral sample (p<0.05), and consistent with values from worked animals (Figure 3.11a). Although DSK nuchal ossification scores appear lower than modern ridden specimens, these values are also inconsistent with those of feral horses, and their distribution is visuall si ila to that of Be d e s di e population. Nasal remodelling provides stronger support for the hypothesis that DSK horses were used for transportation. A one-way ANOVA for medial groove depth between ridden, feral and DSK horses (p = 0.001), followed by Holm-corrected pairwise t-tests, indicates that the DSK sample is similar to ridden horses, and can be distinguished from the feral group (p<0.05, Figure 3.11b). This pattern holds even after controlling for the effects of both size (p<0.05) and age (p<0.01). As in ridden comparatives, deep groove scores correspond with higher nuchal ossification scores among DSK specimens. Lateral groove depth also occurs in the DSK sample at a high frequency similar to that of ridden horses (Figure 3.11c), and despite the small sample size, pairwise t-tests provide some evidence to separate uncorrected DSK lateral groove values from the feral sample (p = 0.10). Most tellingly, deep lateral and medial grooves co-occur in the DSK sample, as they did in the ridden comparatives (Figure 3.12). If the people of the DSK Complex were indeed equestrian pastoralists, these results would support the idea that transport activity is involved in lateral groove formation, and could implicate the use of a bridle or headgear in late Bronze Age Mongolia. When present, elevated levels of nuchal ossification and medial and lateral nasal remodelling appear to be robust indicators of equine transport, and may be useful for evaluating prehistoric horse use in other archaeological contexts. The compelling pathological signature identified in DSK specimens supports the contention that equine transport and increased mobility played a key role in social transformations in Mongolia and East Asia towards the end of the Bronze Age (Honeychurch et al. 2009; Houle 2009; Wright 2014). Chariots and riding artefacts may be absent from the late Bronze Age archaeological record in Mongolia, but the equine crania analysed here suggest that many DSK horses were heavily exerted (and perhaps bridled). Although these osteological techniques cannot reliably distinguish between chariotry, cart traction or horseback riding, the data imply that the horse was used for transport in the Mongolian Steppe as early as 1300 BC. Models for the spread of equine transport into East Asia 45 may thus have greatly underestimated the role played by steppe peoples from the Mongolian Plateau. Figure 3.11. A) Nuchal ossification score for DSK sample (top), as compared with known groups, d i i g a d idi g data from Bendrey (2008); B) normalised and age-corrected medial groove depth for DSK sample (lower left), as compared to known groups; C) normalised and agecorrected lateral groove depth for DSK and comparative horses (lower right). 46 Figure 3.12. Lateral vs medial groove depth across groups, showing co-occurrence of high values in ridden and DSK samples. 47 CHAPTER 4: RECONSTRUCTING EQUINE BRIDLES IN THE MONGOLIAN BRONZE AGE Journal of Ethnobiology 36(3): 554–570 William Timothy Treal Taylor1*, Tumurbaatar Tuvshinjargal2, and Jamsranjav Bayarsaikhan2 A haeozoologi al e ai s p o ide a ke dataset fo u de sta di g ho se o t ol i Mo golia s Deer Stone-Khirigsuur (DSK) Complex, a late Bronze Age culture dating to circa 1300–700 BC. Although no horse tack has been recovered from DSK contexts, archaeological finds from nearby areas of East and Central Asia suggest that a bridle with a noseband, soft organic bit, and rigid cheekpieces was used by late Bronze Age Mongolian herders. Osteological data from a sample of 25 ritually interred horse crania corroborate these inferences. Deformation to the bridge of the nose on several archaeological specimens suggests that DSK bridles incorporated a noseband, while limited damage to the premolars or diastema is consistent with organic mouthpiece use. A preliminary comparison between archaeological and contemporary horses ridden with known bridle equipment imply that osteological changes to the lateral margin of the premaxilla, present in the DSK sample, might have been produced by a bridle cheekpiece. This study highlights the promise of combining multiple lines of skeletal evidence with other archaeological data to reconstruct ancient equine bridles and tack. Keywords: horses, transport, osteology, Bronze Age, Mongolia 1 Department of Anthropology, University of New Mexico, MSC01-1040, Albuquerque, NM 87131. 2 National Museum of Mongolia. *Corresponding author (wtaylor@unm.edu) 4.1 INTRODUCTION The development of effective horse control revolutionized human societies in ancient Eurasia. As early as 3500 BC or before, domestic horses provided a source of milk, meat, and transport to people living in the steppes of western Central Asia (Outram et al. 2009). In the late third millennium BC, horse-drawn wheeled vehicles were interred in burials belonging to the Ural egio s “i tashta ultu e, a d the iddle of the se o d ille iu , had e o e idesp ead across much of the Eurasian continent (Kelekna 2009b:63). The spread of equine transport stimulated new forms of social organization (Anthony et al. 1991), prompted the expansion of trade networks (Christian 2000), and laid the foundation for new lifeways, such as nomadic horse pastoralism (Kuzmina 2003). The domestic horse did not apparently arrive in the eastern steppes of Mongolia until the late Bronze Age, circa 1300 BC (Hanks 2010:475–76). At this time, some scholars hypothesize that innovations in horse transportation enabled the rapid development of mobile pastoralism in the region (e.g., Beardsley 1953:26). The earliest direct evidence for domestic horses in Mongolia comes from ritual inhumations found near stone monuments of the Deer Stone-Khirigsuur (DSK) Complex circa 48 1300–700 BC (Fitzhugh 2009a; Frohlich et al. 2009). Although mounted horseback riding is not clearly evident in Mongolia before the early first millennium BC (Hanks 2010:476–77; Honeychurch et al. 2009), many DSK horses predate this mark by several centuries (Fitzhugh 2009b). A growing body of evidence suggests that these horses were used for transport (Taylor et al. 2015). However, little is known about how they might have been controlled or bridled, or whether they were used in traction or mounted riding. Characterizing DSK horse use is thus an important step towards understanding the development of nomadic societies in the Eastern Steppe. In this paper, we use historical and ethnographic data in tandem with zooarchaeological evidence to explore DSK bridling and horse control. We describe archaeological horse equipment from late Bronze and early Iron Age contexts in Mongolia, China, and South Siberia, which provide helpful analogs for DSK bridle technology. Next, we summarize the various ways in which halters and bridles may be identified through cranial osteology, including new evidence for deformation to the nasal bones caused by a noseband. Using a sample of well-documented contemporary and archaeological horses, we explore the potential osteological effects of bridle hardware on the premaxilla. Finally, we present results from an osteological study of 25 DSK horse crania, suggesting that DSK bridles incorporated a noseband for communication and braking, a soft organic mouthpiece, and a rigid cheekpiece for turning and lateral control. These initial results highlight the value of cranial osteology in the study of early horse equipment and provide a starting point for reconstructing the development of equine transport in the Eastern Steppe. 4.1.1 Late Bronze Age Archaeology and Early Horse Use in Mongolia Horses were an important component of subsistence and ritual in the DSK Co ple . Dee sto e is the te fo a th opo o phi sta di g sto es, hi h might have been memorials for warriors or particular ancestors (Fitzhugh and Bayarsaikhan 2011), while khirigsuurs are stone mounds that at least sometimes served a mortuary function. These monuments are commonly accompanied by ritually interred horse skulls, buried in smaller stone mounds surrounding deer stones or khirigsuurs (Fitzhugh 2009a). The distribution of DSK sites throughout the Mongolian steppe might suggest they played a role in the initial spread of horses into China (Honeychurch 2015:193-4), where chariots and horses appear in late Shang Dynasty burials circa 1180 BC (Kelekna 2009b:136). Previous archaeozoological studies indicate that DSK horses were consumed for meat (Houle 2010:127) and were likely used for transport (Taylor et al. 2015). In this context, equine skeletal remains from the DSK period hold important clues about early horse use in eastern Eurasia. Prior to the first millennium BC, carts and chariots were an important means of horse transport in other Central Asian cultures. Coercing a horse to be ridden requires overcoming a host of o sta les, i ludi g the ho se s ph si al discomfort and panicked flight response (Dietz 2003:190–91). When hitched to a chariot, the presence of another horse would have had a calming effect, the heavy restraints of draft equipment mitigating many other behavioral issues (Dietz 2003:190). At an earlier stage of domestication, such chariots might have been a more reliable form of transport than riding horseback (Dietz 2003:190; Drews 2004). Bronze objects connected to chariotry have been found in second millennium BC archaeological contexts from the Minusinsk Basin, adjoining Mongolia to the northwest (Wu 2013:35–39; Figure 4.1:2). 49 Although few artifacts were intentionally buried in DSK contexts (Frohlich et al. 2009), rock art carvings of chariots attributed to the late Bronze Age are common in central and western areas of Mongolia (Honeychurch 2015:192–94) and a few vehicles are even depicted on western Mongolian deer stones (e.g., Volkov 2002). This scenario raises the possibility that chariots were known and used by DSK people. Beyond circumstantial evidence for chariots, people of the DSK Complex may also have been among the first in East Asia to use the horse for riding. In East Asia, nomadic groups likely began riding horses before sedentary peoples (Mair 2003:181). Herders in late Bronze Age Mongolia had extensive experience with equine management and seasonal mobility, experience which may have provided the necessary skillset to experiment with methods of horse control Ho e hu h : , . A haeologi al ta k fou d i Mo golia sla u ial sites implicates mounted riding in the ninth century BC (Honeychurch et al. 2009:347), concurrent with later dates for deer stones and khirigsuurs. In short, despite a sparse material record pertaining to horse use, the DSK period encompasses a watershed period in the history of horse control—the emergence of sophisticated mounted riding and equestrian societies in East Asia. The archaeological record of other Bronze and Iron Age cultures in the region can help shed light on how DSK horses were bridled during this important transition. 4.1.2 Ancient East and Central Asian Bridles The record of archaeological horse tack from Siberia, China, and Mongolia suggest that DSK bridles incorporated a cheekpiece. As used here, the term heekpie e efe s to a a of etal or organic material situated against the sides of the ho se s fa e, hi h helps to sta ilize the mouthpiece and ensures its proper positioning in the mouth. Bridles with rigid cheekpieces have been recovered from many late Bronze and early Iron Age contexts in eastern and central Eurasia. In such bridles, when the reins are pulled on one side, the rigid bar would have been compressed against the opposite cheek, coercing the horse to turn in the desired direction Littaue .d. . Ka asuk ultu e sites i “i e ia s Minusinsk Basin have produced three-holed bone cheekpieces that probably date to between the eleventh and ninth centuries BC (Honeychurch 2015:257). Although few late Bronze Age chariot burials from China contain preserved bridle equipment, some have yielded rectangular bronze cheekpieces (Cooke 2000:88–89; Wu 2013:54). In Chinese bridles of the early first millennium BC, elongated cheekpieces of antler and bronze were common (e.g., Wu 2013:13). Rigid cheekpieces are also known from first millennium BC, non-DSK archaeological contexts in Mongolia. At the slab burial site of B-007 in the Egiin Gol Valley, antler cheekpieces (Figure 4.2) were found in association with equine skulls and bridle decorations (Honeychurch et al. 2009:347). Hard antler tines running along each cheek would have been secured to the bridle via straps, attached to the small holes visible at each end. A radiocarbon date from this burial feature places it between circa 940–800 cal yrs BC (Honeychurch 2015:129), coeval with later dates for DSK sites in other areas of Mongolia. These artifacts suggest that the cheekpiece was an important bridle element in East Asia during the late Bronze Age. The archaeological record of Late Bronze and Early Iron Age bridles also raises the possibility that DSK bridles used an organic bit. When preserved wood, bone, or bronze cheekpieces are found in situ without a mouthpiece, the original presence of a bit of perishable material can sometimes be inferred (Drews 2004:84). In China, many second millennium BC horse burials 50 yielded in-place bridle decorations or cheekpieces, but no mouthpiece (e.g., Cooke 2000:88– 89;Wu 2013:54). Bitless bridles are one important possible explanation for this scenario (Dietz 2003), but cheekpieces with a thin and flat central aperture probably once accommodated a soft leather strap or cord bit (Drews 2004:84). Several Karasuk cheekpieces from the Minusinsk Basin have such an opening (see Legrand 2006:857). Organic connecting straps can of course also be used to affix a separate, metal mouthpiece (e.g., Dietz 2006). However, this configuration became common in the first millennium BC (Dietz 2006:158), several centuries later than these Karasuk and early Chinese examples. As a result, it is likely that the scarcity of bits in these archaeological contexts reflects the degradation of organic mouthpieces. Figure 4.1. Khirigsuur and deer stone sites included in the study (filled dots), as they relate to contemporary political boundaries, the Minusinsk Basin in South Siberia, and Anyang, China (Shang Dynasty capital). Archaeological materials from Mongolia indicate that organic bits were used in the region well into the first millennium BCE. At BG-007, the cheekpieces shown in Figure 4.2 were recovered with bridle decorations, but no accompanying mouthpiece (Wright 2006:275). Similarly, four sets of bronze cheekpieces were recovered from Jargalantyn Am Structure 3 in Central Mongolia, a slab burial built from repurposed deer stones (Turbat 2011; Volkov 1990). Each cheekpiece boasts three thin holes. According to the initial investigator, these artifacts were found in situ with remnant leather mouthpieces (Sanjmyatav 1993:34). Although the original faunal materials have been lost, a horse tooth recovered from site backfill dates these artifacts to between 790–542 cal yrs BC (2520 +/- 30 14C BP, Beta #363202). In the context of other archaeological finds already noted, this constitutes compelling evidence that organic bits remained in use in Mongolia until at least eighth century BC. Finally, these archaeological comparisons also suggest that DSK bridles incorporated a noseband. A noseband is a bridle strap that runs transverse across the nose of the animal. If attached to the reins, it places pressure on sensitive facial tissues, prompting the horse to 51 Figure 4.2. Interior and exterior views of antler tine bridle cheekpieces from the site of B-007 in Egiin Gol Valley, Mongolia, dated to circa 940–800 BC (Honeychurch 2015:129). Drawing by Dr. Joshua Wright, reprinted with permission. A mouthpiece would likely have been hemmed in between leather straps attached to the two holes, connecting to the bridle headstall. instinctively lower its head and slow (Dietz 2003:192). As a means of braking, a key challenge of early horse transport (Drews 2004:88), the noseband was a major improvement over more rudimentary systems, such as the nose-ring (Littauer 1969). Although nosebands were phased out of some bridle systems after the invention of the jointed metal snaffle (Drews 2004:89), soft organic mouthpieces probably did not produce enough pressure for effective braking on their own (Drews 2004:83–86). In Mongolia and other parts of Central Asia, nosebands have been used to control horses throughout antiquity (Brownrigg 2006:168). As a result, it is reasonable to hypothesize that DSK bridles probably incorporated a similar feature. 4.2 RECONSTRUCTING ANCIENT BRIDLES THROUGH EQUINE OSTEOLOGY This review of archaeological tack raises the possibility that DSK bridles incorporated three key components—a noseband, an organic bit, and a rigid cheekpiece. Because of historical variability in bridle design and the absence of organic components due to poor preservation, bridle reconstructions based on incomplete artifacts sometimes yield controversial results (Dietz 2003:193–97; Drews 2004:15–19). A clear understanding of DSK horse control thus requires consideration of other forms of evidence. One promising line of inquiry comes from faunal remains. Recent studies (e.g., Anthony and Brown 1998; Anthony et al. 2006; Bartosiewicz 2014:135; Bendrey 2007, 2008; Taylor et al. 2015) indicate that various osteological changes to the skull and mandible accompany the use of the horse for transport. Here, we outline several changes to the equine cranium that help identify components of DSK bridle technology. 52 4.2.1 Noseband Use and Nasal Remodeling Defo atio to the idge of a ho se s ose o asio all a o pa ies the use of a halter or bridle noseband. When present, such deformation can be used to identify this bridle component in the archaeological record. In the horse, very little tissue covers the bone at the bridge of the nose, facilitating osteological changes to this area of the skull. As one example, tight haltering of developing juvenile animals can cause the bone to deform as the horse matures (Scott Bender, personal communication, February 12, 2015). An extreme example is illustrated here by the skull of a horse who apparently grew to maturity while wearing an undersized halter (Figure 4.3, left . As the a i al s head grew larger, the halter placed consistent pressure on the nasal bones, resulting in grotesque deformation. In less dramatic fashion, the skulls of constantly bridled or haltered adult animals may also deform under pressures from regular use (Bartosiewicz 2014; Takacs 1985). For example, we discovered a pronounced nasal depression in the skull of several adult male Mongolian riding horses (Figure 4.3, center). Unlike most western bridles, the traditional Mongolian bridle still uses a noseband that is directly connected to the reins (Figure 4.3, right). In cases of chronic use, this configuration alone might be enough to prompt bone remodeling. Nasal divots such as those described above are also common in a group of modern horses from the Altai of western China, who are controlled with a bridle nearly identical to its Mongolian counterpart (Figure 4.4, top; see also Jenkins 2014:77). The Kazakh horseman who owns the animals, Mr. Norbek, suggests that nasal remodeling observed in his horses may be a result of traction work and jugen cart racing (Nils Larsen, personal communication, January 29, 2015). When lo g ei s a e looped th ough te ets o othe ha ess pa ts hi h eak the line of the reins (Figure 4.4, bottom), this likely increases leverage and magnifies the d i e s o igi al pressure (Littauer 1969:290). Paired with a direct attachment between reins and noseband, chronic traction work with long reins is thus one possible explanation for the Altai horse patterns. Whether nasal deformation relates to chronic bridling, traction work, or other factors, in all cases this osteological feature provides clear evidence of noseband or halter use. Figure 4.3. Left: adult horse with facial bones badly deformed around undersized halter, found in Wyoming. Photo courtesy of Dr. Danny Walker. Center: nasal deformation in a contemporary Mongolian riding horse. Right: diagram showing direct connection between reins and noseband in Mongolian bridles. 53 4.2.2 Bit/Mouthpie e Use a d Bit Wea Dental and lower jaw anomalies can shed light on bit/mouthpiece use in antiquity (Bendrey 2007; Outram et al. 2009). A unique pattern of dental wear, o o l efe ed to as it ea , o sists of ha ges to the lo e se o d premolar from interaction between tooth and mouthpiece (Anthony and Brown 1998, 2003; Anthony et al. 2006). According to Anthony and Brown (1998; 2003), riding horses with a mouthpiece regularly produces measurable beveling to the occlusal surface of the lower P2, and wear of more than 3 mm in magnitude can be taken as evidence of bit use. Unfortunately, without careful control over factors such as malocclusion and abrasion, these occlusal bevels can be an unreliable measure of human influence (Bendrey 2007; Olsen 2006b). Nonetheless, other changes to the second premolar, including anterior enamel exposures on the lower P2 as well as damage and new bone formation on the diastema, may also be useful for the identification of bitted horses (Bendrey 2007). 4.2.3 Cheekpieces and Premaxillary Remodeling Recent research links two separate changes to the equine premaxilla with horse transport, and may help to identify the use of bridle cheekpieces. The first feature, a groove forming along the medial aspect of the premaxilla, is probably caused by hypertrophy of the lateralis nasi muscle and its accessory cartilage, which are involved in nasal dilation (Perez and Martin 2001; Vanderwegen and Simoens 2002; Figure 4. :A . ‘efe ed to he eafte as edial e odeli g, this groove appears to be more severe in captive and ridden domestic animals, and is plausibly linked to heavy breathing from stress or exertion during transport (Taylor et al. 2015:863–66). The second feature occurs along the lateral aspect of the same bone (Figure 4.5:B). Referred to he e as late al e odeli g, this g oo e is asso iated ith a internal nasal branch of the infraorbital nerve (Perez and Martin 2001). This feature could also be developmentally related to heavy exertion, wherein the rigid lateralis nasi on a heavily worked horse presses the nerve against the premaxilla and causes remodeling to avoid nerve compression (Perez and Martin 2001). Unlike medial remodeling, however, lateral remodeling appears inconsistently, even among some extensively ridden animals (Taylor et al. 2015:866), and is often highly asymmetric in specimens we observed. One possible explanation for this pattern is that lateral remodeling is exacerbated by bridle equipment. The internal nasal branch of the infraorbital nerve is situated near the facial exterior, where it lies in close proximity to the margin of the premaxilla (Figure 4.5, B). For those horses bridled with a hard cheekpiece, chronic pressure or irritation of this area could exacerbate remodeling of the premaxilla to prevent nerve compression. If so, the presence of this feature would be especially valuable for the reconstruction of ancient horse control technology. Premaxilla morphology in a small sample of contemporary and archaeological horses, ridden with documented equipment, supports that these features of bone remodeling relate to use of bridle cheekpieces. If cheekpieces are involved in lateral remodeling, this feature should be limited or absent from unridden horses, and reduced in horses ridden with bridles that use less exterior pressure. The loose-ring snaffle bit, for example, relies primarily on pressure at the 54 corners of the mouth and unless very large connecting rings are used, places minimal hardware in the relevant areas of the cheek (Figure 4.6, bridle A). Five historical American racehorse specimens, most of which were definitively ridden with a loose-ring snaffle, exhibit limited lateral remodeling depths (0.6 mm or less, Figure 4.7, top). We also observed minimal lateral remodeling in definitively unbridled wild equids. Eleven paleontological specimens (late Pleistocene Equus scotti, 30–75 ka; Pliocene Equus simplicidens, 3.5 Ma: and Pliocene Equus stenonis, 2.8 Ma) all lacked measurable lateral remodeling. Only shallow lateral remodeling was observed in three wild Equus hemionus skulls from Mongolia and this feature was uncommon among wild and feral equids examined in a prior study (Taylor et al. 2015; Figure 4.7, top). Figure 4.4. Top: nasal depressions on working horses in the Altai of Xinjiang, China.Bottom: Altai horses in traction work, showing long reins looped through a body harness and attached to bridle headstall. Photos courtesy of Nils Larsen, Altai Skis. 55 Figure 4.5. Diagram showing remodeling to the medial (A) and lateral (B) aspects of the equine premaxilla. Illustration by Rebecca Tuccillo. Figure 4.6. Left: equipment of horses analyzed for premaxillary remodeling. A) simple loose ring snaffle, B) Turkic-era snaffle bit with S-shaped iron cheekpiece, C) Pazyryk snaffle with wooden cheekpiece, and D) Weymouth bridle under rein pressure. Right: US Cavalry curb bit similar to equipment used on Kidron s We outh idle, ith black line indicating the path of the infraorbital nerve in area of lateral remodeling. 56 Figure 4.7. Top: lateral vs. medial remodeling depth across a sample of wild extant and fossil equids, feral domestic horses, captive E. przewalskii, and ridden horses with documented equipment (A simple loose ring snaffle, B and C- archaeological snaffle with rigid cheekpiece, DWeymouth or double bridle). Bottom: lateral vs. medial remodeling depth for DSK horses (black) as compared to a sample of wild extant and fossil equids, feral domestic horses, captive E. przewalskii, and ridden horses with documented equipment (A- simple loose ring snaffle, B and C- archaeological snaffle with rigid cheekpiece, D- Weymouth or double bridle). 57 I o t ast, idles usi g full heeks pla e o e p essu e o the sides of the fa e D ape et al. 2014:427) and horses ridden with such bridles should exhibit heavier premaxillary remodeling. Historical bridles used by Pazyryk (circa 600–300 BC) and Turkic Khaganate (circa 600–700 AD) cultures used a large rigid cheekpiece, which ran perpendicular to the mouth (Figure 4.6, bridles B and C). Two Pazyryk and Turkic horses from western Mongolian burials exhibited comparatively deeper lateral remodeling (0.7 and 1 mm in depth, respectively). “i ila l , t o t e tieth e tu spe i e s i ludi g Ge e al Joh Pe shi g s a ho se Kidron) were ridden with a Weymouth bridle, a style of tack that uses both a snaffle and a curb bit simultaneously. The curb bit amplifies rein pressure using an arm or shank attached to the reins Ma Fa la d : . This it s p i a a tio is o the oof of the ho se s outh, ut the curb shank extends dorsally, where it it may pressure the infraorbital nerve directly, or irritate the face and cause swelling and compression (Figure 4.6, bridle D). These two specimens exhibited marked lateral remodeling (Figure 4.7, top).A larger comparative sample is necessary to confirm the validity of these patterns and assess the impact of potentially confounding a ia les, su h as a ho se s age, ho se a ship o idi g st le, a d o k histo . Ho e e , these preliminary data raise the possibility that a bridle cheekpiece exacerbates lateral remodeling of the premaxilla. As such, this feature is useful to consider in tandem with other lines of osteological evidence to evaluate late Bronze Age equine bridling in Mongolia (summarized in Table 4.1). 4.3 METHODS To test the hypothesis that DSK bridles incorporated a noseband, rigid cheekpiece, and soft organic bit, we analyzed a sample of 25 horse crania for evidence of nasal depression, bit wear, and premaxillary remodeling. Although the majority of these skulls were badly fragmented, two skulls had complete preservation of the upper nasal area.We scanned these at a resolution of 2000 DPI with a NextEngine 3D Scanner, visually inspecting them for recesses related to noseband use. To identify whether bitting damage affected DSK horses, we assessed lower premolar beveling following the protocol outlined by Anthony and Brown (2003). Excluding deciduous or broken specimens, teeth from 13 individual horses remained for analysis. For each lower P2, we measured bevel depth at the anterior-most border of the tooth perpendicular to the occlusal surface, using scale profile photographs in the open-source measurement program, ImageJ. When measurable bevels could be identified, we compared tooth row morphology of both upper and lower jaws to identify malocclusion-related causes. For each lower P2, we also analyzed the anterior morphology for exposed cementum or enamel following Bendrey (2007). In cases of exposed enamel, we sought to identify parallel-sided wear exposures that might be indicative of bit use and compared anterior exposures with the lingual sides of the tooth to rule out natural tooth wear. For DSK specimens with sufficient preservation (n = 8), we scored diastema bone changes according to the categorical ranking system provided by Bendrey (2007). Finally, we compared previously published digital measurements of nine DSK premaxillary fragments (Taylor et al. 2015) with data from the documented specimens outlined above in order to explore implications for cheekpiece use. 58 Feature Anatomical mechanism Equipment implications Premolar beveling Contact between mouthpiece and lower premolars OR Dentistry/malocclusion Presence of a mouthpiece OR None Diastema bone formation Contact between mouthpiece and bars of the mouth Presence of a mouthpiece Nasal bone depression Medial remodeling to the premaxilla Lateral modeling to the premaxilla Downward pressure on the nasal bones during development OR chronic work Hypertrophy of lateralis nasi and accessory cartilage (due to heavy exertion and stress) Developmentally related to medial remodeling OR Lateral pressure/irritation of the face near point of contact between prexaxilla and branch of infraorbital nerve Presence of a bridle noseband or tight halter None hypothesized Unknown, possibly exacerbated by chronic use of a rigid cheekpiece Table 4.1. Osteological features of the skull and their potential significance for equine harness equipment. 4.4 RESULTS One of two horses with sufficient preservation for morphological study provided unequivocal evidence of a bridle or halter noseband (Figure 4.8, left). This specimen, an elderly animal of indeterminate sex from the site of Khushuutiin Gol, in northern Mongolia, was radiocarbon dated to 1224–980 cal yrs BC (calibrated usi g I tCAL Fitzhugh . The a i al s skull exhibits a deep recess to the bridge of the nose, situated above the third premolar. This depression is similar to pathologies identified by Takacs (1985) in sixteenth century AD specimens from Hungary, as well as the animals from contemporary Xinjiang and Mongolia discussed above. While the second complete horse exhibited no deformation, a third partially complete specimen, a juvenile between 2–2.5 years in age from the site of Tsatstain Khushuu exhibits thinning and possible deformation to the nasal bones (Figure 4.8, right). Unfortunately, taphonomic damage reduces confidence in this assessment. Nonetheless, at least one specimen provides compelling evidence that DSK horses were controlled with a noseband. Diastema and premolar form revealed no diagnostic evidence of bit wear in the DSK sample. Among adult lower P2s (n = 13), eight lacked a measurable bevel and only one specimen produced a value greater than 3 mm (Supplementary online material, Appendix B). When fitted with the opposing upper jaw, this feature was clearly attributable to malocclusion, caused by a badly impacted upper molar. Anterior premolar wear and diastema bone formation were also nearly absent from the DSK sample. Ten DSK lower P2 specimens show some form of anterior cementum or enamel loss. However, most instances of apparent enamel wear also affected other portions of the tooth beyond the anterior edge, and none exhibited the parallel sides that 59 are characteristic of metal bit wear. As a result, it remains unclear whether this anterior wear is entirely natural or relates to the use of a softer bit. In any case, none of the seven DSK specimens with diastema preservation demonstrated more than faint bony changes to the diastema, falling within the range of variation observed by Bendrey (2007) in unbitted horses. Several of the DSK horses displayed lateral remodeling depths similar to those observed in horses bridled with a rigid cheekpiece. Four of the DSK specimens with measurable premaxillae had no lateral remodeling (that is, levels comparable to feral horses and other wild equids). However, five DSK premaxillary fragments had measurable external grooves of greater than 0.5 mm (Figure 4.7, bottom). In particular, the Khushuutiin Gol specimen, previously noted for its noseband depression, had pronounced lateral remodeling of nearly 2 mm in depth. Two other DSK specimens fell between 0.7 and 0.9 mm, measurements comparable to those observed in the Pazyryk, Turkic, and modern Weymouth bridle specimens. 4.5 DISCUSSION Remodeling of the nasal bones in at least one DSK horse skull corroborates the inference of a noseband or halter, with several possible modes of formation. In contemporary Mongolia, young foals are often haltered and tied for long periods during the summer, when the mares are milked to produce airag(fermented horse milk; Figure 9). Young horses that will be used for riding also begin their training at around one year of age (Enkhtuvshin and Tumurjav 2011:173– 74). Either of these practices could produce chronic pressure to the bridge of the nose, at an age when the nasal bones would be developing and thus particularly sensitive to deformation. The extreme deformation on the specimen from Khushuutiin Gol probably indicates that DSK bridles used a direct connection between reins and noseband, as seen in the Altai and contemporary Mongolian examples. A compelling additional possibility is that traction work using long reins (i.e., chariotry or carts) increased the pressure on this point of the skull. Figure 4.8. 3D model showing facial deformation from noseband use in a specimen from the site of Khushuutiin Gol (left, center) radiocarbon dated to 2910 +/- 40 14C BP (1224–980 cal yrs BC), and possible deformation in a young horse from Tsatstain Khushuu (right, 2920 +/- 40 14C BP). 60 Figure 4.9. Young foal haltered and tied to a rope line with other foals during summer milking season, Bayankhongor province, Mongolia. The absence of appreciable diastema or P2 damage to sample DSK specimens may reflect organic bit use. While experimental efforts have come to conflicting conclusions about the effect of softer organic bits on equine dentition (Anthony et al. 2006; Brownrigg 2006), a less abrasive mouthpiece of leather or hemp could have a reduced effect on the diastema and lower P2 (Bendrey 2007:1048). It should be noted that the absence of bitting damage in our sample does not necessarily rule out use of a metal bit; factors such as bridle design and style of horsemanship probably alter the skeletal impact of even hard metal mouthpieces (Bendrey et al. 2013:98). DSK horsemen could also have used a bitless bridle, a technology that remained common throughout the Bronze Age (Dietz 2003). However, as archaeological tack demonstrates the use of organic bits well into the first millennium BC in other Mongolian archaeological contexts, we suggest that the minimal bitting damage in the DSK sample reflects the use of leather, wood, or bone mouthpieces. 61 If future study validates the link between osteological changes to the premaxilla and lateral bridle pressure, our sample would also support the presence of a hard cheekpiece in DSK bridles. Although Weymouth or curb-bit bridles are not known from the archaeological record of eastern Eurasia in prehistory, this style of bridle pla es o e ha d a e i the a i al s heek region than many other configurations, and thus could cause more regular interaction with the infraorbital nerve near the premaxillary margin. Grooving to the lateral margin of the premaxilla in DSK specimens is similar to contemporary horses ridden with a Weymouth bridle and archaeological specimens controlled with a rigid bar cheekpiece. Bars of metal, bone, or wood situated alo g the sides of the ho se s face were an important element of Mongolian bridles for centuries following the DSK period and would have enabled DSK riders or charioteers to turn the horse effectively. 4.6 CONCLUSION Several key challenges complicate the osteological study of equine bridles. Foremost among these is the difficulty of acquiring specimens with suitably detailed histories, as well as the complex and continuous nature of osteological remodeling processes in the equine skull. However, our results indicate that osteology can be used to identify particular bridle components when skulls are preserved in the archaeological record, even in the absence of preserved tack. Drawing on both osteology and analogy, it is likely that DSK bridles incorporated a noseband, organic bit, and hard cheekpieces. This style of bridle could have been used to drive chariots/carts or for experiments in early horseback riding (Bokovenko 2000; Honeychurch 2015:128). In either case, DSK bridle technology would have been critically important for ancient nomadic activities, and may have facilitated the development and spread of mobile pastoralism in the Eastern Steppe. With an improved understanding of osteological formation processes, the cranial changes identified here may one day be useful for evaluating temporal patterns in bridle technology across the late Bronze Age. This approach will help to clarify the changing role of the horse in ancient societies, as well as the ways Mongolian nomads may have affected the transition from chariotry to mounted horseback riding in East Asia. 62 CHAPTER 5: HORSEBACK RIDING, ASYMMETRY, AND ANTHROPOGENIC CHANGES TO THE EQUINE SKULL: EVIDENCE FOR MOUNTED RIDING IN MONGOLIA “ LATE B‘ONZE AGE In review, Oxbow Books – Proceedings of the 6th Animal Paleopathology Working Group, International Council for Archaeozoology, Budapest, Hungary William Taylor1 and Tumurbaatar Tuvshinjargal2 A primary obstacle facing the study of early horse transport is the challenge of identifying ridden horses in the archaeological record. Although changes to the equine skull and dentition may help identify animals that were bridled and used for transport in the archaeofaunal record, these features are insufficient to distinguish riding mounts from animals used to pull vehicles. This paper presents evidence that asymmetry in cranial deformations may be anthropogenic, and useful for identifying ridden horses on the basis of skeletal remains. Many contemporary nomadic herders in Mongolia use a bridle which directly pressures the skull in several places, riding with the reins typically held in the left hand. Preliminary analysis of equine crania from modern and historical contexts suggests that this riding style may leave an asymmetric osteological signature, including deformation and thinning of the left nasal bone and remodelling of the right margins of the premaxilla. A small sample of late Bronze Age horses from Mongolian archaeological contexts also displayed this asymmetric deformation to the left nasal bone and right premaxilla, a pattern which may help understand the early chronology of equine transport in the region. This research suggests that consideration of cranial asymmetry may help to distinguish riding mounts in the archaeological record, and supports other evidence that horseback riding was established in Mongolia by circa 1200 BCE. Keywords: horseback riding, osteology, asymmetry, Late Bronze Age, Mongolia 5.1. INTRODUCTION Although horses were domesticated in Central Asia as early as the fourth millennium BCE (Olsen 2006; Outram et al. 2009), the timing of the emergence of mounted horseback riding is less well understood. Horses were used to pull chariots across much of the contintental interior during the second millennium BCE (Drews 2004:50-51; Kelekna 2009b:63), and some scholars suggest that horses must also have been ridden by Central Asian nomads at this time (Anthony et al. 1991; Anthony and Brown 2003). However, the first unequivocal historical records of competent mounted riding appear to date to the early first millennium BCE (Argent 2011:31; Drews 2004:66). Some of the earliest domestic horses known from Mongolia come from an archaeological culture known as the Deer Stone-Khirigsuur (DSK) complex (Fitzhugh 2009:189; Honeychurch 2015:121). Named for the carved stone megaliths (deer stones) and burials (khirigsuurs) that were constructed across much of Mongolia and eastern Central Asia during the late Bronze Age (ca. 1200-700 BCE), this culture has been linked with a mobile herding lifestyle and the emergence of social inequality in the region (Houle 2009:372). Many DSK ritual sites are characterized by large numbers of small stone mounds, containing the heads and hooves of sacrificed horses. Despite their ubiquity, however, it is unclear how these DSK horses would 63 have been used. Deformation to the nasal bones of both old and young horses recovered from such contexts suggests that many were bridled or haltered (Taylor et al. 2016), while changes to the premaxilla linked to heavy exertion suggest they were also used for transport (Taylor et al. 2015). A few deer stones depict chariot images, and many researchers assume that DSK people used horses to pull chariots (e.g. Erdene-Ochir and Khyadkov 2016:23-30). However, it is likely that nomadic peoples began mounted riding prior to its widespread emergence during the early first millennium BCE (Mair 2003:181). The rapid spread of horse ritual features across Mongolia ca. 1200 BCE hints at a major, horse-related social transformation that is perhaps linked with the adoption of horseback riding (Taylor et al. in review). Other than the faunal remains of sacrificed horses, however, no grave goods and few artifacts have been recovered from DSK sites with which to evaluate directly how these horses were used by late Bronze Age people. 5.1.1 Identifying Mongolian riding: an ethnoarchaeological approach One potential avenue for distinguishing horses used to pull chariots or carts from those used for mounted horseback riding may be found in the riding style employed by many nomadic people in Central Asia. Just as in some other riding traditions where the right hand must be used often for other purposes (such as historic American cavalry, or competition riding in the western United States), Mongolian herders typically ride with the reins held in the left hand (Figure 5.1). For the right-handed rider, this approach enables the more dexterous hand to hold other tools such as a whip (Mongolian: tashuur), lasso, or lasso pole. Mongolian bridles traditionally incorporate a noseband directly attached to the reins. When under tension, this noseband directly pressures the bridge of the nose. The bit used in most Mongol bridles is a kind of jointed snaffle, with somewhat unique elongated, curved canons that protrude far beyond the margins of the mouth, and focus pressure on the corners of the mouth. These bridles are used in conjunction with many different styles of cheekpiece, including small rings, large rings, and rings with vertical bar extensions. When riding at a gallop, Mongolian riders stand in the stirrups, using the reins for stability (Figure 5.2). Figure 5.1. A group of Mongolian riders watch a horse race in Khuvsgul province, northern Mongolia. Image shows the ubiquity of the left-handed riding posture. 64 This combination of tack and riding style might be expected to produce osteological deformations to the equine nasal bones. Deformation caused by bridling or chronic halter use has been previously recognized on archaeological horse specimens (e.g. Bartosiewicz 2014:132; Takács 1985), and modern Central Asian horses (Taylor et al. 2016). Many factors likely influence the formation of this feature, including the age of the animal when bridled, type and fit of the harness, and the frequency and type of transport for which the animal is used (Taylor et al. 2016). The constant presence of an overtight halter may be one especially important factor influencing deformation (Takács 1985:312). However, unlike a tight halter, the left-handed riding technique employed by Mongolian horsemen should produce more frequent or pronounced pressure on the left side of the nose while under rein pressure. Consequently, this fact should be reflected in asymmetric remodeling of modern Mongolian horse skulls. Bridle cheekpieces may also produce changes to the skulls of horses ridden by contemporary Mongolian nomads. Equine crania we observed in museum collections often display a groove to the lateral aspect of the premaxilla (Figure 5.3), which appears linked to use in transport (Taylor et al. 2015). This groove is associated with a nasal branch of the infraorbital nerve, which runs close the margin of the premaxilla. Bone remodeling of the premaxilla at thie location probably occurs to protect the nerve and blood vessels from compression (Perez and Martin 2001:358). The source of this compression is unclear. Although this feature was initially linked to the development of rigid nasal muscles in heavily trained animals, a recent comparative study raises the possibility that grooving is exacerbated by chronic pressure or irritation caused by bridle equipment (Taylor et al. 2016). According to our ethnographic informants, the contemporary Mongolian zuuzai or cheekpiece functions not only as a turning aid, but also to prevent the bit from passing entirely through the mouth when the reins are pulled. Experimental work shows that when a single rein is pulled on a snaffle bridle, the opposing cheekpiece is moved medially, where it contacts the side of the mouth and face, opposite the pulled rein (Clayton and Lee 1984). With the reins held in the left hand, the cheekpiece is thus more likely to contact the right facial exterior during riding activities unless an appropriate countering force is consistently applied. If premaxillary remodeling is indeed reliably linked to some form of interaction between the cheekpiece and the infraorbital nerve, contemporary and ancient horses ridden with a Mongolian bridle or similar configuration should display more severe remodeling on the right premaxilla. Lastly, the use of a metal bit can cause oral damage which may help identify left-handed horse riding. One of the most well-known anthropogenic features connected with horse transport is it ea , o e eli g of the se o d lo e p e ola aused etal it use A tho et al. 1991; Anthony and Brown 2003; Anthony and Brown 2006). Because natural problems with dental occlusion can also alter the second premolar, the validity of bit wear as evidence for horseback riding has been questioned over concerns about equifinality (Bendrey 2007; Levine 1999; Olsen 2006). Some scholars have even doubted the very premise that a metal bit can interact with the premolars long enough to cause recognizable alterations (e.g. Sasada 2013). Nonetheless, bit chewing has been observed under clinical examination (Clayton and Lee 1984; Manfredi et al. 2010). Some contemporary equine dentists link this behavior with specific changes to the teeth - displacement, remodeling of the alveoli, formation of a smooth, domeshape to the upper premolars, and a flat, smooth ramp on the lower premolars (e.g. Johnson and Porter 2006). Horses that chew the bit sometimes exhibit a preference for one side of the 65 mouth, but they more usually favor both sides equally (Johnson and Porter 2006). Consequently, any occlusal beveling caused by bit chewing behavior would be unlikely to exhibit asymmetry. Ho e e , ho se a k idi g a also ause i ide tal o ta t ith the ho se s p e ola s du i g use that could produce asymmetric dental damage. Single reining produces a caudal displacement of the bit, bringing it closer to the premolar margins on the side being pulled (Clayton and Lee 1984). Unlike most Euro-American bridle systems, the Mongolian bit protrudes sig ifi a tl e o d the sides of the ho se s outh – enabling looser movement – and during ethnographic observations we often observed the bit being pulled back to a point of contact with the lower second premolars. Such contact can damage the anterior margin of the lower second premolar in a characteristic fashion. In contrast to occlusal damage, this anterior wear occurs only in very rare cases among unbitted animals, making it a more reliable index of bit use (Bendrey 2007:1041,1049). As a result, horses ridden with a Mongolian bridle and riding style might exhibit an increased frequency of dental changes to the left side of the mouth. Figure 5.2. Mongolian herder riding left-handed, leaning to one side and stabilizing himself with the reins, with visible pressure the left nasal area. Herder using lasso pole visible in background. Photo by Orsoo Bayarsaikhan photography. 66 Figure 5.3. Asymmetric lateral remodelling to the premaxilla caused by remodeling of the bone in the area of the infraorbital nerve, shown on an archaeological specimen from Mongolia. Insofar as they hold for modern animals, these predictions should also hold for ridden horses from archaeological contexts. Finds of ancient of horse tack indicate that ancient Mongolian bridles also used a direct noseband attachment, with cheekpieces as turning aids (Taylor et al. 2016). Images and artifacts also suggest that the left-handed riding style may have great antiquity in eastern Central Asia. Through to the present day, depictions of mounted nomadic warriors from the early and late middle Ages (e.g. Yatsenko 2015) nearly always show riders holding the reins in the left hand (Figure 5.4). Although not directly related to riding posture, well-p ese ed idles f o the Paz k ultu e of the fi st ille iu BCE so eti es had a lead ei o the a i al s left side A ge t : , suggesting that riders handled and mounted horses from the left. Prior to the invention of stirrups in the first millennium CE, late Bronze and 67 early Iron Age riders may have also used the reins to stabilize and balance themselves to an even greater degree. Figure 5.4. Statue depicting a warrior from the Great Mongol Empire, 13th-14th centuries CE, outside the Parliament building in the capital city of Ulaanbaatar. Thus, if asymmetric cranial changes characterize contemporary Mongolian horses, similar patterns should also be observed on archaeological horses from the first millennium BCE and after. Would such a finding indicate prehistoric riding, as opposed to chariot driving? It is possible that ancient chariot horses might also have experienced asymmetric rein pressures. Carvings on Mongolian rock panels (Honeychurch 2015:121) and deer stones themselves (Nyambat and Odbaatar 2010:63-64) indicate that late Bronze Age Mongolian people used chariots. These light, two-horse carts would have been controlled by a single driver, with separate sets of reins for each animal on the left and right. If a horse was kept in the same team 68 position over the course of its lifetime, it might conceivably experience unequal chronic tension from a single direction, despite never having been used for mounted riding. In the aggregate, though, this scenario should produce roughly equal numbers of horses used on on either the left or the right, rather than the consistent asymmetry anticipated from left-handed horseback riding. A second, more problematic possibility is that the behavior of the chariot driver could also result more regular pressure from the left direction. Based carvings with visible reins, it seems that early Mongolian chariots were at least sometimes controlled by reins running freely from the bridle to the driver. If a charioteer favored left-sided turns for tactical reasons or due to natural instincts, rein activity could conceivably produce chronic asymmetric pressures favoring the a i al s left side. Futu e study will be necessary to explore this possibility in depth. However, petroglyphs consistently depict reins held by the charioteer with both hands or tied to the center of the driving box, and sometimes apparently running through terrets or guides attached to the central draught pole (Figure 5.5). Subsequently, it seems that pressure asymmetry should be less frequent and less pronounced among chariot animals. In short, the presence of marked asymmetry in archaeological horses would strongly support the hypothesis that said horses were ridden, rather than driven. Figure 5.5. Petroglyphs from Tsagaan Gol in western Mongolia, showing driver holding two sets of reins, and reins running through a terret affixed to the pole (right). Photographs: Gary Tepfer. Copyright: Mongolian Altai Inventory Collection, University of Oregon. Reprinted here with permission. 5.2. MATERIALS AND METHODS To test the hypothesis that assymetry is related to left-handed riding, we conducted osteological analysis of modern and archaeological horse skulls from Mongolia, comparing these specimens to previously analyzed control samples of wild and domestic horses from museum collections (Table 1). As Mongolian horses begin training for riding between one and two years of age (Enktuvshin and Tumurjav 2011:173- , a d e ause a ho se s pe a e t p e ola s e e ge at around 2.5 years, we excluded all animals younger than three years (estimated by dental eruption following Evans et al. [2006]) from our analysis. 69 5.2.1 Contemporary Mongolian horses We collected a sample of 15 adult horse skulls via surface collection in the Mongolian countryside in several regions of central and western Mongolia, including Tuv, Uvurkhangai, and Gobi-Altai provinces. Because similar bridle styles are used by herders across the region, these crania should effectively characterize deformation patterns among animals ridden with a Mongolian bridle. Age and sex estimates of these (and all subsequent) sample specimens are provided in Appendix IV. 5.2.2 Iron and Middle Age Mongolian horses We analyzed the cranial remains of 8 adult archaeological horses buried with riding tack, and dating to the era of mounted horseback riding cultures – the Iron Age and early Middle Ages. These included one Pazyryk (ca. 600-200 BCE) horse from western Mongolia, two Xiongnu (ca. 200 BCE-100 CE) specimens from north-central Mongolia, one Xianbei horse from Orkhon province in central Mongolia (ca. 150-250 CE), three horses from the time of the Turkic Khaganate in western and central Mongolia (ca. 6th-8th centuries CE), and one male horse from the Khitan Period (10th century CE). Of these, two consisted of only a lower mandible, and lacked relevant portions of the cranium (nasal bones or premaxilla). 5.2.3 Contemporary American and Przewalski horses We also studied a sample of previously collected, high resolution 3D scans of 12 domestic American racehorses, farm horses, and military horses from American museum collections (Appendix IV). Some of these sample specimens (the war horse Kidron ridden by John Pershing, the racehorses Lexington, Hanover, Sysonby, Haleb, and the competition horse Indraff) had photographic documentation of tack. However, none of the bridles we identified had a link between the noseband or halter and the reins, making them unlikely to produce pronounced deformation of the nasal bones or left-biased asymmetry. Moreover, although some of these animals were controlled with rigid cheekpieces, the primary effect of these documented bits (such as a curb or snaffle) is on the palate or bars of the mouth. Consequently, these animals should exhibit a greater degree of bilateral symmetry in cranial deformations linked to human activity. We compared these to a group of 13 adult animals which had never been bridled or ridden, including six feral domestic horse skulls – museum specimens recovered from areas occupied by feral herds on Assateague island in Virginia/Maryland, and northwestern New Mexico – along ith se e P ze alski s ho se skulls. 5.2.4 Bronze Age Mongolian horses Finally, we compared our compiled data to a large sample of 46 adult specimens recovered from individual horse burial features at deer stone and khirigsuur sites across Mongolia. Twelve of these horses had sufficient preservation to assess premaxillary morphology, while and only two skulls had sufficient preservation to assess the presence of nasal remodeling (described in Taylor et al. 2016:563). 70 5.2.5 Data collection protocols 5.2.5.1 Nasal and premaxillary remodeling Using a NextEngine3D desktop laser scanner, we created a digital 3D model of each specimen at a resolution of 2000 DPI. For contemporary American domestic and feral animals, as well as P ze alski s ho ses, e used p e iousl olle ted D data. We used these odels to easu e the maximum depth of premaxillary and nasal remodeling on both the left and the right aspect of the skull using open-source measurement software (GOM Inspect). Specimens Contemporary Mongolian horses Iron and Middle Age Mongolian horses Contemporary American domestic horses Contemporary American feral horses Contemporary Przewalski horses Bronze Age Mongolian horses Number Examined With cranial data With dentition data 15 15 6 8 6 8 12 12 --- 7 7 --- 6 6 --- 48 13 48 Table 5.1. Samples used in this study, along with number of specimens analyzed for cranial deformations and oral bitting damage. 5.2.5.2 Dental pathologies linked to bitting We measured the beveling to the occlusal surface of the lower premolars on all specimens except the contemporary American and Przewalski horses, for which only previously collected 3D models of the upper crania were available. We followed the protocol outlined by Anthony and Brown (2003) and used Mitutoyo digital calipers. For specimens exhibiting a measurable premolar bevel, we refit the skull with the lower jaw to identify cases caused by malocclusion. We also e o ded the p ese e o a se e of the G ea es effe t , he ei the e a el a d cementum wear naturally at different rates due to differential composition and hardness. When the Greaves effect is absent – meaning that the enamel and cementum have worn evenly and flat, and the jaw shows few signs of malocclusion when refit – an occlusal premolar bevel may be indicative of bit wear (Olsen 2006:100-101). For each horse, we recorded instances of parallel-sided enamel exposure to the anterior P2 margin (Bendrey 2007), along with cases of non-diagnostic enamel exposure, premortem tooth fractures, alveolar remodeling, and other abnormalities for both upper and lower premolars. 71 Figure 5.6. The nasal bones of a horse from Uvurkhangai province in central Mongolia, showing p o ou ed tapho o i eathe i g to the a i al s left side i the a ea of asal defo atio . Figure 5.7. Asymmetric deformation to the nasal bones on a mummified horse dating to the Middle Ages from Ulaan-Uneet (left), and similar feature on a late Bronze Age horse from the site of Khushuutiin Gol in northern Mongolia (right). 72 Figure 5.8A (top), showing measured left vs. right maximum premaxilla groove depths for feral American horses (n = 6), Przewalski horses (n = 7), contemporary American horses (n =11), contemporary Mongolian horses (n=13), post-Bronze Age archaeological horses (n = 7), and those from deer stones and khirigsuurs (DSK, n = 12). B (bottom), shows left minus right 73 maximum premaxilla groove depths for feral American horses, Przewalski horses, contemporary American horses, contemporary Mongolian horses, post-Bronze Age archaeological horses, and those from deer stones and khirigsuurs (DSK). Observations above the dotted line indicate deeper grooves on the left side of the skull, while those below the line represent a deeper groove on the right. Specimens without both left and right premaxillary measurements were excluded from Figure 5.8B. Each observation is represented by a black dot. 5.3 RESULTS 5.3.1 Nasal remodeling Three of the 13 analyzed contemporary Mongolian horses displayed a marked concavity to the bridge of the nose, and two of these also displayed asymmetric bone thinning. One adult male horse from Morin Mort, Bayankhongor province displayed especially dramatic nasal deformation of 3.7 mm in depth, which has been described elsewhere (Taylor et al. 2016:558). This feature is nearly perfectly symmetric, and may have been caused by a halter. Both other horses with significant depressions also appear to have asymmetric thinning of the nasal bones, as suggested by the greater degree of weathering to the left side (Figure 5.6). It should be noted that as all modern Mongolian specimens were acquired through surface collection in different areas of central Mongolia, they likely experienced variable taphonomic conditions that cannot easily be controlled for. Among the Mongolian archaeological horses dating to the Iron Age and onwards, very few (n = 4) had sufficient preservation to assess nasal morphology. One of these specimens, a partially mummified young male horse from the site of Ulaan-Uneet, dating to the time of the 10th century CE, exhibited marked and asymmetric remodeling - a deep recess of over 4 mm in depth on the left side of the nasal bones (Figure 5.7, left). Within our control group of 12 contemporary American domestic horses (which included military horses, race horses, farm horses, and other animals definitively used for riding) none exhibited any clear evidence of deformation to the bridge of the nose. Only one specimen, Kidron, exhibited a possible instance of deformation: a slight, symmetric interruption in the nasal profile which could relate to the chronic use of a U.S. Cavalry bridle, incorporating a detached noseband only indirectly connected to the reins. Nasal deformation features were e ti el a se t f o o te po a P ze alski s ho ses a d fe al do esti ho ses. Nasal deformation on one of two complete Bronze Age horses, from the site of Khushuutiin Gol in northern Mongolia, was markedly asymmetric – with a deep depression of more than 4mm in depth on the left nasal bone, and only a shallow groove of around 1mm on the right. Symmetric nasal remodeling was also visible on two juvenile horses (1-3 years old) which exhibited moderate deformation. 5.3.2 Premaxillary remodeling Seven of the 13 contemporary Mongolian horses with measurable premaxillae displayed deeper premaxillary grooves on the right side, in some cases showing a discrepancy of more than 1mm between the left and right premaxilla. In contrast, only three observed specimens were 74 symmetrical (i.e. no premaxilla grooving), and three exhibited a slightly deeper groove on the left (Figure 5.8). Fewer specimens were available to characterize archaeological riding mounts dating to the Iron Age and onwards (7 adult horses), premaxilla grooving also appears asymmetric in this sample (Figure 5.8, second from right). Interestingly, two animals dating to the early Turkic period (6th-8th centuries AD) had an appreciably larger groove on the left premaxilla (Figure 5.8B). Still, the mean depth was larger for right premaxillary grooves –one specimen from the Xiongnu culture (ca. 200 BCE- 100 CE) had a negative groove differential of nearly 1mm. Because of the small sample and effect size, these apparent patterns could not be statistically validated. Among the studied sample of contemporary American domestic horses from museum collections, grooving to the exterior of the premaxilla is remarkably symmetrical, even among those with comparatively severe remodeling. The same trait also characterizes the lateral grooves observed in unridden feral and captive wild horses. A one-way analysis of variance (ANOVA) between feral American horse (n = 6), Przewalski horse (n = 7), and contemporary American domestic horse (n = 11) samples provides some evidence that the contemporary Mongolian horse sample (n =13) has more negative groove differentials -- i.e. deeper right premaxilla grooves – tha the othe g oups p < . . Mo eo e , a Ba tlett s test for equal variance suggests that these four groups have different variances (p < 0.01), with the modern Mongol horse sample displaying the greatest variation in premaxilla groove depth. Because study samples were drawn from museum and opportunistic surface collections, it is unclear how reliably they may represent the larger populations. A test of sample normality in the statistical package R using the qqplot() function also shows a heavy-tailed distribution, suggesting that the assumption of normality underpinning these tests may not be entirely justified. Only a handful of late Bronze Age specimens had both left and right margins present and sufficiently preserved for analysis (n = 4), invalidating attempts at statistical comparison. Three of these exhibited deeper grooves to the right premaxilla, and among all measured specimens (n = 12), the mean depth for right premaxilla grooves (0.83mm) was higher than that for left premaxilla grooves (0.49 mm). 5.3.3 Oral and bitting damage Bitting damage on contemporary Mongolian horse specimens consisted primarily of severe, parallel-sided wear to the anterior margin of the lower second premolar, similar to that noted by Bendrey (2007). Of the six total mandibular specimens analyzed, four exhibited this type of damage, but there was no discernable pattern in the length or severity of anterior enamel wear between the left and right. Two horses exhibited severe premortem enamel chips and erosion of the lower margin of the left P2. We observed no instances of occlusal premolar beveling in modern horses, but one specimen collected from Gobi-Altai province in southwestern Mongolia had a strange occlusal concavity on the anterior portion of the lower left P2. Archaeological horse specimens from the Iron and Middle Ages exhibited a variety of tooth damage which may be related to metal bits. These include occlusal beveling with even cementum and enamel wear, parallel anterior enamel exposure on the upper and lower premolars, enamel chips and cracks, and bone formation (Table 6.2). While anterior enamel 75 wear indicative of bitting was common on both the left and right lower premolars, all three cases of occlusal beveling to the lower P2 were more extreme on the left side. Several horses also displayed damage to the upper premolars, including concave wear of the upper occlusal surface similar to that linked by Johson and Porter (2006) with bit-chewing (Figure 5.9A) as well as flat, even wear to both upper and lower premolars (Figure 5.9B). This occlusal damage removed a significant portion of the anterior part of the tooth, and must have been caused by either intentional rasping/dentistry, or direct wear to the tooth margin during periods of extreme rein tension. Finally, one specimen displayed new bone formation to the diastema at the corner of the lower left P2 (Figure 5.9C). Previous analysis of bit wear in DSK archaeological samples showed little evidence of occlusal beveling or other bitting trauma (Taylor et al. 2016). However, this study did reveal a high instance of premortem chipping and cracking with asymmetric frequency. Nine of 48 DSK specimens exhibited non-taphonomic chips to the left premolars, with only four of 48 specimens displaying a similarly damaged right premolar. One specimen, a horse from the site of Zeerdegchingiin Khoshuu in northern Mongolia, displayed chips of identical size and placement to the anterior margin of both the left and right lower premolar (Figure 5.10). Specimen LP2 Bevel Historical period Greaves effect Left Right Malocclusion caused? Anterior enamel wear Left Right Other Upper occlusal wear (UL and UR P2) Upper occlusal wear (UL and UR P2) Concavity in LRP2 anterior margin Cracked enamel (LLP2) Upper occlusal wear (UL and UR P2) NMM 013 Pazyryk (6th2nd cent. BCE) 7.35 NA Yes No U --- NMM 071 Xiongnu (2nd cen. BCE- 1st cent. CE) 6.86* 5.37 Yes No U U NMM 080 Xiongnu (2nd cen. BCE- 1st cent. CE) 0 0 No --- L U L NMM 094 Xiongnu or Xianbei (1st3rd cent. CE) 0.66 0.79 No --- L --- NMM 011 Turkic (6-8th cent. CE) 9.81* 8.22 Yes No L L NMM 081 Turkic (6-8th cent. CE) 6.83* 4.84 Yes Unknown --- --- --- L Bone formation (Left), Enamel chip (ULP2) NMM 082 *de otes deepe Turkic (6-8th cent. CE) 3.12 0.77 No --- e el o a tooth ith o G ea es effe t e e e a el a d e e tu L ea Table 5.2 Possible Bit-related oral damage among adult horses from post-Bronze Age archaeological contexts. 76 Figure 5.9A (top), showing concave wear to the upper P2 occlusal surface and flat beveling of the lower P2 in a Pazyryk horse from western Mongolia. B (center), flat beveling of both lower and upper premolars in a Xiongnu period horse from western Mongolia. C (bottom), bone formation on the left mandibular exterior on a horse dating to the Turkic period, likely caused by a bit. Figure 5.10. Identical enamel chips on the anterior surface of the lower second premolars of a horse from the site of Zeerdegchingiin Khoshuu in northern Mongolia, which may have been caused by a hard bar bit. 77 5.4. DISCUSSION These data provide support for the idea that riding horses with a bridle placing chronic pressure on key areas of the nose and mouth produces asymmetric deformations to the equine skull, and therefore this practice might be identifiable archaeologically. Nasal remodeling likely caused by a bridle noseband was identified in contemporary Mongolian horses, but largely absent from the analyzed sample of race, military, and farm horses from American museum collections. By itself, such deformation is not indicative of horseback riding, as it can occur on haltered but unridden animals, those used for pulling carts or sleds, or in agriculture (Taylor et al. 2016). Nonetheless, the presence of larger taphonomic holes on the left side of the nose suggests preferential thinning of the nasal bones, which may be related to asymmetric pressures during use. A young horse recovered from a 10th century CE burial at the site of Ulaan-Uneet, in association with riding equipment, exhibited an extreme example of this left-side deformation, strengthening the argument that this feature is caused by riding activities. Another adult male horse also dating to this period, recovered from Bayan-Ulgii province, displayed a small depression on its left side. Asymmetric premaxillary grooving is also a feature of the contemporary Mongolian domestic horse skulls examined, a trend which might relate to riding activity. This grooving, which is associated with a branch of the infraorbital nerve, might develop in response to chronic p essu e o i fla atio aused the idle s heekpie e o heek i g du i g idi g. Those Mongolian horses analyzed in this study displayed consistently deeper grooves to the right premaxilla margin. The archaeological riding horses from Iron and Middle Age burials also showed more asymmetry than contemporary domestic and wild horses from American museum collections, although this pattern was less consistent in terms of deeper right premaxilla grooves. In early nomadic bridles, the large rigid metal, antler, or wooden bars flanking the bit on either side of the cheek probably placed even more substantial pressure on this area of the ho se s a ato o oth sides of the fa e tha do o te po a idles, hi h ay help explain this variability. A larger dataset will be necessary to clarify whether premaxillary grooving asymmetry identified in these samples is indicative of broader trends in contemporary and ancient populations. The dentition of 15 modern Mongolian horses analyzed for this study reveals contact between bit and premolar, but little evidence of asymmetry. Most horses in this group exhibited severe wear to the anterior margin of the second premolar with morphology diagnostic of bit use. A localized concavity on the lower left P2 of one modern horse could have been caused by bit chewing: a behavior commonly observed among contemporary Mongolian horses, but not otherwise connected with osteological changes in the study sample. Despite extreme bit pressures observed on the teeth during our ethnographic study, none of the analyzed specimens exhibited definitive occlusal bit wear. Modern Mongolian bits have a unique structure in comparison to their ancient counterparts – with large, curved canons that may alte the it s position in relation to teeth under rein tension, and make it more difficult to produce occlusal changes during horseback riding. This difference may explain the general absence of occlusal tooth wear, and comparatively high frequency of anterior damage in this group. 78 In contrast, several archaeological riding horses from the Iron and Middle Ages displayed severe, flat occlusal damage to the lower premolars, with more invasive beveling on the lower left premolar. An expanded sample of historical horses from Mongolian contexts will help explore whether this asymmetry is meaningful, or related to horseback riding. Other kinds of tooth wear, including anterior enamel exposure, upper premolar wear, and bone formation were also common in this group, although few directional trends emerged. Two kinds of upper premolar beveling identified here are also worth of further study, and could relate to bit chewing or direct wear during mounted riding. 5.4.1 Horse monument at Arvaikheer One issue with quantitative comparison of contemporary and ancient horses is that of sampling bias. Most of the analyzed horses interred in nomadic burials from the Iron Age and onwards were recovered along with bridles, saddles, stirrups, and other tack. Consequently, they may have been used quite intensively as both herding animals and war mounts. In contrast, contemporary animals recovered from the Mongolian countryside may not have been ridden as often nor as hard as their historical counterparts. It is also likely that the rigid cheekpieces used i a ie t idles had a o e se e e effe t o the ho se s fa e tha the si ple i gs o i gwith-bar configurations used today. To check our results against a sample of intensively trained and ridden animals, we visited the racehorse monument near Arvaikheer in Arkhangai province, central Mongolia. For many years, local people have placed the head of successful race horses or favored personal riding horses in lo g o s ehi d the o u e t s ho se statues. Due to thi k egetation and differential specimen preservation, the total number of horses at this monument is difficult to calculate, but we observed at least 296 individual horse skulls. These horses are revered – many adorned with special khadag or prayer scarves. Because of this, their skulls could not be disturbed or handled from their original position. Moreover, only a limited number were fully exposed at the surface, with preservation levels ranging from freshly deceased and fully fleshed to indistinct bone scatters. Quantitative conclusions cannot thus be reliably drawn from this sample regarding the actual prevalence of cranial changes in a live population of Mongolian riding horses. Nonetheless, several inferences can be drawn from observations of Arvaikheer horse skulls. First, even allowing for taphonomic degradation, the absolute frequency of nasal deformations appears relatively low. We noted visible premaxillary grooving deformations on only ten animals, and only ten cases of demonstrable nasal remodeling or thinning. This suggests that other factors – perhaps related to equipment fit and riding style – influence cranial remodeling. Most skulls with visible premaxillae also had exposed and badly damaged nasal bones, while crania with intact nasalia were partially buried. Consequently, it was not possible to assess whether any specimens displayed both kinds of remodeling. Still, these specimens exhibited consistent patterns in asymmetry which provide independent support for our hypothesis. Those specimens with premaxillary remodeling visible to the naked eye all showed apparently deeper g oo es o the a i al s ight side Figu e 5.11, top). Moreover, of the eight specimens with visible thinning of the nasalia, seven displayed appreciably larger holes on the left side of the nose (Figure 5.11, bottom). This suggests a pattern of greater bone thinning on the left side of contemporary riding horses. These results provide independent support for the idea that contemporary Mongolian horsemanship produces recognizable, asymmetric deformations to the equine skull. 79 Figure 5.11. Racehorse skulls at Arvaikheer displaying premaxillary remodeling (top), and nasal thinning (bottom). 5.4.2 Assessing late Bronze Age horse use In light of these finds, the asymmetric cranial bony changes identified on horses recovered from deer stones and khirigsuurs may be plausibly linked to horseback riding. In the context of contemporary and historic comparative samples, the presence of a markedly asymmetric nasal deformation in the best-preserved DSK skull is noteworthy, and must have been caused by either left-handed rein tension, or an alternative source of similarly consistent and asymmetric 80 pressures on the equine skull. Sample size and preservation prevent robust quantitative comparison with modern samples, but the Bronze Age horses considered in this study also displayed generally deeper grooves on the right premaxilla. The anaylzed sample of Bronze Age Mongolian horses does not display occlusal beveling, consistent with the inference that organic bits were used in the region at this time (Taylor et al. 2016). However, the analyzed sample did exhibit a high frequency of chips and cracks with a slightly higher frequency on the left. This result could have any number of natural or taphonomic causes, could relate to riding strategy. To a k the ho se s e te io tooth e a el with such regularity, it is likely that at least some DSK bits were made of a solid bar of bone, rather than wood or leather. This inference is supported by two identically-placed chips to the midsection of the anterior margin of the lower second premolars on a horse from the site of Zeerdegchingiin Khoshuu in northern Mongolia (Figure 5.10), possibly caused by a single contact event with a hard bar bit. If enamel chips on late Bronze Age Mongolian horse teeth are anthropogenic, the strong left-biased asymmetry may be attributable to horse riding. Horseback riding in the DSK complex would have several key implications for the development of nomadic culture and horse transport in eastern Eurasia. As most other conclusive traces of equestrianism can be dated to the early first millennium BCE (Drews 2004), DSK culture might have been among the earliest in eastern Eurasia to engage in reliable, widespread riding. Because DSK culture is also linked with the emergence of mobile pastoralism (Houle 2010:180), evidence for DSK riding would support the idea that early pastoralism was linked to the development of mounted riding (Beardsley 1953). Future applications of the techniques presented here will help evaluate the generalizability of patterns identified here, along with hypothesized links to mounted horseback riding and lefthanded reining. Doing so will require an expanded sample of modern and ancient skulls, and detailed comparison with specimens used only in ancient wheeled vehicles – to rule out the possibility that chariot use could produce similar results. Nonetheless, it appears that study of equine cranial asymmetry is a particularly fruitful line of inquiry for tracing the transition to mounted riding using only prehistoric archaeofaunal material. This method may help resolve key debates related to the chronology of horse transport in other early Eurasian archaeological contexts, where skeletal remains are often the only direct evidence for how horses were used in antiquity. The approach outlined here expands the scope of archaeozoological inquiry to include fine-grained aspects of transport type and riding style, and should be investigated in other archaeological contexts relating to horse use. 5.5 CONCLUSIONS Ethnoarchaeological study among contemporary Mongolian herders raises the possibility that the bridle and riding style used in Mongolia today causes asymmetric effects to the skull of the horse. These assymetric features can be identified through osteological study. Detailed comparison of modern and historical horse skulls points to preferential impacts to the left side of the nasal bones and right side of the premaxilla, which may be caused by horseback riding. Future research will be necessary to investigate whether such cranial changes also develop in the context of ancient chariot use. A sample of 46 late Bronze Age horses from deer stones and khirigsuurs exhibits asymmetric deformation of the premaxilla and nasal bones that may be 81 consistent with mounted riding. Methodological refinement will move archaeologists closer towards reliable identification of mounted horseback riding in other prehistoric contexts, and improve our understanding of how equine transport shaped human societies. 82 CHAPTER 6: A BAYESIAN CHRONOLOGY FOR EARLY DOMESTIC HORSE USE IN THE EURASIAN STEPPE In review, Journal of Archaeological Science William Timothy Treal Taylor1*, Jargalan Burentogtokh2, K. Bryce Lowry3, Julia Clark4, Tumurbaatar Tuvshinjargal5, and Jamsranjav Bayarsaikhan6 A haeologi al ho se e ai s f o Mo golia s late B o ze Age Dee “to e-Khirigsuur (DSK) culture present some of the oldest direct radiocarbon dates for horses in northeast Asia, hinting at an important link between late Bronze Age social developments and the adoption or innovation of horse transport in the region. However, wide error ranges and imprecision associated with calibrated radiocarbon dates obscure the chronology of early domestic horse use in Mongolia and make it difficult to evaluate the role of processes like environmental change, economic interactions, or technological development in the formation of mobile pastoral societies. Using a large sample of new and published radiocarbon dates, this study presents a Bayesian chronological model for the initiation of domestic horse sacrifice at DSK culture sites in Mongolia. Results reveal the rapid spread of horse ritual over a large portion of the eastern Steppe circa 1200 BCE, concurrent with the first appearance of chariots and horses in China during the late Shang dynasty. These results suggest that key late Bronze Age cultural transformations – specifically the adoption of mobile pastoralism and early horseback riding – took place during a period of climate amelioration, and may be linked to the expansion of horses into other areas of East Asia. Keywords: horse domestication, pastoralism, Bayesian modeling 1 University of New Mexico, MSC01-1050, Albuquerque, NM 87131 2 Yale University 3 University of Chicago 4 American Center for Mongolian Studies 5 National Museum of Mongolia *corresponding author 6.1 INTRODUCTION Researchers studying eastern Eurasia have considered a wide range of potential processes to explain the first formation of mobile pastoral groups in the region. These include prolonged drought and/or climate deterioration, a growing dependency on agricultural economies in China, and the invention of the bronze snaffle bit (Honeychurch 2015:128; Khazanov 1984:9394). Horses increase the mobility of hunters and herders, and provide critical subsistence advantages in arid and cold environments (Anthony et al. 1991). As a result, the innovation or adoption of horseback riding has also been connected with the emergence of migratory herding societies in eastern Eurasia during the late Bronze and early Iron Age (e.g. Beardsley 1953; Lattimore 1940). However, as these different social, technological and environmental processes 83 took place at different times and scales in prehistoric Eurasia, assessing the relationship between horseback riding, incipient mobile pastoralism, and other hypothesized causes requires a precise and reliable chronological framework for domestic horse use. A variety of archaeological evidence places a shift towards highly mobile pastoralism in Mongolia during the late Bronze Age. Although people in eastern Central Asia practiced mixed hunting, agriculture, and cattle-breeding as far back as the Neolithic, important social transformations appear to have occurred towards the end of the second millennium BCE (Houle 2010:4-10). At this time, large stone mounds known as khirigsuurs were first constructed across the steppes of Mongolia, southern Tuva, eastern Kazakhstan, and northern Xinjiang (Bayarsaikhan 2016). These mounds often contain human burials, and are sometimes accompanied by anthropomorphic deer stones – tall standing stones decorated with weapons, tools, and often elaborate deer images. Together, these two types of monument are referred to as the Deer Stone-Khirigsuur (DSK) Complex (Fitzhugh 2009a). Analysis of campsites and faunal remains from this period point to residential mobility and domestic sheep, goat, and cattle consumption (Broderick et al. 2014; Houle 2010). Importantly, ritual inhumations of horse skulls, hooves, and neck bones, oriented to face east, are also common at deer stones and khirigsuurs (Allard and Erdenebaatar 2005). Characteristic osteological changes to the skull indicate that many of these horses were bridled and heavily exerted, while demographic data from dentition suggest that adult male animals were buried in prominent ritual locations (Taylor et al. 2015; Taylor 2016). Together, these data suggest an increasingly important role for horses in DSK society, concurrent with the adoption of mobile herding lifeways. 6.1.1 DSK horse use in chronological context In addition to the association between DSK culture and early mobile herding, the DSK period spans a critical transition in the history of horse use in eastern Eurasia. Horses were likely domesticated in the steppes of western Central Asia ca. 3500 BCE (Outram et al. 2009). By at least the second millennium horse-drawn vehicles were employed by semi-nomadic, agropastoral people in limited seasonal migrations in the western Central Asian steppes (Khazanov 1984:93-94). Although wild E. przewalskii persisted in Mongolia until the 20th century, there is no direct archaeological evidence for domestic horses in Mongolia prior to the late Bronze Age (Honeychurch 2015:121). Nonetheless, sites attributed to the Afanasievo culture, which in other regions have been linked with horses and wheeled vehicles, are found in some areas of Mongolia as far back as the third millennium BCE (Kovalev and Erdenebaatar 2010; Houle 2010:4). Additionally, a large corpus of chariot petroglyphs can be found on Mongolian rock art panels, variously attributed to the 3rd through the 1st millennium BCE (Erdene-Ochir and Khodyakov 2016: 23-30). By ca. 1200 BCE, horses and chariots had reached central China, appearing in oracle bone records and elite burials at the site of Yinxu in Henan province. Other archaeological data demonstrate the emergence of mounted horseback riding during the late Bronze Age, prior to most estimates for the end of the DSK period (ca. 700 BCE, Fitzhugh 2009a). Early consensus evidence for mounted riding in eastern Asia includes horse tack interred i the ku ga of A zha I, dati g to a. BCE, a d si ila fi ds f o sites of the sla u ial culture in Mongolia (Honeychurch et al. 2009: 347). These dates for archaeological horse tack from Mongolia also correspond closely to the first historical mentions of mounted warriors in classical histories from western Eurasia (Argent 2011:31). If, as some suggest, nomadic peoples were among the first to adopt mounted riding (Mair 2003:181), the emergence of horsemanship 84 in East Asia must have occurred in the preceding decades or centuries – concurrent with the construction of deer stones and khirigsuurs. Due to challenges with monument dating and aggregation, however, the exact relationships between the DSK complex, changes in horse transport, and processes of social or environmental change are difficult to distinguish. Horse burials can be found at both deer stones and khirigsuurs. However, the earliest deer stones appear to postdate the earliest khirigsuurs by at least a century (Fitzhugh 2009a:189; Honeychurch 2015:117). Consequently, viable estimates for incipient horse use may fall across a relatively wide interval, between ca. 1500-1200 BCE (e.g. Fitzhugh 2009a; Honeychurch 2015:112-121). The precise timing of DSK horse use within this interval has critical implications for the role of environmental change in early mobile pastoralism. For example, many explicitly link the development of East Asian mobile herding societies with a prolonged period of drought during the second millennium BCE (e.g., Khazanov 1984:94). Recent syntheses of paleoclimate data place the end of this drought period at ca. 3000 14 C Yr BP, or the 13th century BCE (Wang et al. 2011), a date which is difficult to interpret given the wide range of estimates for DSK cultural developments. Improved chronological precision is also necessary to evaluate links between late Bronze Age horse use in Mongolia and the broader Eurasian region. For example, some researchers have hypothesized that DSK sites may have greater antiquity in northern Mongolia than elsewhere in the eastern steppes (Clark 2014:72; Fitzhugh 2009b:402). If valid, this might reflect a gradual diffusion of horses into the region out of southern Siberia and Kazakhstan circa 1400 BCE, before reaching China during the 12th century BCE (Honeychurch 2015:121). Thus, clarifying the implications of DSK horse ritual for the spread of horses into East Asia, as well as the ecological and environmental context of early pastoralism in the region, requires improved chronological precision. 6.2 MATERIALS AND METHODS To estimate the timing of DSK horse use, we compiled new and published radiocarbon dates from late Bronze Age archaeological sites. Radiocarbon dating is based on the ratio of carbon isotopes in organic materials. The proportion of 14C (the unstable isotope of carbon) present in a o ga is s tissues e ai s i e uili iu ith its e i o e t u til its death. F o this point on, the ratio changes as the unstable isotope decays away exponentially. By measuring the remaining proportion of 14C in archaeological material and then correcting (or calibrating) this result for past variations in the concentration of 14C in the environment, archaeologists can o tai a ale d i al esti ate fo the o ga is s death. U fo tu atel , these ali ated date ranges often span several centuries and produce irregular probability distributions, complicating fine-grained temporal analysis or aggregate analysis of multiple dates (Dee et al. 2013). One solution to this issue is the application of Bayesian statistics, which use prior archaeological information to model cultural phenomena (Ramsey et al. 2009). Bayesian techniques constrain error ranges and increases the precision of aggregate radiocarbon date analysis. 6.2.1 Aggregating published 14C dates We surveyed Mongolian and English language academic publications for direct dates on archaeological Horse bones from deer stones or khirigsuurs. Excluding measurements or samples of equivocal association to DSK monuments, those identified by the excavator as problematic, or those publications which did not report original uncalibrated measurements, we identified a total of 45 published dates on horse bone or tooth specimens from satellite burial features located at deer stones and khirigsuurs across Mongolia. 85 6.2.2 New 14C analysis We collected new archaeological horse material from 16 DSK horse features at localities in underrepresented regions, including Uvs, Zavkhan, Bulgan, and Bayankhongor provinces, to correct for oversampling in other areas of Mongolia (especially Khuvsgul province in northern Mongolia). For each newly processed sample, we demineralized each bone or tooth sample using hydrochloric acid, before removing humics using sodium hydroxide and rinsing with weak acid. We then gelatinized collagen from the bone solids, froze and lyophilized each using a low vacuum pump, and measured the carbon isotope ratios using the Accelerator Mass Spectrometry laboratory at the University of Arizona, Tucson, AZ. 6.2.3 Modeling DSK horse use Using the resulting sample of 61 uncalibrated dates on DSK horse remains (Appendix V), we produced a single-phase Bayesian model with a uniform prior using the program OXCAL and the INTCAL13 radiocarbon calibration curve (Bronk Ramsey and Lee 2013). To identify potential spatio-te po al g adie ts i ho se use, e p odu ed ti e sli e aps sho i g odeled posterior probabilities for each calibrated date, using GPS coordinates from published field reports or personal correspondence from the original investigator. We repeated this analysis o e ith a outlie odel hi h ide tifies a d do -weighs anomalous measurements (Bronk Ramsey 2009:356). To accommodate the possibility that a uniform prior does not adequately characterize the likelihood of sampling horse remains from the DSK period, we also ran the model using a trapezoid prior (Lee and Bronk Ramsey et al. 2012). This approach allows for greater sampling likelihood to a ds the e te of the phase i.e. attleship u e , a featu e that is often more characteristic of archaeological phenomena than a uniform probability across time. All models ran successfully until completion. 6.2.4 Modeling deer stone and khirigsuur construction To assess the relationship between early domestic horse use and late Bronze Age cultural developments, we also modeled construction of non-horse features at deer stones and khirigsuurs, using 48 new and published radiocarbon dates. These included 13 dates from charcoal excavated from within satellite features at deer stones, 29 dates on human bone from inside khirigsuurs, one date on charcoal from a khirigsuur satellite feature, and five dates on sheep or unspecified animal remains from deer stone and khirigsuur satellite features (Appendix V). Because most khirigsuurs appear to have been constructed as mortuary features (Littleton et al. 2012), radiocarbon dates on human remains from these features can be reasonably assumed to date the time of feature construction. Linking dates from ritual features surrounding DSK monuments to the event of construction is more problematic. Some researchers have argued that mounds accumulated around monuments over a prolonged period of time (Wright 2014:154). However, most stone circles and mounds associated with DSK monuments show evidence of feasting probably associated with monument dedication ceremonies (Fitzhugh 2009a:189), and radiocarbon dates from different features at the same monument often produce nearly indistinguishable radiocarbon measurements (Fitzhugh and Bayarsaikhan 2009:219). Moreover, ritual features seem to produce dates more reliably connected with DSK cultural activity than other alternatives, such as organic material recovered near the stone base 86 (Fitzhugh 2004:14-17). It is important to note that because the khirigsuur dates used here are predominantly drawn from Khuvsgul province in northern Mongolia (Frohlich et al. 2009), this phase model may not characterize the timing of cultural developments across a broader region. However, these data provide an independent sample with which to compare the chronology of horse ritual with other cultural activities at deer stones and khirigsuurs. Using the modeled posterior probabilities, e tested the h pothesis that sta t ou da ies fo ea h t pe of o u e t dee sto es a d khi igsuu s p e eded the sta t ou da fo D“K ho se itual usi g O Cal s O de fu tio . 6.3 RESULTS Estimates for the start of horse ritual at deer stones and khirigsuurs across all models place the onset of domestic horse burials at DSK sites at ca. 1200 BCE. Our trapezoid model produced a boundary estimate of 1262-1127 cal. BCE (95% probability) or 1227-1159 cal. BCE (68% probability), with a median value of 1193 cal. BCE (Figure 6.1). The substitution of a uniform prior resulted in a similar but earlier boundary midpoint estimate of 1271-1151 cal. BCE (95% probability) or 1240-1171 cal. BCE (68% probability). Because the algorithm used by OxCal allows the tails of the t apezoid dist i utio to a a o di g to the dist i utio of the data, the similarity of these two modeled start boundaries implies that the uniform prior assumption of a rapid phase transition was robust (Lee and Ramsey 2012:121). Although a few samples yielded agreement indices below the arbitrary threshold of 60%, the application of an outlier model did not significantly alter boundary estimates. Trapezoid model estimates for the termination of DSK horse ritual place this boundary between 870 and 684 cal. BCE. Comparison of the horse model with estimated trapezoid phase start dates drawn from human remains, non-horse livestock, and charcoal from deer stones and khirigsuurs supports the idea that by the time ritual inhumation of horses was initiated at DSK sites, khirigsuurs had been constructed in Mongolia for a century or more (start boundary estimate of 1439-1345 cal. BCE, 68% probability, Figure 6.2). In contrast, the estimated start boundary for deer stone construction derived from non-horse material does not differ significantly from the horse start boundary estimate (1291-1141 cal. BCE, 68% probability). A test of the relative ordering of these th ee sta t ou da ies usi g the O de fu tio i O Cal o fi s these i te p etatio s (Table 6.1), indicating a high probability that khirigsuur construction preceded the first deer stones (95.02%) as well as the first construction of horse ritual features (99.93%). On the other hand, the Order function does not provide strong evidence to differentiate the first deer stone construction from horse ritual (66.58%). Spatial pattern analysis indicates that very soon after its initial adoption, horse ritual was practiced over a wide geographic expanse of the Mongolian steppe. Figure 6.3 shows time slice maps of all horse remains with associated GPS or geographic location information, with the diameter of each circle corresponding to the relative percentage of total posterior probability for each date. The map shows that by just after 1200 BCE, horse ritual was practiced across much of the territory of modern Mongolia (Figure 6.3, A). Importantly, this process does not seem to have had a recognizable spatial gradient, with some of the earliest dates occurring on samples from Dundgovi province in southeastern Mongolia, Bayankhongor province in central Mongolia, and the Darkhad basin along the Siberian border. The youngest dates on deer stone horses appear to have persisted in the northern reaches of the country until ca. 750 BCE (Figure 6.3, C and D bottom). 87 Figure 6.1. Posterior calibrated probability ranges for 14C dates from horse remains at deer stones and khirigsuurs. Prior distribution i di ated i light g a . Dist i utio la eled D“K Ho se ep ese ts the output of O Cal s “u fu tion, and summarizes the general chronological spread of the data. 88 Figure 6.2. Modeled start and end dates for DSK horses, Khirigsuurs, and deer stones. Dashed line indicates median modeled start date for DSK horse ritual, falling within the 1-sigma range for deer stones but outside the modeled probability distribution for khirigsuurs. Event 1 Event 2 DS KS Horse DS --- KS 4.977% Horse 66.58% 95.02% --- 99.92% 33.42% 0.0783% --- Table 6.1. Probability that t1 (left column) precedes t2 top o 89 usi g OXCAL s O de fu tio . Figure 6.3. Spatial distribution of DSK horse radiocarbon dates with available geographic provenience. For each date, the diameter of each circle corresponds to the percentage of the date s poste io p o a ilit distribution which falls within the time-slice. 90 6.4 DISCUSSION These models suggest a rapid adoption of horse ritual across the Mongolian steppe circa 1200 BCE and provide several important clues to the origins of nomadic societies in East Asia. The modeled start date for DSK horse ritual is remarkably consistent with estimates for the arrival of domestic horses in China, which archaeologists typically place somewhere between ca. 12501150 BCE (e.g. Kelekna 2009b; Linduff 2003; Wu 2013). In recent years, several researchers have advanced the idea that these chariots and horses came to the region via the Mongolian steppe (e.g. Honeychurch 2015; Shelach 2009). Our results are consistent with this hypothesis, and suggest that the first appearance of horses and chariots in China may have been linked to the expansion of horse ritual and the erection of deer stones in Mongolia. The suggestion that the earliest khirigsuur construction preceded the earliest DSK horse ritual features by nearly a century implies the incorporation of new domestic horse ritual practices into an existing culture, rather than the immigration of a new people into the region. It must be noted, however, that most of the earliest dates on khirigsuurs driving this pattern are derived from human bone (Appendix V). A significant dietary contribution from freshwater resources can produce an offset in radiocarbon measurements, and is known to bias human remains to yield an older date on the order of several centuries (Phillipsen 2013). Given the apparent dietary emphasis on domestic livestock in DSK culture (Houle 2010) and lack of evidence for meaningful fish or aquatic resource consumption, this explanation appears unlikely. Nonetheless, many of the oldest khirigsuur dates in the analyzed sample come from human bone derived from lake-rich areas of northern Mongolia - Khuvsgul and Uvs provinces – making freshwater reservoir effects a valid methodological concern. Resolving the issue further may require a detailed study of contemporary isotope signatures. In any case, the rapid, widespread appearance of DSK horse burials across the Eastern Steppe, without a recognizable spatial gradient, seems inconsistent with a gradual diffusion of domestic horses into the region. Many cultures have used horses, and conducted equine funerary ritual, without these processes leaving an archaeological signature (e.g., Mitchell 2015:110). Subsequently, the spread of horse burials during the DSK period could simply indicate a resurgence in ritual practice, rather than a meaningful shift in horse use. However, a proliferation of associated ritual activity often reflects increased investment in managing a domestic resource (Zeder 2016:334). Moreover, the similarity between modeled dates for deer stones and horse ritual features suggests that this period also witnessed important social transformations, which are often linked with the adoption of horse transport (Anthony et al. 1991). The appearance of horses in ritual features ca. 1200 BCE is also mirrored in another late Bronze Age culture found in southern and eastern regions of Mongolia, known variously as the Ulaa zuukh, Te sh, “hape Bu ial, o “ho goolji Bulsh A t-shape u ial ultu e. Bu ials of this culture began as early as the middle Bronze Age (1739-1528 cal. BCE, 95% probability, Tumen et al. 2012). Ulaanzuukh features often contain domestic fauna, and the people who erected them appear to have lived a pastoral lifestyle (Honeychurch 2015:122-3). However, the small handful of directly dated features containing horse remains all yielded radiocarbon dates of ca. 3000 14C yr BP or later (Table 6.2), precisely coeval with the earliest horse dates from deer stones and khirigsuurs. The consistency of this pattern across two different LBA cultures is difficult to reconcile with endogenous changes in horse ritual practices. Instead, it suggests a major, pan-cultural change in domestic horse use at this time, such as the introduction of horses as a livestock animal or the development of horseback riding. 91 Given the remarkable consistency of model results with the first dates for horses in China (Figure 6.4), the spread of DSK horse ritual practices could correspond to the very first introduction of both horses and chariots into the Eastern Steppes. Further research will be necessary to assess how the earliest domestic horses were used by DSK people. However, the ubiquity of chariot petroglyphs in some areas of Mongolia (Honeychurch 2015:121) combined with the presence of older cultures known to have used horse chariots, such as the Afanasievo, suggest that horses could have been present in Mongolia long before the DSK period. In any case, our model indicates that DSK horse ritual persisted in northern Mongolia until at least 800 BCE, coeval with the burials at Arzhan 1 (Figure 6.3). At this site in southern Siberia, archaeologists discovered bronze snaffle bits, cheekpieces, and other evidence for mounted riding (Jacobson-Tepfer 2015:245), along with a portion of an inhumed deer stone (Rolle 1980:44).This means that for at least the latest portion of the DSK period, horseback riding was actively practiced in adjoining areas of northeast Asia, by a culture with ties to the DSK complex. Whether or not the spread of DSK horse ritual was directly linked with horseback riding, the modeled date of ca. 1200 BCE suggested by our Bayesian analysis provides compelling evidence against a causative role for drought or environmental scarcity in the formation of East Asian mobile pastoral groups. A synthesis of regional paleoclimate data (Wang et al. 2011:82) indicates that DSK horse ritual began after a prolonged period of regional drought, and concurrent with the onset of an apparently wetter climate regime that began during the 13th century BCE (Figure 6.4). If the DSK complex indeed represents the emergence of mobile pastoralism in Mongolia, this chronology casts serious doubt on causative links between drought and the adoption of nomadic herding lifeways (e.g. Khazanov 1984:94), instead suggesting that the spread of horse herding and transport in Mongolia took place in the context of climate amelioration. Table 6.2. Radiocarbon dates from Ulaanzuukh/Tevsh/Shorgooljin Bulsh features containing horse remains. 14 C Date Ref C Date (BP) Uncertainty σ IAAA103370 3054 29 IAAA103373 3006 30 Dundgovi aimag, BGC, Baga Mongol EX 07.23 Not reported 2990 40 Bayankhongor aimag, Ulziit sum, Bulgan Uul AA108307 2482 27 14 ID Sukhbaatar aimag, Dornod Mongol Ulaanzuukh Burial C Sukhbaatar aimag, Dornod Mongol Ulaanzuukh Burial 3 Material Human Bone (assoc. w/ horse) Human Bone (assoc. w/ horse) Human Bone (assoc. w/ horse) Horse tooth Monument type Reference Ulaanzuukh/Shorgooljin Tumen et al. 2012 Ulaanzuukh/Shorgooljin Tumen et al. 2012 Ulaanzuukh/Shorgooljin Nelson et al. 2009 Ulaanzuukh/Shorgooljin This study In several recent works, scholars have posited a relationship between wetter climate intervals, grassland productivity, and the expansion of nomadic polities (Kradin 2015:45-46; Putnam et al. 2016; Pederson et al. 2014). The short Mongolian summer features mild temperatures, and is wet enough to sustain relatively productive grasslands (Goulden et al. 2011:91). However, due to the extreme seasonality of precipitation, plant cover regenerates slowly and is particularly susceptible to damage from grazing. Consideration of these factors raises the possibility that ecological decision-making played an important role in the origins of East Asian mobile pastoralism. The logistical difficulties of using light horse chariots for transport in mountainous steppe terrain have been noted by those studying Central Asian chariot images (Jacobson-Tepfer 92 2012:7). Horseback riders can move 2-3 times as far per day as those moving on foot alone, permitting mobile pastoralists to tend larger herds over larger pastures and allow grasses to replenish (Anthony et al. 1991; Anthony and Brown 2003). In the context of improving climate and its accompanying opportunities for grazing, the development of horse transport– whether through riding skill, better horse equipment, or by breeding horses with a more manageable temperament – would have enabled herders to tend more animals and capitalize on new ecological opportunities. Given the increasing returns to scale associated with livestock herding (e.g. Borgerhoff-Mulder 2010), this process might have favored rapid territorial expansion in DSK society. In this case, changes in horse transport among early Mongolian nomads might have prompted interaction between new groups of people, and facilitated the spread of horses and chariots to China ca. 1200 BCE. Results from the Bayesian analysis of archaeological radiocarbon dates indicate that late Bronze Age people rapidly adopted horse ritual practices across much of the Mongolian steppe circa 1200 BCE, concurrent with the first appearance of horse chariots in China and the cessation of a prolonged drought across much of Mongolia. Although it remains unclear whether this process was linked to the initial introduction of domestic horses or the adoption of horseback riding, contextual evidence links the spread of horse ritual practices with the construction of deer stone monuments, and a major change in horse use during the late Bronze Age. The proposed chronology is inconsistent with a link between drought and increased mobility or nomadic lifeways, and invites further study into the ecological context of early horseback riding. Figure 6.4. Modeled cultural phase start dates, as compared to large-scale climate data from Wang et al. 2011 (yellow), and important regional events in horse use. 93 CHAPTER 7: CONCLUSIONS 7.1 OVERVIEW Previous research linked the emergence of nomadic herding life in Mongolia to the late Bronze Age, in conjunction with the construction of stone monuments of the Deer Stone-Khirigsuur culture. Although the ubiquity of horse burials at DSK sites hints at an important role for horses in late Bronze Age social and subsistence transformations, data with which to evaluate this premise are limited. Moreover, without a precise chronological framework, the relationships between DSK culture, horse transport, and hypothesized environmental or climate processes linked with the first nomadic pastoralism remain unclear. In this dissertation, I used archaeozoological data from deer stones and khirigsuurs to provide direct evidence for domestic horse herding and transport in late Bronze Age Mongolia. Demographic profiles show that DSK people actively bred and managed horses as livestock, selecting young animals and those beyond the age of reproduction for sacrifice in ritual activities. People of this culture also buried adult male horses in special locations at the eastern edge of deer stones and khirigsuurs, demonstrating a prominent role for horses in DSK culture and ritual practices at this time. When considered alongside previous studies linking this period with residential mobility and the consumption of sheep, cattle, and horse, these data support the idea that DSK culture represents one of the earliest highly mobile pastoral societies in eastern Eurasia. Osteological techniques developed through this dissertation help to effectively identify horses used for riding or chariotry using equine cranial remains, even in the absence of tack or historical records. Detailed comparison of contemporary and archaeological horses ridden with known tack indicates that some specific bridle components – including a noseband, metal bit, and perhaps a rigid cheekpiece – can be inferred from osteologi al ha ges to the ho se s skull visible in archaeological specimens. While these changes alone cannot distinguish mounted riding from other transport uses, such as pulling carts, asymmetry in these features characterizes the skulls of ridden horses from contemporary Mongolia, and might help identify horseback riding. The identification of these anthropogenic changes in a sample of DSK horses indicates clearly that some of these animals were bridled and heavily exerted in transport, and may have been ridden. Carvings on Mongolian rock art attributed to the late Bronze Age, as well as a handful of carvings made directly on deer stones depict horse chariots. Consequently, it is almost certain that DSK people used horses to pull wheeled vehicles. However, the sample of late Bronze Age horse remains analyzed here display asymmetric deformation patterns consistent with those seen on contemporary ridden animals, raising the possibility that DSK nomads were among the earliest to engage in mounted riding in East Asia. This evidence corroborates reports of human skeletal pathologies linked with horseback riding in human burials from khirigsuurs (Frohlich et al. 2009:107), as well as cultural links between DSK culture and first millennium BCE equestrian peoples in Tuva and south Siberia (Hanks 2012). Interestingly, these DSK horses show little of the oral bitting damage typically associated with metal bit use – indicating that they were controlled with a bitless bridle or organic bit made of leather, bone, or other material. If these animals were indeed used as mounts, this finding would contradict hypothesized links between the invention of particular horse equipment – such as the metal snaffle- and the initial spread of mounted horseback riding (e.g. Drews 2004:90; Honeychurch 2015:128, 210-11). 94 Bayesian statistical analysis places these developments in chronological context, suggesting that a widespread emergence of domestic horse ritual took place circa 1200 BCE. This period was one of climate amelioration and increased rainfall in the eastern Eurasian steppe, weakening explanatory frameworks which prioritize resource scarcity or climate stress as causal factors in the origins of East Asian nomadic pastoralism. The apparent rapidity of this spread of horses was also concurrent with the first construction of deer stones, and the initial arrival of domestic horses in other areas of the continent – such as Shang Dynasty China. This model thus links the spread of DSK horse ritual with key social changes and the territorial expansion of horses. Because similar processes have been linked with horseback riding in other historical and archaeological contexts (Anthony et al. 1991), these data provide further evidence connecting DSK culture with the spread of mounted horseback riding in the Eastern Steppe. 7.2 IMPLICATIONS FOR THE ORIGINS OF HORSE RIDING AND NOMADIC SOCIETIES IN EAST ASIA One of the key puzzles in the story of horse domestication is the question of when, and why, horseback riding was first practiced on a wide scale – and why this process apparently lagged centuries behind the development of horse carts and chariots. Due to the logistical challenges involved in herding horses, some researchers argue that mounted riding must have been practiced since the early days of horse domestication (e.g. Anthony 2007). Nonetheless, apart from the occasional image or textual reference suggestive of horse riding (e.g. Khazanov 1984:92), neither the historical nor the archaeological record provides unambiguous evidence for reliable or widespread horse riding until the first millennium BCE (Drews 2004). Archaeologists have yet to settle on a consensus explanation for this discrepancy, but the case study provided by late Bronze Age Mongolia may provide one useful explanation for the delayed emergence of mounted riding in eastern Eurasia. While horse riding may seem intuitively easier than chariot or cart traction to the modern observer, horse transport introduced a series of behavioral and logistic obstacles which may be more easily solved by chariots than by riding (Dietz 2003). Early horses likely had an aggressive disposition, and unique anatomical issues that would have made riding dangerous for the rider and difficult for the horse. Chariots would have placed less anatomical strain on the back and front quarters, while the presence of a second horse in a chariot team, presuming the animals were agreeable to one another, would soothe the a i al s atu al pa i ked flight espo se (Dietz 2003:190). Overcoming these barriers to riding would likely have required not only learning and experimentation by humans, but also sustained, multigenerational genetic changes in horses. In fact, detailed full- o e age ge o i o pa iso of do esti ho ses ith P ze alski s ho se – their closest living relative— indicates that many divergent alleles between these two species are related to metabolism, cardiac function, and musculature. These differences probably reflect changes related to anthropogenic selection in domestic horses (der Sarkissian et al. 2015:2579. Breeding and genetic changes can be considered a form of long-term, direct investment in the improvement of a resource – a behavior usually practiced in conditions of stable, abundant resources by those with detailed traditional ecological knowledge, an important component driving the evolutionary trajectories of domestic animals (Zeder 2016:332-4). Based on the admittedly sparse archaeological record of pastoral groups in the early and middle Bronze Age, it is reasonable to infer that by the second millennium BCE, Mongolian groups had prolonged familiarity with several kinds of domestic livestock, including horses. Selective culling of young, 95 likely male horses – evidenced by mortality profiles from DSK sites – would have reduced mate competition, and meant that the selective choices of Bronze Age herders had a lasting genetic lega o e ologi al i he ita e )ede : . Recent full-genome DNA analyses of early Iron Age horses Pazyryk from the Altai Mountains show that few deleterious mutations related to anthropogenic selection had accumulated by the mid-first millennium BCE (Pennisi 2016). This result seems consistent with the idea that sophisticated horse riding first began during the DSK period, in which case targeted efforts at breeding horses for riding may have only been a few centuries old. Because traditional ecological knowledge is often directly reflected in ritual practices and the built environment (Zeder 2016:332), broad-scale patterns in horse ritual may help explain the relative chronology of late Bronze Age horse use in Central Asia. Although horses still played an important role in ritual activities, archaeological sites from Kazakhstan and western Central Asia show a gradual decrease in the visibility of domestic horses in the archaeological record from their initial domestication until ca. 900-700 BCE (Frachetti 2008:160; Outram et al. 2010:119120). This decline may have accompanied the transition from sedentary hunting and herding to a specialized, semi-nomadic herding lifestyle. In this scenario, horses may have been used for transport by middle Bronze Age groups who herded cattle, sheep, and goat, but were no longer the key to subsistence (Outram et al. 2010:126) – a striking contrast to their economic importance in later eastern Eurasian nomadic societies. A delayed chronology for the emergence of horseback riding offers a solution to this puzzle. Perhaps in the centuries before the advent of riding, smaller numbers of horses were kept for use as draft (chariot) animals and other economic activities, like dairying. The absence of horseback riding before the DSK period may also help explain some interesting aspects of the Bronze Age archaeological record in Mongolia. The high Altai and Khangai mountain fronts capture rain moving westward into the eastern Steppe, making precipitation more seasonally stable than adjoining areas of Mongolia. Montane pastures and alpine meadows in interior Central Asia are inaccessible much of the year because of snow, but can be incredibly productive in summer (Frachetti 2008:88-98). If Mongolian groups began to adopt domestic livestock in the 3rd millennium BCE, the limited mobility offered by wheeled carts may have been sufficient for limited seasonal migration in these more reliable mountainous regions – particularly during the dry interval preceding the late Bronze Age. The lack of mounted transport might thus explain apparent clustering of chariot petroglyphs in the mountains of central and western Mongolia (Honeychurch 2015:193). The decreasing archaeological visibility of horses in western Central Asia into the early first millennium BCE is in marked contrast to the expansion of megalithic structures, ritual, and feasting activities related to horse ritual in DSK culture circa 1200 BCE. Similar proliferation of ritual activity often indicates increased investment in managing a domestic resource (Zeder 2016:334). Consequently, a viable explanation for the rapid geographic spread of DSK horse ritual circa 1200 BCE observed in Bayesian modeling is that this period witnessed the initial development of horseback riding and mobile pastoralism. Armed with detailed traditional ecological knowledge of domestic livestock, and prompted by the ameliorating climate, late Bronze Age herders in Mongolia may have been uniquely positioned to overcome the inherent challenges of using horses for mounted riding. By incorporating horses into domestic livestock herds, pastoralists could maintain larger herds with better winter survival in the arid Eastern Steppe. The grazing opportunities from an improved 96 precipitation regime would have provided strong to experiment with new uses of the horse. As grasslands expanded, reliable horseback riding would enable DSK nomads to move herds over larger distances, capitalizing on the new pasturelands that were still susceptible to overgrazing because of climate seasonality. Mounted transport would have lowered the costs of defending large swaths of territory and productive pasture. Most importantly, the increasing returns to scale documented for ethnographic livestock herds (Borgerhoff Mulder et al. 2010), and the unexploited niche presented by these pastures would have provided a mechanism for rapid territorial expansion across the Eastern Steppe. 7.2.1 Connectivity, communication, and trade In human prehistory, changes in the scale of interaction and communication often prompted renegotiation of political boundaries, economic relations, and social relationships (Frachetti 2008:28). Among pastoralists, even minor changes in the scale of movement can spark the development of new networks of interaction, and influence the spread of ideas and materials (Frachetti 2008:24). The adoption of sophisticated horseback riding by DSK society might have stimulated these processes on a continental scale. Predating the formal trade route known as the Silk Road, continental trade flourished through i fo al “teppe ‘oads egotiated B o ze Age pasto alists Ch istia . I deed, the more southerly routes now popularly associated with the Silk Road may have been a deliberate creation in the late first millennium BCE, as the Han Dynasty sought relief from the control of steppe routes by the nomadic Xiongnu (Christian 2000:85). Linguistic and archaeological data o e t the fi st ho ses a d ha iots appea i g i elite u ials du i g Chi a s late “ha g D ast in the late second millennium BCE likely derived from steppe regions to the north (Mair 2003:182). These horses were also accompanied by a surge in steppe-provenanced artifacts and designs, such as animal-headed knives (Wu 2013:39). A desirable trade good themselves, horse transport and increased mobility would have positioned Eastern Steppe nomads to capitalize on trade through the continental interior like never before. Early trans-Eurasian trade networks saw an intensification of exchange and an expansion of the scale of social interaction during the early first millennium BCE (Christian 2000:82-83; Frachetti 2008:141). As one key example, the western Eurasian domesticates wheat and barley made their first appearance on the Tibetan plateau during the late second millennium BCE, perhaps in conjunction with the spread of pasto alis a d ho ses i to adjoi i g a eas d Alpoi Guedes et al. : -6). The new trade networks facilitating such critical ecological exchanges could have been prompted by the innovation of horse riding and the concurrent spread of mobile pastoralism in Mongolia (Honeychurch 2015:69). In a changing global system brought about by increasing connectivity, those groups able to successfully leverage their advantages (such as geographic location, ecologically favorable circumstances, or technology) often benefit greatly at the expense of those who are not (Mann 2011). As the steppe improved under a cooler, wetter climate, ridden horses would have provided DSK people with all three. The speed of mounted horseback riding and leap in connectivity associated with DSK horse use and mobile pastoralism –developed or adopted for pastoral purposes – might have nonetheless drastically reshaped the trade networks, political relationships, and social systems of the region during the late Bronze Age. 7.2.2 Social legacy 97 The innovation of mounted riding may also explain changes in social dynamics behind the construction of large stone monuments and the scale of animal sacrifice at deer stones and khirigsuurs. Because niche construction produces lasting ecological modifications that improve esou e a aila ilit , it i eases the oppo tu it fo f ee ide s to e ploit e esou es without contributing to their production (Zeder 2016:333). These additional resources provide incentive for the development of cooperative behaviors, as people move to monopolize access a d e lude heate s. As ou ted idi g ha ged the e ologi al pa a ete s of the d Eastern Steppe, new difficulties protecting herds and newly viable areas of grassland would also have emerged. Seasonal pastures might now lie at great distances, introducing the danger of neighbors grazing a favorite pasture or rustling livestock. These risks might have encouraged DSK people to develop a means of monopolizing territory through monument construction, and engage in cooperative behaviors like the large feasts and animal sacrifices associated with deer stone and khirigsuur construction. The new demands and incentives provided by horse riding, and the ecological niche it opened up, probably stimulated monument building and the development of larger, cooperative social networks. Even as horse riding brought new opportunities, the unique challenges associated with mobile pastoralism in Mongolia likely facilitated social stratification. Recent simulation research suggests that over repeated generations, the unequal geographic distribution of quality pasture combined with intermittent, extreme winter livestock losses (a regular feature of mobile pastoralism in the Eastern Steppe, referred to as a dzud) would have been enough to some herding groups comparatively wealthy (Borgerhoff-Mulder 2010; Shultz and Costopolous n.d.). This incipient inequality is more likely to develop under conditions of heterogeneous resource distribution and increased carrying capacity, such as the wetter interval observed during the early DSK period (Shultz and Costopolous n.d.). Consequently, the expansion of Mongolian herding groups into drier and riskier areas of the Eastern Steppes enabled by horseback riding created a new need for social networks and cooperative behavior, while also indirectly stimulating inequality in animal wealth. Together, these processes may help to explain the emergence of enormous monument complexes and associated animal sacrifices like those at Urt Bulagyn, where nearly 2000 individual horses were sacrificed (Allard and Erdenebaatar 2005), as well as the emergence of complex, hierarchical social structures associated with early steppe states linked with the late Bronze Age (e.g. Rogers 2012:213). In the centuries following the DSK period, horses would influence the social landscape in other, equally influential ways. As horse control became increasingly sophisticated, horse cavalry underwrote pastoral military successes, laying the foundation for the spread of nomadic horse cultures across most of the Eurasian interior (Drews 2004:73) and the periodic emergence of massive steppe empires (Rogers 2012:209). Concurrent with the fluorescence of horse culture was the sp ead of a i al-st le a t as fa as easte Eu ope, hi h a ha e its ea liest o igi s in deer stone designs (Fitzhugh 2009a:193). The mobility offered by mounted horseback riding likely affected the religious landscape, too, influencing factors such as the distance between khirigsuur clusters and the location of sacred places (Houle 2010:192; Rogers 2012:211). DSK ritual sites would continue to be reused as important burial locations for later cultures, serving as important foci in the ritual landscape (Clark 2014:152). In addition to establishing patterns of connectivity and interaction, the changes in horse use effected by DSK people thus appear to have had a lasting impact on both religious practices and social structure in the Mongolian steppe (Fitzhugh 2009b: 406). 98 7.3 CONCLUSION Archaeofaunal evidence from ritual horse burials at late Bronze Age archaeological sites point to the adoption of horse herding and transport in the Eastern Steppe circa 1200 BCE. Previous studies have associated this period with the emergence of nomadic pastoralism in eastern Asia, and demonstrated that the influence of sedentary agricultural societies is a poor explanatory framework. Osteological data suggest that DSK horses may have been used not only for pulling chariots, but also for early mounted riding, predating other direct evidence for equestrianism in eastern Eurasia by several centuries. A high resolution chronological model for DSK horse use suggests that major changes in equine transport in late Bronze Age Mongolia are not attributable to drought or resource scarcity, but instead took place during a period of increased precipitation and improved grassland productivity. These results align well with predictions from cultural niche construction theory, and suggest that horseback riding could have been developed by late Bronze Age herders as a means of capitalizing on new opportunities presented by climate amelioration during the late second millennium BCE. Consequent changes to communication and connectivity may have altered the course of Eurasian prehistory, prompting the development of nomadic social hierarchy, the expansion and reorganization of transcontinental trade networks, and the exchange of horses to new areas of the continent. 99 APPENDIX I. DATA USED IN DENTAL ESTIMATES OF AGE AND SEX, ALONG WITH PROVENIENCE DATA FOR ANALYZED DSK SAMPLE (CH.2). CH = crown height, ER = eruption, GS = grinding surface/wear ID Location Provenience DS/ KS Median age est Range Canines Est. basis Crown Height( mm) Tooth CH2 Tooth CH3 Tooth Notes Grinding surface Source --- Central incisors twice as thick as are wide (~24), nearly acute angle of incidence (90◦ Triangular (18+) Bayarsaikhan (2008), Fitzhugh (2008) --- --- --- Bayarsaikhan (2008), Fitzhugh (2008) Triangular (18+) Bayarsaikhan (2008), Fitzhugh (2008) Rectangular Bayarsaikhan (2008), Fitzhugh (2008) Triangular (18+) Fitzhugh (2005) Rectangular Fitzhugh (2005) NMM 1 Khuvsgul, Tsagaan Uul, Khushuut in Gol F2 E? (missing stone) DS 27.5 20--35 Fewer than four CH, ER, GS 12.94 LRP2 5.2 UP2 NMM 2 Khuvsgul, Tsagaan Uul, Bogd Uul /Bor Hujiriin Gol F2 W exterior DS 14.5 14-15 indet CH 13.35 LRP2 --- --- NMM 3 Khuvsgul, Tsagaan Uul, Khushuut in Gol F3 E interior DS 27.5 20-35 Fewer than four CH, ER, GS 18.94 LLM3 --- --- --- --- NMM 8 BayanUlgii, Tsengel, On Khad E interior KS 6.5 6-7 YES ER, GS --- --- --- --- --- --- East of DS DS 27.5 20-35 YES CH, ER, GS 13.78 ULM2 14.26 ULM3 --- --- E DS 2.25 2-2.5 indet ER --- --- --- --- --- --- NMM 9 NMM 10 Khuvsgul, Renchinl khumbe, Tsatstain Kh. F2 Khuvsgul, Renchinl khumbe, Tsatstain Kh. F1 105 --- --- Central incisors twice as thick as are wide (~24), acute angle of incidence < ◦ Cups missing from lower central incisors (6), intermediat e angle of incidence (110◦ Central incisors twice as thick as are wide (~24) M2 in, central incisors and P2 not in, steep angle of incidence (160◦ NMM 14 NMM 15 NMM 16 NMM 17 NMM 18 NMM 20 Zavkhan, Telmen, Uguumor DS 2 Zavkhan, Telmen, Uguumor DS 8 Zavkhan, Shiluustei , Daagan Del Khuvsgul, AnarErdene, Ulaan Tolgoi DS 4 F6 Khuvsgul, AnarErdene, Ulaan Tolgoi DS4 F5 Tuv, Zaamar, AD 40-3 Bayarsaikhan and Tuvshinjargal (2013) Bayarsaikhan and Tuvshinjargal (2013) NE of DS DS 2 2 indet ER --- --- --- --- --- --- M2 in Rectangular E DS 1.75 1.5-2 indet ER --- --- --- --- --- --- M2 erupting --- Rounded(918) Bayarsaikhan and Tuvshinjargal (2013) 20.91 LLM3 --- --- --- --- Nearly acute angle of incidence (incisor malocclusio n). Nippers of upper jaw twice as thick as broad (27) SE exterior DS 27.5 20-35 NO CH, ER, GS E/SE of DS DS 14.5 11-18 indet CH 38.59 LRM3 23.78 URM3 --- --- --- --- Fitzhugh (2005) NW of DS DS 0.875 0.75-1 indet ER --- --- --- --- --- --- M1 erupting --- Fitzhugh (2005) E interior KS 11 10-12 indet CH, GS 33.3 LP3 41.8 LM3 --- --- --- Rectangular National Museum (2013) --- National Museum (2013) --- National Museum (2013) Rectangular/r ound (9-18) National Museum (2013) Rectangular Bayarsaikhan (2011) NMM 21 Tuv, Zaamar, AD 40-1 E interior KS 2.5 2.5 indet ER --- --- --- --- --- --- P2 erupted, deciduous P3 and P4 (2.5) NMM 22 Tuv, Zaamar, AD 40-4 E interior KS 11 8-14 indet CH 31.47` LM3 38.47 LM2 46.24 LP3 --- NMM 23 Tuv, Zaamar, AD 40-5 E interior KS 12 11-13 YES CH, GS 17.07 LP2 41.95 LM3 37.21 LM1 NMM 24 Arkhanga i, UndurUlaan, Jargalant yn Am F5 E of DS DS 0.75 0.75 indet ER --- --- --- --- --- --- 106 Cups worn from upper incisors (12+), Upper enamel ring present (< 13-16) Corner deciduous incisors present, M1 nearly NMM 25 NMM 26 NMM 27 Arkhanga i, UndurUlaan Jargalant yn Am F4 Arkhanga i, UndurUlaan, Urt Bulagyn KYRI 22 Khuvsgul, AnarErdene, Ulaan Tolgoi DS5F2 E of DS DS 8 7-9 YES CH, ER, GS 31.16 URP3 50.28 URM3 --- --- SE exterior mound KS 2.25 2-2.5 indet ER --- --- --- --- --- --- YES CH, ER, GS E/SE of DS DS 7.5 6-9 43.76 ULP2 --- --- --- erupted (~9 months) Cups gone from lower central and middle incisors (78) Rectangular Bayarsaikhan (2011) M2 erupted, P2 still unerupted --- Fitzhugh (2006, 2009b), Bayarsaikhan (2006) --- Cups going from lower central incisors (67) Rectangular Fitzhugh (2005) Rounded Fitzhugh (2006), Bayarsaikhan (2006) NMM 28 Khuvsgul, Renchinl khumbe, Zeerdegc hingiin Khoshuu M1 E interior mound KS 10.5 9-12 YES CH, ER, GS 35.54 ULM3 --- --- --- --- Cups gone from upper central and middle incisors (910), Intermediat e angle of incidence NMM 29 Khuvsgul, Galt, Nukthiin Am M1 F1 E interior mound KS 8 7-9 YES CH, ER, GS 57.23 LRM3 58.48 ULM3 --- --- All cups present in upper incisors (<9) Rectangular Fitzhugh (2006), Bayarsaikhan (2006) Rounded/Tria ngular Fitzhugh (2006, 2009b), Bayarsaikhan (2006) NMM 30 NMM 31 NMM 32 NMM 33 Arkhanga i, UndurUlaan, Urt Bulagyn KYRI 21 Khuvsgul, Galt, Nukhtiin Am DS F1 Khuvsgul, AnarErdene, Ulaan Tolgoi DS5F3L2 Khuvsgul, Anar- SE interior mound KS 13.5 13-14 YES CH, ER, GS 30.17 ULP3 --- --- --- --- Smooth mouthed (>12), 11-yr hook (> 11), No enamel ring (13-16), Intermediat e angle of incidence E interior mound DS 12 11-13 indet CH 33.32 ULP3 30.52 URP2 --- --- --- --- Fitzhugh (2006), Bayarsaikhan (2006) NE of DS DS 0.3125 0.1250.5 indet ER, GS --- --- --- --- --- --- Corner incisors and M1 not in (<6 months) --- Fitzhugh (2005) NW of DS DS 0.3125 0.1250.5 indet ER, GS --- --- --- --- --- --- Corner incisors and Rectangular Fitzhugh (2005) 107 Erdene, Ulaan Tolgoi DS5F4 M1 not in (<6 months) 108 APPENDIX II. AGE AND SEX DATA UTILIZED IN THIS STUDY (CH. 2) from Allard et al. (2007), Benecke (2007), and Takahama et al. (2006) ID Site Location Feature type Median est. age Sex estimate 2 Urt Bulagyn E exterior KS 6.5 Indet. 4 Urt Bulagyn E interior KS 19.5 Male 8 Urt Bulagyn SW exterior KS 15+ Female 9 Urt Bulagyn SW exterior KS Juvenile Indet. 10 Urt Bulagyn SW exterior KS 15+ Female 11 Urt Bulagyn SW exterior KS Juvenile Indet. SC5 Ushigiin Ovor SE DS Adult Indet. SC7 Ushigiin Ovor E/NE? (missing deer stone) DS Adult Male SH10 Ushigiin Ovor E exterior KS ~0.75 Indet. SH11 Ushigiin Ovor E interior KS Adult Male. SH12 Ushigiin Ovor E exterior KS Adult Indet. SH14 Ushigiin Ovor E interior KS Juvenile Indet. SH15 Ushigiin Ovor E interior KS Adult Male SH16 Ushigiin Ovor E exterior KS Juvenile Indet. SH17 Ushigiin Ovor E exterior KS ~0.75 Indet. SH18 Ushigiin Ovor E exterior KS Adult Indet. SH20 Ushigiin Ovor E exterior KS ~0.75 Indet. SH3 Ushigiin Ovor NE exterior KS Juvenile Indet. SH4 Ushigiin Ovor E interior KS Juvenile Indet. SH9 Ushigiin Ovor E interior KS Adult Male 1 --- --- --- 12.5 Male 2 --- --- --- 9.5 Male 3 --- --- --- 13.5 Male 4 --- --- --- 12.5 Male 5 --- --- --- 13.5 Male 6 --- --- --- 8.5 Male 105 Source Allard et al. (2007) Allard et al. (2007) Allard et al. (2007) Allard et al. (2007) Allard et al. (2007) Allard et al. (2007) Takahama et al. (2006) Takahama et al. (2006) Takahama et al. (2006) Takahama et al. (2006) Takahama et al. (2006) Takahama et al. (2006) Takahama et al. (2006) Takahama et al. (2006) Takahama et al. (2006) Takahama et al. (2006) Takahama et al. (2006) Takahama et al. (2006) Takahama et al. (2006) Takahama et al. (2006) Benecke (2007) Benecke (2007) Benecke (2007) Benecke (2007) Benecke (2007) Benecke (2007) 7 --- --- --- 19.5 Male 8 --- --- --- 12.5 Male 9 --- --- --- 12.5 Male 10 --- --- --- 12.5 Male 11 --- --- --- 15.5 Male 12 --- --- --- 13.5 Male 13 --- --- --- 19.5 Male 14 --- --- --- 19.5 Male 106 Benecke (2007) Benecke (2007) Benecke (2007) Benecke (2007) Benecke (2007) Benecke (2007) Benecke (2007) Benecke (2007) APPENDIX III. AGE AND PATHOLOGY MEASUREMENTS/SCORES FOR HORSES INCLUDED IN THE STUDY (CH. 3) Institution/ID Category Estimated Age Estimated Sex Nuchal ossification score (1-6) Medial depth (mm) Lateral depth (mm) Nasoincisve width Era NMM 12 Ridden 2 indeterminate 1 0.698 NA NA archaeological (Pazyryk) Personal collection Ridden 4 female 3/2 1.17 0.466 10.462 modern NMNH 174960 Ridden 6 male 5 2.071 0.828 13.881 modern KNNP 2 Ridden 6.5 female NA 0.603 0.558 9.815 modern NMM 11 Ridden 8 male 4 1.459 1.03 9.193 archaeological (Turkic) NMNH 172454 Ridden 8 male 3/2 1.955 0 8.382 modern NMNH 843 Ridden 9 male 5 1.107 0.239 11.293 modern MSB 56502 Ridden 10 male 5 0.471 0 9.65 modern MSB 193704 Ridden 12 male 5 1.081 0.943 11.408 modern NMM 13 Ridden 16 male NA 1.174 0.712 11.039 archaeological (Pazyryk) NNV 6 Ridden 20 male 6 1.741 0 9.233 modern NMNH 174968 Ridden 30 female 5 1.656 0.435 12.555 modern NMNH 270900 Ridden 30 male 6 1.756 0.82 9.117 modern Ridden mean --- 12.42308 --- 4.272727 1.303231 0.502583 --- --- NMM 18 DSK 1 indeterminate NA 0.696 0 11.075 archaeological (Bronze Age) 2 indeterminate 2 0.505 0 6.517 archaeological (Bronze Age) 7 male NA 0.954 0.613 11.692 archaeological (Bronze Age) 7 male 4 1.488 0.871 9.546 archaeological (Bronze Age) 11.5 male 4 0.813 0 9.081 archaeological (Bronze Age) 13 male NA 1.662 0.702 10.842 archaeological (Bronze Age) 14.5 indeterminate 5 NA NA NA archaeological (Bronze Age) 23 female 4 1.821 1.879 11.726 archaeological NMM 10 NMM 27 NMM 8 NMM 28 NMM 30 NMM 2 NMM 3 DSK DSK DSK DSK DSK DSK DSK 107 (Bronze Age) NMM 1 NMM 9 DSK DSK 25 indeterminate NA 0.61 0.551 11.564 archaeological (Bronze Age) 25 male 4 1.576 0 7.98 archaeological (Bronze Age) DSK mean --- 12.9 --- 3.833333 1.125 0.512889 --- --- MSB 198378 Feral E. caballus 1.5 indeterminate 3/2 0.164 0 8.138 modern NNV 4 Feral E. caballus 2.5 indeterminate 2/1 0.518 0 8.865 modern NNV 2 Feral E. caballus 4 male 2 1.281 0 12.837 modern MSB 146537 Feral E. caballus 5 female 1 0.493 0 10.533 modern NMNH 395180 Feral E. caballus 5.5 indeterminate 1 0.735 0 10.658 modern MSB 150587 Feral E. caballus 6 female 4 0.693 0 10.761 modern NMNH 268938 Feral E. caballus 8 female 2/1 0.413 0.272 10.826 modern NNV 1 Feral E. caballus 8 male 5 0.577 0 10.252 modern NNV 3 Feral E. caballus 8 male 3/2 0 0 8.518 modern NNV 5 Feral E. caballus 9 male 1 0.739 0.709 11.239 modern NMNH 302898 Feral E. caballus 11 female 2 0.261 0.279 10.412 modern NMNH 395432 Feral E. caballus 20 female 3/2 1.034 0.25 11.468 modern Feral mean --- 7.375 --- 2.208333 0.575667 0.125833 --- --- MSB 54743 Zoo E. Przewalskii 5.5 male 4 0.998 0 11.405 modern NMNH 23811 Zoo E. Przewalskii 9 male 5 0.942 0 8.578 modern NMNH 311033 Zoo E. Przewalskii 25 female 2 1.097 0 9.254 modern NMNH 582088 Zoo E. Przewalskii 6 female 3/2 1.343 0.056 9.613 modern NMNH 582467 Zoo E. Przewalskii 8 female 5 0.928 0 10.115 modern NMNH 582910 Zoo E. Przewalskii 7 male 5 0.837 0.794 11.262 modern Zoo mean --- 10.08333 --- 3.916667 1.024167 0.141667 --- --- 108 KNNP 1 Wild E. Przewalskii 7 male 3 0.393 0 11.909 modern Wild E. przewalskii mean --- 7 --- 3 0.393 0 --- --- MSB = Museum of Southwestern Biology, NMM = National Museum of Mongolia, NMNH = Smithsonian National Museum of Natural History, KNPP = Khustai Nuruu National Park, Mongolia, NNV = Navajo Nation Veterinary Clinic, Chinle, AZ. Sex estimated based on presence/absence of 4 adult canines. 109 APPENDIX IV. DENTAL AND CRANIAL OSTEOLOGICAL DATA FOR HORSES USED IN THIS STUDY (CH. 5) Premaxilla depth ID NMM 001 Era Bronze Age-DSK Provenience Khuvsgul, Tsagaan Uul sum, Khushuutin Gol DS Feature 2 Estim. age Estim. sex 14-15 (wear) LIKELY MALE NMM 002 Bronze Age-DSK Khuvsgul, Tsagaan Uul sum, Bor Hujiriin Gol 1 Feature 2 14-15 (wear) indet. NMM 003 Bronze Age-DSK Khuvsgul, Tsagaan Uul sum, Khushuutin Gol DS Feature 3 20+ (wear) LIKELY MALE L 0.186 R 0.882 LeftRigh t Nasal remod eling -0.696 LP2 Bevel L R 0 0 0 1.762 NMM 008 Bronze Age-DSK Bayan Ulgii, Tsengel sum, On Khad KS 6-7 (wear) MALE 0.758 NMM 009 Bronze Age-DSK Khuvsgul, Renchinlkhumbe sum, Tsatstain Khushuu DS Feature 2 20+ (wear) MALE 0.356 NMM 016 Bronze Age-DSK Zavkhan, Shiluustei sum, Daagan Del DS 20+ (wear) FEMALE NMM 017 NMM 020 NMM 022 NMM 023 Bronze Age-DSK Bronze Age-DSK Bronze Age-DSK Bronze Age-DSK 11-18 (wear) 10-12 (wear) 10-14 (wear) 10-15 (wear) NMM 025 Bronze Age-DSK NMM 027 Bronze Age-DSK NMM 028 Bronze Age-DSK 11-12 (wear) MALE NMM 029 Bronze Age-DSK Khuvsgul, Alag-Erdene sum, Ulaan Tolgoi DS 4 feature 6 Tuv, Zaamar sum, Ulaan Khadnii Am 40 feature 3 Tuv, Zaamar sum, Ulaan Khadnii Am AD 40 feature 4 Tuv, Zaamar sum, Ulaan Khadnii Am AD 40 feature 5 Arkhangai, Undur-Ulaan sum, Jargalantyn Am F4 2011 Khanui Valley Khuvsgul, Alag-Erdene sum, Ulaan Tolgoi DS 5 feature 2 Khuvsgul, Renchinlkhumbe sum, Zeerdegchingiin Khushuu KS F1 (2006) Khuvsgul, Galt sum, Nukhtiin Am Mound 1 feature 1 (8-9),(78) MALE NMM 030 Bronze Age-DSK Arkhangai, Undoor Ulaan sum, Urt Bulagyn KS KYRI 21 11-16 (wear) MALE NMM 031 Bronze Age-DSK Khuvsgul, Galt sum, Nukhtiin Am Deer Stone Site Feature 1 11-13 (wear) LIKELY MALE NMM 034 Bronze Age-DSK Khuvsgul, Tumurbulag sum, Zunii Gol A1F3 11-14 (wear) Indet. NMM 035 NMM 036A NMM 036B Bronze Age-DSK Bronze Age-DSK Bronze Age-DSK Bulgan, Khutag-Undur sum, Uurgiin Gol 4 1-064-03 Bulgan, Khutag-Undur sum, Uurgiin Gol 4 1-064-04A Bulgan, Khutag-Undur sum, Uurgiin Gol 4 1-064-04B 10-16 (wear) 13-15 (wear) 11-13 (wear) LIKELY MALE LIKELY MALE LIKELY MALE NMM 037 Bronze Age-DSK Bulgan, Khutag-Undur sum, Uurgiin Gol 4 1-064-02 8-12 (wear) NMM 038 Bronze Age-DSK Bulgan, Khutag-Undur sum, Uurgiin Gol 4 1-064-05 8-12 (wear) 2.331 0.889 -0.569 + -0.131 UP2 bevel Bevel cause + (R) Yes + (L, R) Anthr opog (uppe r) 0 Yes 1.8 2.27 Nat. maloc c. 2.4 6 1.2 Nat. maloc c. Enamel exposure L R Bit mor phol ogy? No Enamel chip Notes L R + (L) + (L) Greaves effect (UL) No + (L) + (L) No - LIKELY MALE indeter minate + (L) No + (L) 0 indet. MALE 0 8-13 (wear) indet. 0 6-9 (wear) MALE 0.613 0 0 0 0 0 2.41 Indet. 0.702 8.5 24 5.26 Nat. maloc c. + (L) + (L) + (L) + (L) No No + (L) No + (L) + (L) + (L) 2.3 4 1.44 Nat. maloc c. + (U) 0 + (L) 0 0 Indet. 1.5 9 1.33 Indet. LIKELY MALE 0 1.61 Indet. 105 Concavity (LL) + (U) + (L) + (L) Yes No + (L) + (L) NMM 040 Bronze Age-DSK Khuvsgul, Tumurbulag sum, Zunii Gol K3 10-14 (wear) MALE 3.7 3 2.25 + (L, R) Anthr opog (uppe r) NMM 041 Bronze Age-DSK Khuvsgul, Tumurbulag sum, Zunii Gol DS 4 9-14 (wear) NO 1.6 4 1.29 + (L, R) Indet. NMM 042 Bronze Age-DSK Khuvsgul, Tumurbulag sum, Zunii Gol DS 7 NMM 043 Bronze Age-DSK Khuvsgul, Galt sum, Khushuutin Am F. 18 NMM 045 Bronze Age-DSK Bayankhongor, Erdenetsogt sum, Bor Shoroonii Am HM1 NMM 047 Bronze Age-DSK Khuvsgul, Renchinlkhumbe sum, Zeerdegchingiin Khushuu FA NMM 048 Bronze Age-DSK Khuvsgul, Renchinlkhumbe sum, Targan Nuur Feature 1 NMM 049 Bronze Age-DSK Uvs, Zuunkhangai sum, ZK 1-4 3.5-4 (eruptio n) 6-8 (wear) 14-17 (wear) 10-15 (wear) 11-12 (wear) 2.5-3 (eruptio n) 14-17 (wear) MALE MALE + (L) + (L) No 0.368 0 0 0 0.83 Nat. maloc c. No + (U, L) 0 + (L) Concavity (LL) No Indet. LIKELY MALE Uvs, Zuunkhangai sum, ZK 257-3 NMM 051 Bronze Age-DSK Uvs, Zuunkhangai sum, ZK 1-3 NMM 055 Bronze Age-DSK Uvs, Zuunkhangai sum, ZK2571 ~4 (eruptio n) Indet. NMM 056 Bronze Age-DSK Bayankhongor, Erdenetsogt sum, Shatar Chuluu DS 1 13-14 (wear) LIKELY MALE NMM 057 Bronze Age-DSK Bayankhongor, Erdenetsogt sum, Shatar Chuluu KS 1 6-11 (wear) LIKELY MALE NMM 058 Bronze Age-DSK Bulgan, Khutag-Undur sum, K75-3 LIKELY MALE NMM 059 Bronze Age-DSK Bulgan, Khutag-Undur sum, K75-2 NMM 062 Bronze Age-DSK Bulgan, Khutag-Undur sum,K75-4 6-10 (wear) 3-4 (eruptio n) ~10 (root morph.) NMM 066 NMM 067 Bronze Age-DSK Bronze Age-DSK NMM 083 Bronze Age-DSK NMM 084 Bronze Age-DSK NMM 090 NMM 091 Bronze Age-DSK Bronze Age-DSK Uvs, Zuunkhangai sum, Khurdleyiin Am Bayankhongor, Shatar Chuluu DS 2 Khuvsgul aimag, Alag-Erdene sum, Khushuutiin Devseg DS F-3 Khuvsgul aimag, Alag-Erdene sum, Khushuutiin Devseg DS F-1 Tuv, Zaamar sum, Ulaan Khadnii Am AD 41 feature 3 Tuv, Zaamar sum, Ulaan Khadnii Am AD 41 feature 2 0 0.131 -0.131 No LIKELY MALE 0 2.14 Indet. 0 0 2.6 4 1.37 Possibl e 0.332 0 1.98 1.53 LIKELY MALE Indet. Indet. Indet. 7-10 (wear) Indet. Nat. maloc c. No Nat. maloc c. Nat. maloc c. Indet. + (U) No No No LIKELY MALE 9-14 (wear) 7-9 (wear) ~5 (eruptio No Indet. Bronze Age-DSK 7-8 (wear) 9-10 (wear) + (U) + (L) LIKELY MALE NMM 050 20+ (wear) No Indet. Indet. 106 + (L) NMM 092 Bronze Age-DSK Bulgan, Khutag-Undur, Uurgiin Gol Khirisguur 75/76 NMM 093 Bronze Age-DSK Bulgan, Khutag-Undur sum,K76-7 IMH 2008.1 01 IMH Hanov er IMH Indraff IMH Lexing ton IMH Syson by MSB 19370 4 NMNH 17245 4 NMNH 17496 0 NMNH 17496 8 NMNH 27090 0 NMNH 843 CMH 1 CMH 2 CMH 3 Contemp . America n Contemp . America n Contemp . America n Contemp . America n Contemp . America n Contemp . America n Contemp . America n Contemp . America n Contemp . America n Contemp . America n Contemp . America n Contemp . Mongolia n Contemp . Mongolia n Contemp . Mongolia n Lexington, KY Lexington, KY n) 7-10 (wear) 5-9 (eruptio n/wear) 18+ (wear) LIKELY MALE 1.9 6 1.68 Indet. 3.1 9 3.43 Nat. maloc c. + (L) No Indet. LIKELY MALE 25 (hist. records) 1.022 1.172 -0.15 - 0.291 0.278 0.013 - 1.182 0.91 0.272 - 0.507 0.484 0.023 - NA 0.637 NA - 0.862 0.684 0.178 - 0.259 0.314 -0.055 1.103 1.169 -0.066 0.351 0.348 0.003 0.793 0.821 -0.028 0.282 0.309 -0.027 0.279 0.552 -0.273 - 0.214 1.203 -0.989 + 0.522 0.592 -0.07 - MALE Lexington, KY 25 (hist. records) MALE Lexington, KY 25 (hist. records) MALE Lexington, KY Albuquerque, NM roadside find (gelding) Haleb Percheron Stallion US College of Veterinary Surgeons, Washington DC Kidron Baird Farm, Carlisle, PA Ulaanbaatar Bayankhongor, Gurvanbulag sum Tuv, Khustai Nuruu 4 (hist. records) ~10-11 (wear) MALE MALE 8 (hist. records) MALE ~6 (wear) MALE 18+ (wear) FEMALE 35 (hist. records) MALE ~9 (wear) MALE ~4 (eruptio n) FEMALE 18+ (wear) MALE 5+ (eruptio n) MALE 107 + (L) Yes + (U, L) + (U, L) Pronounced concavity to LLP2 anterior margin CMH 4 CMH 5 CMH 6 CMH 7 CMH 8 CMH 9 CMH 10 CMH 11 CMH 12 CMH 13 CMH 14 CMH 15 KNNP Przew alski MSB 54743 NMNH 23811 1 NMNH 31103 3 Contemp . Mongolia n Contemp . Mongolia n Contemp . Mongolia n Contemp . Mongolia n Contemp . Mongolia n Contemp . Mongolia n Contemp . Mongolia n Contemp . Mongolia n Contemp . Mongolia n Contemp . Mongolia n Contemp . Mongolia n Contemp . Mongolia n Contemp . Przewals ki Contemp . Przewals ki Contemp . Przewals ki Contemp . Przewals Tuv, Gun-Galuut 16+ (wear) MALE 0.326 0.639 -0.313 - Tuv, Terelj (near Erdene sum) ~4.5 (eruptio n) MALE 0.401 0.272 0.129 - Tuv, Terelj 5+ (eruptio n) MALE 0.641 0.537 0.104 - Tuv., Terelj 13-16 (wear) MALE 0.561 0.686 -0.125 + Tuv, Khustai Nuruu 11-12 (wear) MALE 0.413 0.763 -0.35 - Uvurkhangai, Bat-Ulzii 16+ (wear) FEMALE 0 0 0 - Uvurkhangai, Bat-Ulzii 5+ (eruptio n) FEMALE 0 0 0 + Uvurkhangai, Bat-Ulzii ~7 (wear) FEMALE NA NA NA Uvurkhangai, Bat-Ulzii 4 (eruptio n) MALE 0.395 0.271 0.124 + Uvurkhangai, Bat-Ulzii 5+ (eruptio n) MALE 0 0 0 - Gobi-Altai, Biger 5+ (eruptio n) MALE NA NA NA Tuv, Khustai Nuruu 8-9 (wear) MALE 0.422 0.627 -0.205 Possibl e Tuv, Khustai Nuruu ~6 (wear) MALE 0 0 0 - Albuquerque, NM ~5-6 (wear) MALE 0.289 0.194 0.095 - Washington, D.C - National Zoological Park ~5-6 (wear) MALE 0 0 0 - Washington, D.C - National Zoological Park 18+ (wear) FEMALE 0.188 0.248 -0.06 - 108 + (L) + (L) + (L) + (L) 0 + (L) + (L) 0 + (L) 1.5 3.13 0 0 2.5 9 1.71 Nat. maloc c. + (L) Yes Yes + (L) + (L) Concavity to lower LP2 anterior margin Bowing out of lower teeth No No + (L) Yes Occlusal concavity LLP2 NMNH 58208 8 NMNH 58246 7 NMNH 58291 0 MSB 14653 7 MSB 15058 7 NMNH 26893 8 NMNH 30289 8 NMNH 39543 2 NMNH 29518 0 ki Contemp . Przewals ki Contemp . Przewals ki Contemp . Przewals ki Feral America n Feral America n Feral America n Feral America n Feral America n Feral America n Washington, D.C - National Zoological Park 6 (hist. Records) FEMALE 0.342 0 0.342 - Washington, D.C - National Zoological Park ~8 (wear) FEMALE 0.086 0.144 -0.058 - Washington, D.C - National Zoological Park ~7 (wear) MALE 0.754 0.678 0.076 - Albuquerque, NM ~5 (eruptio n/wear) FEMALE 0 0 0 - Albuquerque, NM ~6 (wear) FEMALE 0 0 0 - Chincoteague Island 7-8 (wear) FEMALE 0.238 0.301 -0.063 - Assateague Island 10-11 (wear) FEMALE 0.182 0.15 0.032 - Assateague Island 18+ (wear) FEMALE 0.302 0 0.302 - Assateague Island ~5 (eruptio n/wear) Indet. 0.61 0.45 0.138 14-15 (wear) MALE 1.306 0.95 0.356 NMM 013 Pazyryk Bayan Ulgii, Tsengel sum, Khuiten Gol Delta 2-1 Pazyryk burial (Biluut 2-1) NMM 071 Xiongnu Khovd, Mankhan sum, Takhilt THL 64.65 Royal Tomb 9-11 (wear) MALE 0.667 0.856 -0.189 Possibl e 6.8 6 5.37 NMM 080 Xiongnu Bulgan aimag, Dashinchilen sum, Elst-Ar, Burial 11 15+ (wear) FEMALE 0 0.911 -0.911 Possibl e 0 0 NMM 094 Xianbei Orkhon, Jargalant sum, Airagiin Gozgor, Burial 84 8-12 (wear) MALE 0.57 NA NA - 0.6 6 0.79 NMM 011 NMM 081 NMM 082 NMM 087 Turkic Khaganat e Turkic Khaganat e Turkic Khaganat e 10th Century/ Khitan 7.3 5 Bayan Ulgii, Tsengel sum, Khuiten Gol Delta Turkic burial 8-9 (wear) MALE 1.03 0.852 0.178 Bulgan aimag, Dashinchilen sum, Dundgozgor, Burial 10 8-13 (wear) MALE NA NA NA MALE 0.714 0 0.714 - MALE 0.206 NA NA + Arkhangai aimag, Khotont sum, Ovur Khartsaliin Am Khovd, Myangad sum, UlaanUnet Cave Burial 4-5 (eruptio n) 4-4.5 (eruptio n) 109 + + (L, R) Anthr opog + (L, R) Anthr opog Nat. maloc c. 9.8 1 8.22 + (L, R) Anthr opog 6.8 3 4.84 Indet. 3.1 2 0.77 Nat. maloc c. + (U) Yes + (U) + (U) + (L) + (U, L) + (L) + (L) + (L) + (L) Greaves effect (lower) Yes Yes Concavity (LR) + (L) Yes No + (L) + (L) Yes + (U) APPENDIX V. RADIOCARBON DATES FROM DEER STONE AND KHIRIGSUUR ARCHAEOLOGICAL SITES (CH.6) ID Khuvsgul aimag, Alag-Erdene sum, Ulaan Tolgoi DS4 F2 Khuvsgul aimag, Alag-Erdene sum, Ulaan Tolgoi DS4 F3 Khuvsgul aimag, Alag-Erdene sum, Ulaan Tolgoi DS4 F5 Khuvsgul aimag, Alag-Erdene sum, Ulaan Tolgoi DS4 F6 Khuvsgul aimag, Alag-Erdene sum, Khyadag W DS1 F1 Khuvsgul aimag, Alag-Erdene sum, Ulaan Tolgoi DS 4 F1 Khuvsgul aimag, Bayanzurkh sum, Khorigiin Am F1 Arkhangai aimag, Bayantsagaanii Khundii, Deer Stone 38 Feature 1 Arkhangai aimag, Bayantsagaanii Khundii, Deer Stone 38 Feature 95 Arkhangai aimag, Khavtsaliin Am, Deer Stone 40 Feature 4 Arkhangai aimag, Ulaan Tolgoin Ar Shil, Deer Stone 62 Feature 17 Uvurkhangai aimag, Khujirt sum, Khirigsuur 12.3 Uvurkhangai aimag, Khujirt sum, Khirigsuur 4.32 Khuvsgul aimag, 54.22 Khuvsgul aimag, S49.2 Khuvsgul aimag, Renchinlkhumbe sum, Tstatstain Khushuu DS1 F2 Bayankhongor aimag, Erdenetsogt sum, Bor Shoroonii Am DS F1 Bayankhongor aimag, Erdenetsogt sum, Shatar Chuluu 14 C Date (BP) Uncertainty σ 2950 40 Horse bone Deer Stone Satellite mound Fitzhugh and Bayarsaikhan 2009 B-193740 AMS 2810 40 Horse bone Deer Stone Satellite mound Fitzhugh and Bayarsaikhan 2009 B-207205 RAD 2790 70 Horse bone Deer Stone Satellite mound Fitzhugh and Bayarsaikhan 2009 B-207206 RAD 2740 70 Horse bone Deer Stone Satellite mound Fitzhugh and Bayarsaikhan 2009 B-246623 AMS 2610 40 Horse bone Deer Stone Satellite mound Fitzhugh and Bayarsaikhan 2009 B-193738 AMS 2530 40 Horse bone Deer Stone Satellite mound Fitzhugh and Bayarsaikhan 2009 AA106947 2438 33 Horse bone Deer Stone Satellite mound This study B-323806 2660 30 Horse bone Deer Stone Satellite mound Gantulga et al. 2016 B-323808 2580 30 Horse bone Deer Stone Satellite mound Gantulga et al. 2016 B-389402 2780 30 Horse bone Deer Stone Satellite mound Gantulga et al. 2016 B-389401 2880 30 Horse bone Deer Stone Satellite mound Gantulga et al. 2016 COL2032.1.1 2887 38 Horse bone Khirigsuur Satellite mound Yeruul-Erdene et al. 2015 KIA-49219 2880 25 Horse bone Khirigsuur Satellite mound Yeruul-Erdene et al. 2015 Not reported 2843 41 Horse bone Khirigsuur Not reported 2730 50 Horse bone Khirigsuur B-207207 AMS 3000 40 Horse tooth Deer Stone Satellite mound Fitzhugh and Bayarsaikhan 2009 AA106948 2977 30 Horse tooth Deer Stone Satellite mound This study AA106950 2953 31 Horse tooth Deer Stone Satellite mound This study 14 C Date Ref B-193739 AMS Sample Material 110 Monument type Context Satellite mound Satellite mound Reference Frohlich et al 2009 Frohlich et al 2009 DS 1 Khuvsgul aimag, Renchinlkhumbe sum, Tstatstain Khushuu DS 1 F1 Khuvsgul aimag, Tsagaan Uul sum, Khushuutiin Gol A3 F3 Khuvsgul aimag, Galt sum, Khushuutiin Am F18 Khuvsgul aimag, Tumurbulag sum, Zunii Gol A1 F3 Khuvsgul aimag, Tumurbulag sum, Zunii Gol A3 F1 Bayankhongor aimag, Erdenetsogt sum, Shatar Chuluu DS 2 Khuvsgul aimag, Burentogtokh sum, Ulaan Ushig 5 (Kh1 SH-11) Khuvsgul aimag, Alag-Erdene sum, Ulaan Tolgoi DS5 F2 Khuvsgul aimag, Galt sum, Nukhtiin Am DS1/2 F1 Khuvsgul aimag, Alag-Erdene sum, Ulaan Tolgoi DS5 F1 Khuvsgul aimag, Renchinlkhumbe sum, Targan Nuur F1 Khuvsgul aimag, Shin Ider sum, Tsokhiotin Am A1 DS2 F2 Khuvsgul aimag, Tsagaan Uul sum, Khushuutiin Gol F2 Khuvsgul aimag, Tumurbulag sum, Zunii Gol A2 DS4 Zavkhan aimag, Telmen sum, Ogomoor DS8 F1 Khuvsgul aimag, Alag-Erdene sum, Khushuutiin Devseg F3 Arkhangai aimag, Undur-Ulaan sum, Jargalantyn Am DS Khuvsgul aimag, Tsagaan Uul sum, Bor Hujiriin Gol A1 F2 Zavkhan aimag, Shiluustei sum, Daagan Del F1 Khuvsgul aimag, B-207208 AMS 2920 40 Horse tooth Deer Stone Satellite mound Fitzhugh and Bayarsaikhan 2009 B-246618 AMS 2910 40 Horse tooth Deer Stone Satellite mound Fitzhugh and Bayarsaikhan 2009 B-272763 AMS 2880 40 Horse tooth Deer Stone Satellite mound Fitzhugh and Bayarsaikhan 2009 B-272756 AMS 2870 40 Horse tooth Deer Stone Satellite mound Fitzhugh and Bayarsaikhan 2009 B-272758 AMS 2860 40 Horse tooth Deer Stone Satellite mound Fitzhugh and Bayarsaikhan 2009 AA106953 2846 30 Horse tooth Deer Stone Satellite mound This study MTC-12817 2835 57 Horse tooth Deer Stone Satellite mound Toshio 2013 B-222535 AMS 2830 40 Horse tooth Deer Stone Satellite mound Fitzhugh and Bayarsaikhan 2009 B-222534 AMS 2830 40 Horse tooth Deer Stone Satellite mound Fitzhugh and Bayarsaikhan 2009 B-215694 AMS 2800 40 Horse tooth Deer Stone Satellite mound Fitzhugh and Bayarsaikhan 2009 AA106946 2800 31 Horse tooth Deer Stone Satellite mound This study B-272760 AMS 2790 40 Horse tooth Deer Stone Satellite mound Fitzhugh and Bayarsaikhan 2009 B-246617 AMS 2750 40 Horse tooth Deer Stone Satellite mound Fitzhugh and Bayarsaikhan 2009 B-272757 AMS 2710 40 Horse tooth Deer Stone Satellite mound Fitzhugh and Bayarsaikhan 2009 B-363204 2710 30 Horse tooth Deer Stone Satellite mound This study B-243716 AMS 2680 40 Horse tooth Deer Stone Satellite mound Fitzhugh and Bayarsaikhan 2009 Not reported 2670 30 Horse tooth Deer Stone Satellite mound J. Bayarsaikhan, unpublished B-246614 AMS 2640 40 Horse tooth Deer Stone Satellite mound Fitzhugh and Bayarsaikhan 2009 B-363203 2620 30 Horse tooth Deer Stone Satellite mound This study B-240690 2610 40 Horse Deer Stone Stone Fitzhugh and 111 Alag-Erdene sum, Khyadag E DS pav 7 Khuvsgul aimag, Alag-Erdene sum, Khyadag E A3 F32 Khuvsgul aimag, Alag-Erdene sum, Khushuutiin Devseg F2 Khuvsgul aimag, Alag-Erdene sum, Khushuutiin Devseg F1 Khuvsgul aimag, Burentogtokh sum, Ulaan Ushig 3 (SC 5) Khuvsgul aimag, Burentogtokh sum, Ulaan Ushig 1 (SC 7) Uvs aimag, Zuunkhangai sum, ZK-1-1 Bayankhongor aimag, Erdenetsogt sum, Shatar Chuluu KS1 Khuvsgul aimag, Alag-Erdene sum, Ulaan Tolgoi KS M1 F3 tooth Khuvsgul aimag, Renchinlkhumbe sum, Zeerdegchingiin Khoshuu FA Bayan-Ulgii aimag, Tsengel sum, On Khad Khushuu Uvs aimag, Zuunkhangai sum, ZK-1-3 Bulgan aimag, Khutag-Undur sum, Uurgiin Gol KS 64-4 Khuvsgul aimag, Alag-Erdene sum, Ulaan Tolgoi KS Mound 1 F2 tooth Uvs aimag, Zuunkhangai sum, ZK-257-4 Uvs aimag, Zuunkhangai sum, ZK-257-1 Arkhangai aimag, Khanuy Valley, Urt Bulagyn KYRI 22 tooth Arkhangai aimag, Undur-Ulaan sum, Urt Bulagyn KYRI 21 tooth Khuvsgul aimag, Galt sum, Nukhtiin Am M1 F1 tooth Khuvsgul, Zunii Gol, K3 F42 AMS tooth pavement Bayarsaikhan 2009 B-246620 AMS 2520 40 Horse tooth Deer Stone Satellite mound Fitzhugh and Bayarsaikhan 2009 B-240688 AMS 2450 40 Horse tooth Deer Stone Satellite mound Fitzhugh and Bayarsaikhan 2009 B-243716 AMS 2410 40 Horse tooth Deer Stone Satellite mound Fitzhugh and Bayarsaikhan 2009 MTC-12815 2923 59 Unspec. horse Deer Stone Satellite mound Toshio 2013 MTC-12531 2749 50 Unspec. horse Deer Stone Satellite mound Toshio 2013 AA106955 2963 31 Horse tooth Khirigsuur Satellite mound Houle 2016 AA106951 2955 31 Horse tooth Khirigsuur Satellite mound This study B-215693 AMS 2950 60 Horse tooth Khirigsuur Satellite mound Fitzhugh and Bayarsaikhan 2009 AA106945 2934 31 Horse tooth Khirigsuur Satellite mound This study B-246613 AMS 2930 40 Horse tooth Khirigsuur Satellite mound Fitzhugh and Bayarsaikhan 2009 AA106956 2922 30 Horse tooth Khirigsuur Satellite mound Houle 2016 AA106952 2903 30 Horse tooth Khirigsuur Satellite mound This study B-215692 AMS 2860 40 Horse tooth Khirigsuur Satellite mound Fitzhugh and Bayarsaikhan 2009 AA106958 2850 29 Horse tooth Khirigsuur Satellite mound Houle 2016 AA106957 2836 30 Horse tooth Khirigsuur Satellite mound Houle 2016 B-222533 AMS 2790 40 Horse tooth Khirigsuur Satellite mound Fitzhugh and Bayarsaikhan 2009 B-222532 AMS 2780 50 Horse tooth Khirigsuur Satellite mound Fitzhugh and Bayarsaikhan 2009 B-240685 AMS 2630 40 Horse tooth Khirigsuur Satellite mound Fitzhugh and Bayarsaikhan 2009 B-272759 AMS 2950 40 Horse tooth Khirigsuur Satellite mound Fitzhugh and Bayarsaikhan 2009 112 Dundgovi aimag, Baga Gazaryn Chuluu, EX 04.04 Khuvsgul aimag, Burentogtokh sum, Ulaan Ushig 2 (Kh1 SH-18) Bulgan aimag, Tarvagtai Valley, Site 2 Feature 2 Bulgan aimag, Tarvagtai Valley, Site 2 Feature 1 OS-71705 3040 35 Unspec. horse Khirigsuur Satellite mound Amartuvshin and Jargalan 2010 MTC-12814 2871 59 Unspec. horse Khirigsuur Satellite mound Toshio 2013 UG-18415 2810 25 Unspec. horse Khirigsuur Satellite mound B. Jargalan, unpublished UG-18414 2780 20 Unspec. horse Khirigsuur Satellite mound B. Jargalan, unpublished Khuvsgul aimag, 24 Not reported 3174 53 Khuvsgul aimag, 51 Not reported 3086 41 Khuvsgul aimag, 18 Not reported 3074 49 Khuvsgul aimag, 58 Not reported 3056 46 Khuvsgul aimag, 25 Not reported 3052 50 Khuvsgul aimag, 23 Not reported 3052 52 Khuvsgul aimag, 40 Not reported 3052 51 Khuvsgul aimag, 3 Not reported 3044 50 Khuvsgul aimag, 22 Not reported 3033 49 Khuvsgul aimag, 17 Not reported 3029 49 Khuvsgul aimag, 10 Not reported 2992 48 Khuvsgul aimag, 9 Not reported 2991 48 Khuvsgul aimag, 55 Not reported 2990 38 Khuvsgul aimag, 13 Not reported 2989 48 Khuvsgul aimag, 41 Not reported 2958 42 Khuvsgul aimag, 16 Not reported 2930 50 Khuvsgul aimag, 44 Not reported 2918 51 Khuvsgul aimag, 2 Not reported 2910 52 Khuvsgul aimag, S49 Not reported 2900 50 Khuvsgul aimag, 7 Not reported 2897 55 Khuvsgul aimag, 8 Not reported 2872 48 Khuvsgul aimag, 43 Not reported 2862 51 Khuvsgul aimag, 6 Not reported 2857 54 Khuvsgul aimag, 14 Not reported 2849 49 Khuvsgul aimag, 52 Not reported 2842 42 Khuvsgul aimag, 27 Not reported 2835 50 Khuvsgul aimag, 54 Not reported 2831 41 Human bone Human bone Human bone Human bone Human bone Human bone Human bone Human bone Human bone Human bone Human bone Human bone Human bone Human bone Human bone Human bone Human bone Human bone Human bone Human bone Human bone Human bone Human bone Human bone Human bone Human bone Human bone 113 Khirigsuur Khirigsuur Khirigsuur Khirigsuur Khirigsuur Khirigsuur Khirigsuur Khirigsuur Khirigsuur Khirigsuur Khirigsuur Khirigsuur Khirigsuur Khirigsuur Khirigsuur Khirigsuur Khirigsuur Khirigsuur Khirigsuur Khirigsuur Khirigsuur Khirigsuur Khirigsuur Khirigsuur Khirigsuur Khirigsuur Khirigsuur Human burial Human burial Human burial Human burial Human burial Human burial Human burial Human burial Human burial Human burial Human burial Human burial Human burial Human burial Human burial Human burial Human burial Human burial Human burial Human burial Human burial Human burial Human burial Human burial Human burial Human burial Human burial Frohlich et al 2009 Frohlich et al 2009 Frohlich et al 2009 Frohlich et al 2009 Frohlich et al 2009 Frohlich et al 2009 Frohlich et al 2009 Frohlich et al 2009 Frohlich et al 2009 Frohlich et al 2009 Frohlich et al 2009 Frohlich et al 2009 Frohlich et al 2009 Frohlich et al 2009 Frohlich et al 2009 Frohlich et al 2009 Frohlich et al 2009 Frohlich et al 2009 Frohlich et al 2009 Frohlich et al 2009 Frohlich et al 2009 Frohlich et al 2009 Frohlich et al 2009 Frohlich et al 2009 Frohlich et al 2009 Frohlich et al 2009 Frohlich et al 2009 Khuvsgul aimag, 1 Uvs aimag, Zuunkhangai, SKTB-1 Bayan-Ulgii aimag, Tsengel sum, Khuiten Gol Delta 2 KS Bayankhongor aimag, Erdenetsogt sum, Bor Shoroonii Am KS 1 Khuvsgul aimag, Renchinlkhumbe sum, Evdt 2 DS 2 Bayan-Ulgii aimag, Sagsai sum, Tsagaan Asga F4 Khuvsgul aimag, Alag-Erdene sum, Ulaan Tolgoi DS4 S7 Bayan Ulgii aimag, Sagsai sum, Tsagaan Asga DS F3 Khuvsgul aimag, Tsagaan Uul sum, Khushuutiin Gol F6 Bayan Ulgii aimag, Tsengel sum, East Bay 3 DS 2 Bayan Ulgii aimag, Tsengel sum, Biluut 1C-F1 Bayan Ulgii aimag, Tsengel sum, Khoton Nuur East Bay 1 F7 Khuvsgul aimag, Renchinlkhumbe sum, Hort Azuur DS2 L2 F1 Khuvsgul aimag, Renchinlkhumbe sum, Avtiin Fea 5 Sample 6 Khuvsgul aimag, Tsagaan Uul sum, Bor Hujiriin Gol A2 F1 Dundgovi aimag, Baga Gazaryn Chuluu, EX 07.24 Arkhangai aimag, Shivertiin Am, Deer Stone 33 Feature 6 Uvurkhangai aimag, Khujirt sum, Khirigsuur 4.11 Arkhangai aimag, Ikh Tamir, Monument 341 Not reported 2779 50 Human bone Khirigsuur Human burial Frohlich et al 2009 AA106959 3142 30 Human tooth Khirigsuur Human burial Houle 2016 B-334573 2800 30 Human tooth Khirisguur Human burial Fitzhugh et al. 2013 AA106949 2871 31 Sheep tooth Khirigsuur Satellite mound This study B-215643 AMS 3030 40 Charcoal Deer Stone Stone circle Fitzhugh and Bayarsaikhan 2009 B-246612 AMS 3000 40 Charcoal Deer Stone Stone circle Fitzhugh and Bayarsaikhan 2009 B-182959 AMS 2930 40 Charcoal Deer Stone Stone circle Fitzhugh and Bayarsaikhan 2009 B-246611 AMS 2850 40 Charcoal Deer Stone Stone circle Fitzhugh and Bayarsaikhan 2009 B-246619 AMS 2850 40 Charcoal Deer Stone Satellite mound Fitzhugh and Bayarsaikhan 2009 B-306040 2830 30 Charcoal Deer Stone Stone circle Fitzhugh and Kortum 2012 B-306033 2820 40 Charcoal Deer Stone Stone circle Fitzhugh and Kortum 2012 B-306039 2750 30 Charcoal Khirigsuur Stone circle Fitzhugh and Kortum 2012 B-240691 AMS 2710 40 Charcoal Deer Stone Satellite mound Fitzhugh and Bayarsaikhan 2009 B-242730 AMS 2670 40 Charcoal Deer Stone Stone circle Fitzhugh and Bayarsaikhan 2009 B-246616 RAD 2670 50 Charcoal Deer Stone Stone circle Fitzhugh and Bayarsaikhan 2009 OS-68948 3060 35 Unspec. bone Khirigsuur Satellite mound Amartuvshin and Jargalan 2010 B-323805 2910 30 Unspec. bone Deer Stone Stone circle Gantulga et al. 2016 KA-49218 2830 40 Unspec. bone Khirigsuur Stone circle Yeruul-Erdene et al. 2015 B-290944 2810 40 Unspec. bone Deer Stone Stone circle Gantulga et al. 2016 114 REFERENCES Academy of Equine Dentistry. 2013. Teeth of the Horse- Chart for Accurately Telling the Age from Six Months to Twenty-Nine Years. Academy of Equine Dentistry, Glenns Ferry, ID. Allard, Francis,and Erdenebaatar, Diimaajav. 2005. Khirigsuurs, ritual and mobility in the Bronze Age of Mongolia. Antiquity 79, 547-563. Barfield, Thomas. 2011. Nomadic pastoralism in Mongolia and beyond. In: Sabloff, Paula (Ed.), Mapping Mongolia: Situating Mongolia in the World from Geologic Time to Present. University of Pennsylvania Museum of Archaeology and Anthropology, Philadelphia. Allard, F., Erdenebaatar, D., Olsen, S. Cavalla, A., and Maggiore, E. 2007. Ritual horses in Bronze Ageand present day Mongolia: some preliminaryobservations from Khanuy Valley, in L. Popova, C.Hartley & A. Smith (ed.) Social orders and sociallandscapes (Proceedings of the 2005 University ofChicago Conference on Eurasian Archaeology). Newcastleupon-Tyne: Cambridge Scholars. Allen, Tom. 2008. Manual of Equine Dentistry. Muleicorn Press, New York. Anthony, David. 2007. The horse, the wheel, and language: how Bronze-Age riders from the Eurasian steppes shaped the modern world. Princeton (NJ): Princeton University Press. Anthony, David, and Brown, Dorcas. 1998. Bit wear, horseback riding, and the Botai site in Kazakhstan. Journal of Archaeological Science 25: 331–47. http://dx.doi.org/10.1006/jasc.1997.0242 2003. Eneolithic horse rituals and riding in the Steppes: new evidence, in M. Levine (ed.) Prehistoric steppe adaptation and the horse. Cambridge: McDonald Institute for Archaeological Research. 2011. The Secondary Products Revolution, Horse-Riding, and Mounted Warfare. J World Prehist 24: 131–160 Anthony, D. Telgin, D. and Brown, D. 115 1991. The origin of horseback riding. Scientific American 265(6): 94–99. http://dx.doi.org/10.1038/scientificamerican1291-94 Anthony, D., Brown, D., and George, C. 2006. Early horseback riding and warfare: the importance of the magpie around the neck, in S. Olsen, S. Grant, A. Choyke & L. Bartosiewicz (ed.) Horses and humans: the evolution of the equine-human relationship (British Archaeological Reports International series 1560). Oxford: Archaeopress. Argent, Gala. 2011. At Home With the Good Horses: Relationality, Roles, Identity and Ideology in Iron Age Inner Asia. Ph.D. Dissertation, University of Leicester, Leicester, UK. Bar-Oz, G,, Nahshoni, P., Motro, H., and Oren, E. 2013. Symbolic Metal Bit and Saddlebag Fastenings in a Middle Bronze Age Donkey Burial. PLOSOne 8(3):1-7 Bartosiewicz, L. 2014. Shuffling Nags, Lame Ducks: The Archaeology of Animal Disease. Oxbow Books, Oxford. Batsaikhan, Ts. And Baatarsuren, Ch. 2015. Morin Soyoliin Tonog Heregsel. Munkhiin Useg, Ulaanbaatar, Mongolia. Bayarsaikhan, Jamsranjav. 2006. Bugan Chuluun Khushuu Tusliin 2006-onii Xeeriin Sudaalgaani Tailan. National Museum of Mongolia, Ulaanbaatar. 2008. Bugan Chuluun Khushuu Tusliin 2008-onii Xeeriin Sudaalgaani Tailan. National Museum of Mongolia, Ulaanbaatar. 2011. Excavations at Jargalantyn Am Deer Stone Site, 2011 Report. National Museum of Mongolia, Ulaanbaatar. 2016. Deer Stones of Northern Mongolia. Ph.D. Dissertation, National University of Mongolia, Ulaanbaatar, Mongolia. Bayarsaikhan, Jamsranjav, and Tuvshinjargal, Tumuurbaatar. 2013. Zavkhan Aimagiin Telmen Sumiin Uguumur Bagiin Nutag Dund Shurgaxiin Amnii Bugan Khushuut Dursgalt Gazart Xiisen Arkheologiin Sudalgaa. National Museum of Mongolia, Ulaanbaatar. Bendrey, R. 116 2007. New methods for the identification of evidence for bitting on horse remains from archaeological sites. Journal of Archaeological Science 34: 1036–50. http:// dx.doi.org/10.1016/j.jas.2006.09.010 2008. An analysis of factors affecting the development of an equine cranial enthesopathy. Veterinarija Ir Zootechnika 41(63): 25–31. Bendrey, R., Cassidy, J.P, Bokovenko, N., Lepetz, S., and Zaitseva, G.I. . A possi le ase of Poll-E il i a ea l “ thia ho se skull f o Arzhan 1, Tuva Republic, Central Asia. International Journal of Osteoarchaeology 21:111–18. http:// dx.doi.org/10.1002/oa.1099 Bendrey, R., Thorpe, N., Outram, A., and van Widjngaarden-Bakker, L. H. 2013. The Origins of Domestic Horses in North-west Europe: New Direct Dates on the Horses of Newgrange, Ireland. Proceedings of the Prehistoric Society 79:91–103. Beardsley, Richard K., 1953. Hypotheses on Inner Asian Pastoral Nomadism and its culture area. Memoirs of the Society for American Archaeology 9:24-28 (Asia and North America: Transpacific Contacts). Benecke, N. 2007. The horse skeletons from the Scythian royal grave mound at Arzhan 2. Documenta Archaeobiologiae, Skeletal Series and their Socioeconomic Context:115-131. Borgerhoff-Mulder, M., Fazzio, I., Irons, W., McElreath, R., Bowles, S., Bell, A., Hertz, T., and Hazzah, L. 2010. Pastoralism and Wealth Inequality: Revisiting an old question. Current Anthropology 51(1):35-48 Bokovenko, N. A. 2000. The Origins of Horse Riding and the Development of the Central Asian Nomadic Riding Harness. In Kurgans, Ritual Sites, and Settlements: Eurasian Bronze and Iron Age, J. DavisKimball, E. Murphy, L. Koryakova, and L. Yablonsky (eds.), pp. 304–307. Archaeopress, Oxford. Bold, Bat-Ochir. 2012. Eques Mongolica. Bold & Bodi. Reykjavik. Broderick, L.G., Houle, J.L., Seitsonen, O., Bayarsaikhan, J. 2014. The Mystery of the Missing Caprines: Stone Circles at the Great Khirigsuur, Khanuy Valley. In: Arkheologiin Suudal XXXIV:13. Mongolian Academy of Sciences, Ulaanbaatar. 117 Bronk Ramsey, Christopher. 2009. Bayesian Analysis of Radiocarbon Dates. Radiocarbon 51(1):337-360. Bronk Ramsey Christopher, and Lee, S. 2013. Recent and planned developments of the program OXCAL. Radiocarbon 55(2-3):720-730. Brownrigg, G. 2006. Horse control and the bit, in S. Olsen, S. Grant, A. Choyke & L. Bartosiewicz (ed.) Horses and humans: the evolution of the equine-human relationship (British Archaeological Reports International series 1560). Oxford: Archaeopress. Cai, D., Tang, Z., Han, L., Speller, C.F., Yang. D.Y., and Ma, X. 2009. Ancient DNA provides new insights into the origin of the Chinese domestic horse.Journal of Archaeological Science 36: 835–42. http://dx.doi.org/10.1016/j.jas.2008.11.006 Casey, R.A. 2002. Clinical problems associated with the intensive management of performance horses, in N. Waran (ed.) Animal welfare volume 1: welfare of horses. Netherlands: Springer. Cavalli-Sforza, L. 1996. The spread of agriculture and nomadic pastoralism: insights from genetics, linguistics, and archaeology. In The origins and spread of agriculture and pastoralism in Eurasia, D. Harris (ed.), pp. 51-69. UCL Press, London. Chang, Claudia, and Koster, Harold. 1986. Beyond bones: toward an archaeology of pastoralism. Advances in Archaeological Method and Theory 9:97-148. Christian, David. 2000. Silk roads or steppe roads? The silk roads in world history. Journal of World History 11(1): 1-26. http:// dx.doi.org/10.1353/jwh.2000.0004 Clark, Julia. 2014. Modeling Late Prehistoric and Early Pastoral Adaptations in Northern Mongolia's Darkhad Depression. Ph.D. dissertation, University of Pittsburgh. Clayton, H. and Lee, R. 1984. A Flouroscopic Study of the Position and Actio of the Joi ted “ affle Bit i the Ho se s Mouth. Equine Veterinary Science 4(5):193-196. Cooke, B. 118 2000. Imperial China: The Art of the Horse in Chinese History. International Museum of the Horse, Lexington. Cribb, Roger. 2004. Nomads in Archaeology. Cambridge University Press, Cambridge. d Alpoi Guedes, J., H. Lu, A. Hei , a d A. “ h idt. 2015. Early evidence for the use of wheat and barley as staple crops on the margins of the Tibetan Plateau. Proceedings of the National Academy of Sciences 112(18):5625-5630. Dee, M., Wengrow, D., Shortland, A., Stevenson, A., Brock, F., Link, LG, Bronk Ramsey, C. 2013. An absolute chronology for early Egypt using Bayesian statistical modeling. Proceedings of the Royal Society A 469(2159). Der Sarkissian, Clio, Ermini, L., Schubert, M., Yang, M., Librado, P., Fumagalli, M., Jonsson, H., Bar-Gal, G.K., Albrechtsen, A., Vieira, F.G., Peterson, B., Ginolhac, A., Seguin-Orlando, A., Magnussen, K., Fages, A., Gamba, C., Lorente-Galdos, B., Polani, S., Steiner, C., Neuditschko, M., Jagannathan, V., Feh, C., Greenblatt, C.L., Ludwig, A., Abramson, N.I., Zimmermann, W., Schafberg, R., Tikhonov, A., Sicheritz-Ponten, T., Willerslev, E., Marques-Bonet, T., Ryder, O.A., McCue, M., Rieder, S. Leeb, T., Slatkin, M., and Orlando, L. . E olutio a Ge o i s a d Co se atio of the E da ge ed P ze alski s Ho se. Current Biology 25:2577-2583. Dietz, U.L. . Ho se a k idi g: a s a ess to speed?, in M. Levine (ed.) Prehistoric steppe adaptation and the horse. Cambridge: McDonald Institute for Archaeological Research. . Ci e ia B idles: P og ess in Cavalry Technology? In Horses and Humans: The Evolution of Human-Equine Relationships, S. Olsen, S. Grant, A. Choyke, and L. Bartosiewicz (eds.), pp. 157–163. Achaeopress, Oxford. Draper, J., Sly, D. and Muir, S. 2014. Complete Book of Horses and Riding. Hermes House/Anness Publishing, London. Dyson Hudson R. and Dyson Hudson, N. 1980. Nomadic Pastoralism. Annual Review of Anthropology 9:15-61 Enkhtuvshin, Batbold and Tumurjav, Myatav. 2011. Mongolian Nomads and Nomadic Animal Husbandry. International Institute for the Study of Nomadic Civilizations, Ulaanbaatar, Mongolia. 119 Enloe, J.G., and Turner, E. 2002. Methodological problems and Biases in Age determinations: a view from the Magdalenian. In: Muscillo, D. (Ed.), Recent Advances in Ageing and Sexing Animal Bones. Oxbow Books, Oxford. Eregzen, Gelegdorj. 2016. Anceient Funeral Monuments of Mongolia. Archaeological Relics of Mongolia, Vol. 3. Mongolian National Academy of Sciences, Institute of History and Archaeology, Ulaanbaatar. Evans, P.A., Jack, N., and Jones, S. 2006. Aging horses by their teeth. Fayetteville: University of Arkansas Cooperative Extension Service (Division of Agriculture). Drews, Robert. 2004. Early Riders: the Beginnings of Mounted Warfare in Asia and Europe. Routledge, New York. Erdene-Ochir, N., and Khodyakov, Y.S. 2016. Mongoliin Ertnii Nudelchdiin Zer Zevseg. New Research on Mongolian Archaeology Series, Vol. 2. Institute of History and Archaeology, Mongolian Academy of Sciences, Ulaanbaatar. Fedorova, N. 2003a. Treasures of the Ob. Western Siberia and medieval trade routes. Salekhard & St Petersburg: Yamal-Nenets Shemanovsky Museum and State Hermitage Museum. 2003b. Ust-Polui: 1st Century B.C. Salekhard & St Petersburg: Yamal-Nenets Shemanovsky Museum and State Hermitage Museum. Fitzhugh, W. 2004. The Hovsgol Deer Stone Project: 2003 Field Report. Smithsonian Institution, Washington, D.C. 2005. The Deer Stone Project: Anthropological Studies in Mongolia 2002-2004. Smithsonian Institution, Washington, D.C. 2006. American-Mongolian Deer Stone Project: Field Report 2006. Smithsonian Institution, Washington, D.C. 2009a. The Mongolian Deer Stone-Khirigsuur Complex: dating and organization of a late Bronze Age menagerie, in J. Bemmann, H. Parzinger, E. Pohl & D. Tseveendorj (ed.) Current archaeological research in Mongolia: papers from the 1st International Conference on Archaeological Research in Mongolia held in Ulaanbaatar, August 19th–23rd, 2007 (Bonn Contributions to Asian Archaeology, volume 4). Bonn: Bonn University Press. 120 2009b. Pre-Scythian ceremonialism, Deer Stone Art, and cultural intensification in northern Mongolia, in B. Hanks & K. Linduff (ed.) Social complexity in prehistoric Eurasia: monuments, metals, and mobility. Cambridge: Cambridge University Press. http://dx.doi.org/10.1017/CBO9780511605376 2009c. American-Mongolian Deer Stone project: field report 2009. Washington (DC): Smithsonian Institution. Fitzhugh, W. and Bayarsaikhan, J. 2008. American-Mongolian Deer Stone Project; Field Report 2008. Smithsonian Institution, Washington, D.C. 2009. American-Mongolian Deer Stone Project: Field Report 2009. Smithsonian Institution, Washington, D.C. 2011. Mapping ritual landscapes in Mongolia and beyond: unravelling the Deer Stone-Khirigsuur enigma, in P. Sabloff (ed.) Mapping Mongolia: situating Mongolia in the world from geologic time to present. Philadelphia: University of Pennsylvania Museum of Archaeology and Anthropology. http://doi:10.1017/CBO9780511605376 Fitzhugh, W. and Kortum, R. 2012. Rock Art and Archaeology: Investigating Ritual Landscape in the Mongolian Altai Field Report 2011. Smithsonian, Washington, D.C. Fitzhugh, W., Kortum, R., and Bayarsaikhan, J. 2013. Rock Art and Archaeology: Investigating Ritual Landscape in the Mongolian Altai Field Report 2012. Smithsonian, Washington, D.C. Frachetti, Michael. 2008. Pastoralist Landscapes and Social Interaction in Bronze Age Eurasia. University of California Press, Berkeley. Frohlich, B., Amgalantugs, T., Littleton, J., Hunt, D., Hinton, J., and Goler, K. 2009. Bronze Age Burial Mounds in the Khovsgol Aimag, Mongolia. In Current Archaeological Research in Mongolia: Papers from the 1st International Conference on Archaeological Research in Mongolia held in Ulaanbaatar, August 19th-23rd, 2007, J. Bemmann, H. Parzinger, E. Pohl, and D. Tseveendorj (eds.), pp. 99–115. University of Bonn, Bonn. Fukumoto, Yu, Kashima, Kaoru, Orkhonselenge, A., and Ganzorig, U. 2012. Holocene environmental changes in northern Mongolia inferred from diatom and pollen records of peat sediment. Quaternary International 254, 83-91. 121 Gantulga, J, Yeruul-Erdene, Ch., and Magail, J. 2016. Khoid Tamirin Bugan Khushuu. New Research on Mongolian Archaeology Series, Vol. 3. Institute of History and Archaeology, Mongolian Academy of Sciences, Ulaanbaatar Goulden, Clyde, Nandintsetseg, B., and Ariuntsetseg, L. 2011. The geology, climate, and ecology of Mongolia. In: Sabloff, Paula (Ed.), Mapping Mongolia: Situating Mongolia in the World from Geologic Time to Present. University of Pennsylvania Museum of Archaeology and Anthropology, Philadelphia. Hanks, Bryan. 2010. Archaeology of the Eurasian steppes and Mongolia. Annual Review of Anthropology 39: 469–86. http://dx.doi.org/10.1146/annurev.anthro.012809.105110 2012. Mounted warfare and its sociopolitical implications. In S. Stark and K. Rubinson (eds.), Nomads and Networks: the ancient art and culture of Kazakhstan. Institute for the Study of the Ancient World, New York. Harris, S., Iossa, G., and Soulsbury, C. 2006. A review of the welfare of animals in circuses. London: Royal Society for the Prevention of Cruelty to Animals. Honeychurch, W. 2015. Inner Asia and the spatial politics of empire: archaeology, mobility, and culture contact. New York: Springer. http://dx.doi.org/10.1007/978-1-4939-1815-7 Honeychurch, W., Wright, J., and Amartuvshin, C. 2009. Re-writing monumental landscapes as Inner Asian political process, in B. Hanks & K. Linduff (ed.) Social complexity in prehistoric Eurasia: monuments, metals, and mobility. Cambridge: Cambridge University Press. Honeychurch, W., Fitzhugh, W., and Amartuvshin, C. 2013. Precursor to empire, in W. Fitzhugh, M. Rossabi & W. Honeychurch (ed.) Genghis Khan and the Mongol empire. Washington (DC): W.W. Norton & Company. http://dx.doi.org/10.1017/CBO9780511605376.019 Hosey, G., Melfi, V., and Pankhurst, S. 2013. Zoo animals: behaviour, management, and welfare. Oxford: Oxford University Press. Houle, J-L. 122 2009. Socially integrative facilities and the emergence of societal complexity on the Mongolian Steppe, in B. Hanks & K. Linduff (ed.) Social complexity in prehistoric Eurasia: monuments, metals, and mobility. Cambridge: Cambridge University Press. http://dx.doi.org/10.1017/CBO9780511605376.020 2010. Emergent complexity on the Mongolian Steppe: mobility, territoriality, and the development of early nomadic polities. PhD dissertation, University of Pittsburgh. 2016. Western Mongolia Archaeology Report 2015, University of Western Kentucky, Bowling Green, KY. Jacobson-Tepfer, Ester. 1993. The Deer Goddes of Ancient Siberia: a Study in the Ecology of Belief. Brill, Leiden. 2012. The Image of the Wheeled Vehicle in the Mongolian Altai: Instability and Ambiguity. The Silk Road 10:1-28. 2015. The Hunter, the Stag, and the Mother of Animals. Oxford University Press, London. Janz, Lisa. 2012. The Chronology of Post-Glacial Settlement in the Gobi Desert and the Neolithization of Arid Mongolia and China. Ph.D. Dissertation, University of Arizona. Jenkins, M. 2014. A Wi te s Tale: Chi a s Alta Mountains. National Geographic Traveler, December 2014/January 2015. Available at http://intelligenttravel.nationalgeographic.com/2015/01/13/awinters-tale-chinasaltay-mountains/. Johnson, T., and Porter, C. 2006. Dental Conditions Affecting the Mature Performance Horse (5-15 years). American Association of Equine Practitioners Focus Meeting 2006, Indianapolis, IN, USA. Available online at htttp://www.ivis.org/readings/eqdent/toc.asp Kelekna, Pita. 2009a. The politico-economic impact of the horse on Old World cultures: an overview. SinoPlatonic Papers 190: 1–31. 2009b. The Horse in Human History. Cambridge University Press, London. Khazanov, A.M. 1984. Nomads and the Outside World. Cambridge University Press, London. Kradin, Nikolai. 123 2003. Nomadic Empires: Origins, Rise, Decline. In Nomadic Pathways in Social Evolution, N. Kradin and T. Barfield (eds.). Meabooks, Quebec 2015. The Ecology of Inner Asian Pastoral Nomadism. In The Ecology of Pastoralism, P. Nick Kardulias(ed.), pp. 41-70. University Press of Colorado, Boulder. Kovalev, A, and Erdenebaatar, D. 2010. Discovery of new cultures of the Bronze Age in Mongolia according to the data obtained by the International Central Asian Archaeological Expedition. In Current Archaeological Research in Mongolia: Papers from the 1st International Conference on Archaeological Research in Mongolia held in Ulaanbaatar, August 19th-23rd, 2007, J. Bemmann, H. Parzinger, E. Pohl, and D. Tseveendorj (eds.), pp. 149-170. University of Bonn, Bonn. Kuzmina, E. 2000. The Eurasian steppes: the transition from early urbanism to nomadism, in J. Davis-Kimball, E. Murphy, L. Koryakova & L. Yablonsky (ed.) Kurgans, ritual sites, and settlements: Eurasian Bronze and Iron Age (British Archaeological Reports International series 890). Oxford: Archaeopress. 2003. Origins of Pastoralism in the Eurasian Steppes. In Prehistoric Steppe Adaptation and the Horse, M. Levine, C. Renfrew, and K. Boyle (eds.). McDonald Institute for Archaeological Research, Oxford. 2006. Mythological treatment of the horse in Indo-European culture, in S. Olsen (ed.) Horses and humans: the evolution of human-equine relationships. Oxford: Basingstoke. Lattimore, Owen. 1940. Inner Asian Frontiers of China. Beacon Press, Boston. Lee, S., and Bronk Ramsey, C. 2012. Development and application of the trapezoidal model for archaeological chronologies. Radiocarbon 54:107–22. Lees, Susan, and Bates, Daniel. 1974. The origins of specialized nomadic pastoralism: a systemic model. American Antiquity 39 (2), 187-193. Legrand, S. 2006. The emergence of the Scythians: Bronze Age to Iron Age in south Siberia: the emergence of the Karasuk culture. Antiquity 80: 843–60. http://dx.doi.org/10.1017/S0003598×00094461 Lepetz, Sebastien. 124 2013. Horse sacrifice in a Pazyryk Culture Kurgan: the Princely Tomb of Berel'(Kazakhstan): selection criteria and slaughter procedures. Anthropozoologica 48 (2), 309-321. Levine, Marsha. 1982. The use of crown height measurements and eruption-wear sequences to age horse teeth, in B. Wilson, C. Grigson & S. Payne (ed.) Aging and sexing animal bones from archaeological sites (British Archaeological Reports British series 109). Oxford: Archaeopress. 1983. Mortality models and the interpretation of horse population structure. In: Bailey, G. (Ed.), Hunter Gatherer Economy in Prehistory: a European Perspective. Cambridge University Press, New York. 1999. The origins of horse husbandry on the Eurasian steppe, in M. Levine (ed.) Late prehistoric exploitation of the Eurasian steppe. Cambridge: McDonald Institute for Archaeological Research. Linduff, Katheryn. 2003. A walk on the wild side: late Shang appropriation of horses in China. In: Levine, M., Renfrew, C., Boyle, K. (Eds.), Prehistoric Steppe Adaptation and the Horse. McDonald Institute for Archaeological Research, Oxford. Littauer, Mary. n.d. Evolution of the Bit. Unpublished transcript of lecture given at Sweet Briar College, accessed via Littauer Archives, International Museum of the Horse, Lexington, KY. 1969. Bits and pieces. Antiquity 43:289–300. http://dx.doi.org/10.1017/S0003598×00040734 1971. V.O. Vitt and the horses of Pazyryk. Antiquity 45:293-294. Littleton, J., Floyd, B. Frohlich, B., Dickson, M., Amgalantugs, T., Karstens, S., and Pearlstein, K. 2012. Taphonomic analysis of Bronze Age burials in Mongolian khirigsuurs. Journal of Archaeological Science 39: 3361–70. http://dx.doi.org/10.1016/j.jas.2012.06.004 MacFarland, C. 2013. The Ho se a s Guide to Ta k a d E uip e t: Fo Horseman. Globe Pequot Press, Fort Worth, TX. , Fit, a d Fu tio . Western Manfredi, J., Rosenstein, D., Lanovaz, J, Nawelaerts, S. and Clayton, H. 2010. Flouroscopic study of oral behaviours in response to the presence of a bit and the effects of rein tension. Comparative Exercise Physiology 6(4), 143–14 Mair, Victor. 125 2003. The horse in late prehistoric China: wrestling cultu e a d o t ol f o the a a ia s , in M. Levine (ed.) Prehistoric steppe adaptation and the horse. Cambridge: McDonald Institute for Archaeological Research. Mallory, J.P, and Adams, Douglas. 1997. The Encyclopedia of Indo-European Culture. Fitzroy Dearborn, London. McDonnell, S. 1988. Normal and abnormal behavior of stabled horses. Proceedings of the 1998 Alberta Horse Breeders and Owners Conference. Edmonton: Alberta Agriculture, Food, and Rural Development. Mitchell, P. 2015. Horse Nations: The Worldwide Impact of the Horse on Indigenous Societies Post-1492. Oxford University Press, London. National Museum of Mongolia. 2013. Altan Dornod Mongol Kompaniin 001134A Dugaartai Khashaat Nertei 125, 26-ga Ashiglaltiin Talbaid xiisen Arkheologiin Maltlaga Sudulagaa. National Museum of Mongolia, Ulaanbaatar. Nyambat, M. and Odbaatar, Ts. 2010. Zuunkhangai Sumiin Nutgaas Shineer Oldson Bugan Khushuu. Nomadic Heritage Studies 10(4), 57-66 Olsen, Sandra. 2003. The exploitation of horses at Botai, Kazakhstan, in M. Levine (ed.) Prehistoric steppe adaptation and the horse. Cambridge: McDonald Institute for Archaeological Research. 2006a. Early horse domestication on the Eurasian steppe, in M. Zeder, D. Bradley, E. Emshwiller & B. Smith (ed.) Documenting domestication: new genetic and archaeological paradigms. Berkeley, University of California Press. 2006b. Early horse domestication: weighing the evidence. In: Olsen, S., Grant, S., Choyke, A., Bartosiewicz, L. (Eds.), Horses and Humans: the Evolution of Human-Equine Relationships. Archaeopress, Oxford. Outram, A.K., Stear, N., Bendrey, R., Olsen, S., Kasparov, A., Zaibert, B., Thorpe, N., and Evershed, R. 2009. The Earliest horse harnessing and milking. Science 323, 1332-1335. http://doi:10.1126/science.1168594 Outram, A.K., Stear. N, Kasparov, A, Usmanova, E., Varfolomeev, V., and Evershed, R. 126 2011. Horses for the dead: funerary foodways in Bronze Age Kazakhstan. Antiquity 85(116-128). Outram, A.K., Kasparov, Alexei, Stear, Natalie A., Varfolomeev, Victor, Usmanova, Emma, and Evershed, Richard P. 2012. Patterns of pastoralism in later Bronze Age Kazakhstan: new evidence from faunal and lipid residue analyses. Journal of Archaeological Science 39:2424-2435. Pasquini, Chris, Spurgeon, Tom, and Pasquini, Susan. 2003. Anatomy of Domestic Animals. Sudz Publishing, Pilot Point, TX. Pederson, Neil, Hessl, Amy, Baatarbileg, Nachin, Achukaitis, Kevin, and di Cosmo, Nicola. 2014. Pluvials, droughts, the Mongol Empire, and Modern Mongolia. PNAS 111(12):4375-4379 Perez, W. and Martin, E. 2001. An explanation of a groove found on the nasal process of the equine incisive bone. http://dx.doi.org/10.1046/j.1439Anatomia, Histologia Embryologia 30:357–58. 0264.2001.00348.x Penissi, Elizabeth. 2016. Frozen Scythian stallions unravel mysteries of horse domestication. Science Magazine Online, accessed 11 October 2016. DOI: 10.1126/science.aag0003 Phillipsen, Bente. 2013. The freshwater reservoir effect in radiocarbon dating. Heritage Science 2013 1(24):1-19. Propokenko, Alexander, and Khursevich, Galina K., Bezrukova, Elena V., Kuzmin, Mikhail I., Boes, Xavier, Williams, Douglas F., Fedenya, Svetlana A., Kulagina, Nataliya V., Letunova, Polina P., and Abzaeva, Anna A. 2007. Paleoenvironmental proxy records from Lake Hovsgol, Mongolia, and a synthesis of Holocene climate change in the Lake Baikal watershed. Quaternary Research 68:2-17. Putnam, A et al. 2016. Little Ice Age wetting of interior Asian deserts and the rise of the Mongol Empire. Quaternary Science Reviews 131: 33-50. Renfrew, C. . All the Ki g s ho ses: assessi g cognitive maps in later prehistoric Europe, in S. Mithen (ed.) Creativity in human evolution and prehistory. London: Routledge. Rogers, J. Daniel. 127 2012. Inner Asian States and empires: theories and synthesis. Journal of Archaeological Research 20(3):205-256. Rolle, R. 1980. The World of the Scythians. University of California Press, Berkeley. Rudenko, S.I. 1970. The Frozen Tombs of Siberia. Dent, London. Salzman, Philip. 2004. Pastoralists: Equality, Hierarchy, and the State. Westview Press, Boulder. Sanjmyatav, T. 1993. Arkhangai Aimag Nutag Dakhi Ertnii Tuukh Soyoliin Dursgal. Arkhangai Aimagiin Zasag Dargiin Tamgiin Gazar, Ulaanbaatar, Mongolia. Sasada, Yukiko. 2013. A Alte ati e Theo fo the Bit Wea Fou d o the Lo e “e o d P e ola of the Buhen Horse. In Chasing Chariots: Proceedings of the First International Chariot Conference, Cairo 2012. A. Veldmeijer and S. Ikram (eds.), Leiden, Sidestone Press. Shelach, G. 2009. Prehistoric societies on the northern frontiers of China: archaeological perspectives on identity formation and economic change during the first millennium BCE. London: Routledge. Sheratt, Andrew 2003. The Horse and the Wheel: the Dialectics of Change in the Circum-Pontic Region and Adjacent Areas, 4500-1500 BC. In Prehistoric steppe adaptation and the horse, M. Levine (ed.). Cambridge: McDonald Institute for Archaeological Research. Shultz, D. and Costopolous, A. N.d. Wealth inequality, environmental risk, and growth rates in pastoral nomadic society. Manuscript in preparation. Smith, Bruce. 2012. A Cultural Niche Construction Theory of Initial Domestication. Biological Theory 6(3):1-12 Takacs, I. 128 1985. Deformations on the Skeletons from the Migration Period and the Period of Hungarian Conquest Caused by Horse-Furniture. A Mora Ferenc Muzeum Evkvonyve 1984/85-2: 311–321. Szeged, Hungary. Takahama, Shu, Hayashi, Toshio, Kawamata, Masanori, Matsubara, Ryuji, and Erdenebaatar, D. 2006. Preliminary Report of the Archaeological Investigations in Ulaan Uushig I (Uushigiin Ovor) in Mongolia. University of Kanazawa (Japan) Bulletin of Archaeology 28, 61-102. Taylor, William. In press. Horse demography and use in Bronze Age Mongolia, Quaternary International (2016), http://dx.doi.org/10.1016/j.quaint.2015.09.085 Taylor, William, Tuvshinjargal, Tumurbaatar, Bayarsaikhan, Jamsranjav. 2016. Reconstructing Equine Bridles in the Mongolian Bronze Age. Journal of Ethnobiology 36(3): 553–568 Taylor, William, Bayarsaikhan, Jamsranjav, and Tuvshinjargal, Tumurbaatar. 2015. Equine cranial morphology and the archaeological identification of riding and chariotry: applications to Mongolia's late Bronze Age. Antiquity 89:854-871. http://dx.doi.org/10.15184/aqy.2015.76 Taylor, W., Lowry, B., Jargalan, B., Clark, J., Tuvshinjargal, T., and Bayarsaikhan, J. In review. A Bayesian Chronology for Early Domestic Horse Use in the Eastern Steppe. Manuscript submitted to Journal of Archaeological Science, October 2016. Toshio, H. 2013. The Beginning and the Maturity of Nomadic Powers in the Eurasian Steppes: Growing and Downsizing of Elite Tumuli. Ancient Civilizations from Scythia to Siberia 19:105-141 Turbat, T. 2011. Sacrificial Structures of the Later Period. In Deer Stones of the Jargalantyn Am, p. 124–129. Mongolian Tangible Heritage Association, Ulaanbaatar. Tumen. D., Erdene, M., Khatanbaatar, D., Ankhsanaa, G., and Tsidenova, HB. 2012. Do od Mo golA kheologii Khee ii “udalgaa. Mongolian Journal of Anthropology, Archaeology, and Ethnology 7(1):1-38. Vanderwegen, M. and Simoens, P. 2002. Anatomical study of the notches in the nasal process of the equine incisive bone. Equine Veterinary Journal 32:199–202. http://dx.doi.org/10.2746/042516402776767222 129 Volkov, V.V. 1990. Survey of Bronze and Early Iron Age Monuments of the NITSW4, 1989 Field Report. Mongolia Academy of Sciences Insitute of Archaeology, Ulaanbaatar. 2002 [1981]. Olennye kamni Mongolii (2nd edition). Ulaanbaatar: Academy of Sciences, Moscow; Nauka. Wang, Wei, Ma, Yuzheng, Feng, Zhaodong, Narantsetseg, Ts, Liu, Kam-Biu, and Zhai, Xinwei. 2011. A prolonged dry Mid-Holocene climate revealed by pollen and diatom records from lake Ugii Nuur, Central Mongolia. Quaternary International 229:74-83. Webb, David, and Hemmings, Andrew 2006. Last horses and first humans in North America. In: Olsen, S., Grant, S., Choyke, A., Bartosiewicz, L. (Eds.), Horses and Humans: the Evolution of Human-equine Relationships. Archaeopress, Oxford. Wright, Joshua. 2006. The Adoption of Pastoralism in Northeast Asia: Monumental Transformation in the Egiin Gol Valley, Mongolia. Ph.D. Dissertation, Harvard University, Cambridge. 2012. Inequality on the surface: horses, power, and community in the Mongolian Bronze Age. In: Arbuckle, B., McCarty, S. (Eds.), Animals and Inequality in the Ancient World. University of Colorado Press, Boulder. 2014. Landscapes of inequality? A critique of monumental hierarchy in the Mongolian Bronze Age. Asian Perspectives 51: 139–63. http://dx.doi.org/10.1353/asi.2014.0008 Wu, H-Y. 2013. Chariots in early China: origins, cultural interaction, and identity (British Archaeological Reports international series 2457). Oxford: Archaeopress. Yatsenko, S. 2015. Images of the Early Turks in Chinese Murals and Figurines from the Recently Discovered Tomb in Mongolia. The Silk Road 12:13-24 Yeruul-Erdene, C. et al. 2015. Orkhonii Khundii Dekh Mongol-Ge a ii Kha ta sa Uridchilsan Ur Dun. Studia Archaeologica XXXV(14):198-216 Ba ko Tuslii “uldalgaa i Yuan, J., and Flad, R. 2006. Research on early horse domestication in China, in M. Mashkour (ed.) Equids in time and space: papers in honor of Vera Eisenmann. Oxford: Oxbow Books. 130 Zeder, Melinda. 2016. Domestication as a model system for niche construction theory. Evol Ecol 30:325-348 131

RELATED PAPERS

RELATED TOPICS

Log In

Origins of Horse Herding and Transport in the Eastern Steppe

Origins of Horse Herding and Transport in the Eastern Steppe

Related Papers

RELATED PAPERS

RELATED TOPICS