• 1 Cl. Mo. qosiγun-u nutuγ-un ǰiruγ (map of the territory of a banner); Ch. youmutu 游牧圖 (map/image of (...)
• 2 The Mongol banners (qosiγu) were geographical divisions organized by the Qing, ruled by a hereditar (...)
• 3 Mongolian words are transcribed from Classical Uyghur Mongolian, except for place names of the Repu (...)
• 4 About 22% of the toponyms M. Haltod listed from the 182 German maps of the Berlin collection are na (...)

1The maps of the banners1 of Inner and Outer Mongolia2 drawn to be sent to the central Qing (1644-1911) administration – more precisely, the Lifanyuan 理藩院 (Board of Government of the Outer Regions) – in the late Qing period contain a great number of names and drawings of ovoos (Cl. Mo. oboγa3), and certain mountains are also called “ovoo”: more than a fifth4 of the toponyms have “ovoo” (written oboγa or more phonetically oboo or obo) in their name. What do these banner maps tell us about ovoos?

• 5 They were collected by Walther Heissig and Hermann Consten and were previously preserved in the Wes (...)
• 6 About a thousand banner maps are preserved in Japan, Mongolia, China and Europe. The dates and styl (...)

2The maps under discussion are the 182 maps in the collection of the Berlin State Library (Preussischer Kulturbesitz, Orientabteilung)5. Thanks to their online availability in a high definition, it is possible to study them and read most of their texts (in Mongolian and/or Chinese, and Manchu for some of them)6.

• 7 There were forty-nine Mongol autonomous banners (qosiγu) in Inner Mongolia grouped into six “league (...)

3This paper examines the naming of ovoos, their representation by various symbols, and their location on late 19^th and early 20^th century maps of the banners of Inner and Outer Mongolia7. It first introduces the corpus of banner maps, their conventions and texts, and the regional groups that can be identified in the Berlin collection. In the late Qing period, two main types of ovoos structures were usually distinguished: “boundary-marking ovoos” that punctuated the boundaries with the neighboring banners and the border of the northern Qing frontier, and ovoos on mountain tops that I choose to call “cult ovoos”; in addition, many mountains are called “ovoo”. Do all banner maps make a clear distinction between these two types of ovoos, and which kinds of symbols are used to differentiate them? The following section focuses on the boundary-marking ovoos. In the late 19^th century, the Qing government was primarily preoccupied by the precise localization of the banners’ boundaries. But were boundary-marking ovoos systematically named and located on banner maps, and were their location on maps precise enough to fix the boundaries between banners? Besides, boundary-marking ovoos were generally viewed as simple “border devices” used by the Qing government to delimit the territories of Inner (southern) and Outer Mongolia (Kollmar-Paulenz 2006, p. 368). I will ask whether they may also have had a religious function, and whether cult ovoos located on a banner’s boundary may also have served as boundary-marking ovoos. This would confirm Caroline Humphrey’s concept on “edge-based centricity” – a sacred mountain marking the border and at the same time linking different communities – in Buryatia (2016).

• 8 These written reports are called in Mongolian “nutuγa čese” (report on a territory), “ögöled čese” (...)
• 9 Before the Mongolian People’s Republic, there was no complete map of North (Outer) Mongolia.

6In the Qing period, by imperial decree, every Mongol banner prince (ǰasaγ) was required to submit a map of the banner he ruled, accompanied by a written report8 that precisely indicated its boundaries. The great majority of them represent a single banner, but a few maps portray a group of banners, an ayimaγ (the largest Qalqa administrative division), a series of relay-stations (örtege, Ch. tai 台) connected by a road, or border-posts (qaraγul, ger qaraγul, Ch. kalun 卡倫, < Ma. karun) and ovoos along a buffer zone of the Qing-Russian border9. These maps are therefore political documents produced by the province for the centralized government, and thus allowed symbolical possession and military control of a territory: as expressed by David Harvey, “command over space is a fundamental and all-pervasive source of social power” (1989, p. 226). They were not public and have not been printed.

• 10 See Heissig (1944, pp. 125-130) and Pratte (2021) for early Qing maps and orders to draw maps.
• 11 Dated maps of the Berlin corpus were drawn between 1889 and 1920; among them, seventy-five are date (...)
• 12 Wang (2021) studies the nationalization of pasturelands, the creation of new administrative distric (...)
• 13 The increasing pressure of Chinese immigration was the main cause of the numerous Mongol uprisings (...)

7Except for a few maps produced before the 19^th century10, the first banner maps were made between 1803 and 1805. The great majority of them date from the late 19^th and early 20^th centuries11, a period of great transformations and economic crisis, when Mongol lands were no longer considered as grazing pastures and a buffer region between the Qing and Russian empires, but as an area to develop and exploit. Industrial mining was legalized, ruling princes sold or mortgaged land to Chinese agriculturalists for profit, “and the banner system was failing to address the economic and political challenges of the 19^th century” (High & Schlesinger 2010, pp. 293-294). Following the Boxer Uprising, the Qing undertook a series of reforms known as the “New Policy” (Xinzheng 新政, 1901-1911) that aimed at transforming the Qing empire into a modern nation-state. The New Policy fundamentally changed the status of Mongol lands since it officially opened Inner Mongolia to Chinese colonization in order to finance its modernizing programs as well as to stabilize the frontier (yimin shibian 移民實邊) and consolidate Qing control12. Inner and Outer Mongolia’s place in the empire had changed: they were no longer protected territories untainted by foreign influences. The Inner Mongol territory was transformed by Chinese immigration, agriculture, and the formation of villages, increased exploitation of natural resources, and ethnic tensions between Mongols and Chinese13. The Qing authorities requested information on the following topics: agriculture, forestry, grazing, wildlife, leather, wool, railways, minerals, fishing, salt, army, schools, border-posts, relay-stations and trade (Kamimura 2005, pp. 14-15).

• 14 Toponyms were usually transcribed into Chinese while the other texts were translated. Some maps wer (...)

10The banner maps were drawn in the banner’s administrative center (Ch. yamen 衙門, Cl. Mo. yamun) and painted with pigments. Interestingly, the banner administrations apparently did not receive any training in mapmaking, triangulation or geometry, or help of skilled mapmakers: the mapmakers were inexperienced local Mongols, which guaranteed local perspective (Pratte 2021, chapter 4). When finished, the banner prince(s) (ǰasaγ[s]) added his (their) seals. The maps were then copied so a double was kept in the yamen (and perhaps a second copy in the seat of the league), serving as a model for later maps. Then, the maps along with the reports (čese) were sent to the Lifanyuan in Beijing, where all the inscriptions were translated or transcribed into Chinese (written directly on the map or on small papers pasted near or upon the Mongolian inscriptions, preventing the reader from reading the original content). Some maps were entirely copied and translated into Chinese (Heissig & Sagaster 1961, pp. 337-342)14. The preserved maps show the different steps in their production, from the initial unpainted drawing to the Chinese copies. Most of them have been preserved without their accompanying report.

• 15 Regulations for drawing banner maps were promulgated in 1805, 1864, and 1890 (Futaki 2005, p. 28; K (...)
• 16 1 γaǰar=360 alda, 576 m, 1 qubi=0,1 γaǰar, 1 alda=160 cm.
• 17 The grid made it easier to copy the map.

11The late 19^th and early 20^th century banner maps had to follow specific rules15. They were based on mathematical measurements and topographic surveys, with a fixed scale, and were supposed to 1) indicate the boundaries by ovoos; 2) indicate the four cardinal points and be north-oriented; 3) use the following measurement units: γaǰar (= Ch. li 里), qubi and alda16; 4) use a grid of vertical and horizontal lines facilitating the measure of distances and surfaces17; 5) and (to some extent) use standardized conventions (Kamimura 2005, p. 16; Futaki 2005, p. 29; Chagdarsurung 1975). But these rules were not always respected, and many of the maps, as we will see, neither even use a grid nor depict boundary-marking ovoos.

• 18 Chinese maps were influenced by the European cartography introduced by the Jesuits in China in the (...)
• 19 Pratte (2021) speaks of “shifting worm’s eye view”: a view from the ground on the center of the map (...)
• 20 A few of them are very abstract, devoid of drawings, and accompanied by captions (see cat. 780, dat (...)
• 21 A three-dimensional object is represented by a drawing having all axes drawn to exact scale: all li (...)
• 22 For similar remarks about Chinese maps of the end of the empire: Smith 1998, p. 60; Hearn 2011, p.  (...)
• 23 On maps as “pictures”, see and Kollmar-Paulenz 2006, pp. 357-358.

12The banner maps adopt conventions similar to that of Chinese maps of the same period18, using varying scales, but also have many distinctive features such as multiple viewpoints19. The maps usually combine pictorial elements and abstract symbols20. Buildings and mountains are seen from the front or in axonometric perspective21 while lakes and streams are seen from above (i.e. projected orthogonally). The size of the political, administrative and religious centers of some ovoos or of a mountain or spring as shown on a map may be proportional to its importance in the life of the inhabitants and not to its real size, enlargements also making it possible to show a greater level of detail. Banner maps allow us to locate a great number of monasteries, but some important objects have not been represented while others may seem negligible according to our point of view22. Because of the conveyed emic conceptions of space, the pictorial elements, the practice of variable scale, variable perspective, and use of multiple viewpoints, some authors, following Harvey (1980), prefer using the term “picture-maps”23. However, Mongol banner maps can be contrasted with much more pictorial maps, such as paintings of the pilgrimage destinations of Tibet and China, that use representations of living beings (men, camel caravans, wild animals), and apparitions of buddhas in clouds (Charleux 2015, pp. 170-177).

• 24 Pratte (2021) highlights the profound transformation of maps of Sečen qan ayimaγ after the promulga (...)

13Pratte (2021) argues that pictographic maps focusing on the landscape without showing clear boundaries represent an emic viewpoint and an earlier stage of map making. In the Berlin collection of maps, regional differences are many, and show that the majority of banners obviously resisted state attempts of standardization and ignored the new instructions of 1864 and 189024. We can identify about eleven more or less homogeneous regional groups:

• 26 The Berlin catalogue numbers the maps of the northern Mongol lands from west (Qobdo Uriyangqai) to (...)
• 27 Cat. 807, 809, 810, 812, 813, 814, 820, 822.
• 28 Pratte (2021, p. 334) highlights for Sečen qan ayimaγ the work of the league heads as intermediate (...)

14From west to east26 in Outer Mongolia we notice a trend from the non or partial respect of the above-mentioned rules (no grid [cat. 672-685] and very few or no boundary-marking ovoos depicted in Qobdo and Uriyangqai) to a stricter respect, with the use of a grid and numbered ovoos that are equally distributed all along the boundaries (on maps of the Sečen qan ayimaγ) (Appendix). The maps of the banners of Qobdo Uriyangqai, J̌asaγtu qan, Sayin noyan qan and Tüsiyetü qan pay particular attention to their boundaries; often, the interior of the map only indicates a few landscape features and the location of the banner prince’s or reincarnated lama’s encampment (fig. 1a, b). The maps of the banners of Sečen qan ayimaγ include more drawings of natural landscapes and buildings, and numbered boundary-marking ovoos. The maps of Inner Mongol banners, although they were supposed to follow the same regulations as the maps of Outer Mongolia, rarely use the grid (except for eight maps of banners of Sili-yin γool27), pay much less importance to the delimitation of their boundaries (many have few or no boundary-marking ovoos), and are more pictorial. They show an overabundance of details, human presence being marked by the depiction of walled cities, villages, settlements, roads, monasteries, Chinese temples and churches, shrines to Chinggis Khan, stupas (Skt. stūpas), bridges, tombs, wells, ovoos and so on, especially in the banners of J̌irim, J̌osoto and J̌oo-uda which were more densely populated and relatively more urbanized than the western leagues. Maps of the banners of J̌osoto and J̌oo-uda are painted in Chinese landscape style (fig. 1c). These eleven regional “families” obviously result of a great work of coordination between the different banner princes and the chief of the ayimaγ or league (čiγulγan): the latter played an important role in standardizing the maps of the different banners of their ayimaγ or čiγulγan28.

• 29 Haltod (1966) listed 13 644 Mongol place names in the first volume of Mongolische Ortsnamen (there (...)

15The texts of the banner maps usually give the title of the map, the date, title and rank of the banner prince or reincarnated lama, the cardinal points, a description of the banner, as well as descriptions of boundaries with the adjacent banners, and a number of toponyms. Texts are written vertically, or radiating from the center outside of the boundary line, or following different directions, necessitating that a viewer rotate the map to be able to see and read, or circle around it. Most of the scholars who worked on these maps were interested in toponyms: 13 785 Mongolian place names on the “German” maps were listed29. In average, there are seventy-five toponyms (not counting the additional texts) per map, and maps with one hundred to two hundred place names are not uncommon.

• 30 In fumigation texts, the Mongol landscape is equated with mountains of Tibet and ancient India. See (...)

16The way in which banner maps (and reports) list names of mountains, rivers, springs and ovoos, along with precise descriptions of a banner’s boundaries, could be compared to Mongol literary genres that glorify the homeland (nutuγ), such as songs, oral prayers, and fumigation ritual (sang) texts dedicated to spirit-masters of the land. Some fumigation texts (in Tibetan language, less frequently in Mongolian) mention hundreds of names of sacred places30.

• 31 Many phonetic spellings and “misspellings” of toponyms appear in banner maps, such as baying for ba (...)
• 32 Dui means “pile, heap”. Fengdui is the official term according to imperial edicts of the Daoguang e (...)

17On banner maps, ovoos are named only (written in Cl. Mo. oboγa, oboo, or obo31), named and represented, or represented but not named. In Chinese transcription, oboγa is written ebo 鄂博, more rarely ebo 鄂伯 (cat. 673) – today transcribed aobao 敖包 –, and in maps of Ordos, it is sometimes translated into Chinese fengdui 封堆32 or dui 堆.

• 33 Haltod (1966, p. 110) reads on the Berlin maps many occurrences of ovoos called Modon oboγa/obo, Mo (...)

20Some ovoos were made of wood; on banner maps in most cases wooden ovoos are distinguished by their name only (Modon oboγa)33 but depicted not differently from stone cairns (fig. 4a). The vertical tufts on mountain tops of cat. 818 (fig. 4b), and the branch-like triangular Altan neretü oboγa which is bigger than the nearby monastery (Bandida blama-yin süme) on cat. 847 might be wooden ovoos (fig. 4c).

• 34 As we will see, rectangles might actually represent steles; then they would not be abstract symbols
• 35 On cat. 773, almost all the mountains have a pointed summit, which is not the case of cat. 764 of t (...)

21At the other extremity of the continuum, abstract, geometrical signs include dots, round marks (fig. 5d), ovoid or almond-like shapes (figs 8, 12, 31b), orange triangles topped with a circle, pointed triangles (fig. 21a), rectangles (for boundary-marking ovoos only, fig. 6)34, and so on. Discrete triangles (fig. 18), dots, or one or two dashes crowning mountains (fig. 5b, c) and pointed summits evoking a pole35, as well as, perhaps, big boulders (figs 5a, 21b, 27a) may indicate ovoos or the main peak of a mountain. As we will see, the depiction of boundary-marking ovoos is often more abstract, while cult ovoos are depicted in more pictorial forms.

• 36 Here are some examples: Qoyitu sengkir-yin oboo (cat. 677 and cat. 678) in Qobdo Uriyangqai, Boro γ(...)
• 37 Cyr. Mo. Altan ovoo, in present-day Sühbaatar Province (see Jessica Madison Pískatá’s article in th (...)

22Many mountains (with or without a drawing of ovoo on their summit) are also called “ovoo”36: about 11% of the toponyms having ovoo in their names are mountains (estimations from Haltod 1966) (fig. 23b). One of the most worshiped mountains of present-day Mongolia is the Altan oboγa of the Darigγanγa37. Therefore, ovoo could be synonymous with a “mountain” and more particularly, a “sacred mountain”. A large number of ovoos are depicted and/or named on mountain and hills’ summits; when their name ends with “oboγa” it can be either the name of the mountain or the name of the ovoo, or both.

• 38 Ongγočatu-yin eki kösiγe-yin dabaγan-u oboγa (cat. 726), Ataγan-yin dabaγan[-u] oboγa (cat. 721, ca (...)
• 39 Onon γool-un oboγa (cat. 722), Γool oboγa (cat. 713), Γool-un kebtege-yin oboγa (cat. 716, cat. 725 (...)
• 40 Čaγan nuur-yin oboγa (cat. 719, cat. 835).
• 41 Qudduγ-un oboγa, Bulaγ-yin [-un] oboγa (cat. 777-cat. 779), Aγuyitu bulaγ-un oboγa (cat. 751-cat. 7 (...)
• 42 Aman-u usun-u oboγa (cat. 708), Aru uγtaγal-yin usun-u oboγa (cat. 716).
• 43 Ovoos depicted near rivers: cat. 675, cat. 798, cat. 803, cat. 805. On cat. 778 (fig. 20), boundary (...)

23A minority of ovoos (both boundary-marking ovoos and cult ovoos) are named ovoos of mountain-passes38, rivers39, lakes40, springs and wells41, or “water” ovoos42. These are depicted on a mountain top or on a pass (fig. 6) or near a water body43.

• 45 Fines were in livestock with rates according to social rank (Bello 2015, p. 120).
• 46 Constant (2010, p. 75, note 66) quotes a memoir addressed to the Grand Council (Junjichu 軍機處) dated (...)
• 47 From historical documents preserved in the National Archives of Mongolia, Cholmongerel studied conf (...)
• 48 The “Kotwicz IV” map of the Tüsiyetü qan ayimaγ dated 1805, studied by Inoue (2012, pp. 221-226) ha (...)

25Mongols were forbidden to trespass their banner’s boundaries and incurred severe penalties if they were caught outside of their banner without a permit. All who were found to have crossed a boundary were interrogated by banner authorities and brought back to their banner for further punishment45. Yet, Anne-Sophie Pratte (2021) argues that it is only after the promulgation of new regulations of 1864 that the Qing administration succeeded to draw clear boundary lines between banners46. Before that date, although several edicts had ordered the construction and annual inspection of boundary-marking ovoos (Constant 2010, p. 76), the banners’ boundaries were not precisely marked and the interdiction of crossing them was probably on paper only, except when ovoos were especially built to mark the borders after the resolution of a conflict47. The oldest maps depict very few or no boundary-ovoos. For instance, the 1780 map of Sečen qan ayimaγ (National Library of Mongolia; Pratte 2021, chapter 2) and a map of Γomboǰab Banner in Sečen qan ayimaγ dated between 1793 and 1805 (Futaki & Kamimura 2005, map M012) do not show boundary-marking ovoos48.

• 49 Heissig 1944, pp. 130-131; Constant 2010, pp. 76, 79.
• 50 See Chomongguriru 2014, pp. 102-103. Van Hecken (1960) recounts a century of territorial conflicts (...)

26After 1864, boundary-marking ovoos were more systematically built, which impacted land-use and migrations in the Qalqa banners (Pratte 2021). Their location was marked on banner maps together with precise descriptions of the boundaries, and distances to places of the neighboring banners. The banner officials were required to check annually that the ovoos were in place and rebuild them49, and to check that their location corresponded to their location on the banner maps. Ovoos could collapse (especially when built in other materials than stones) or be purposefully destroyed during territorial conflicts and rebuilt at another place50. It was sometimes necessary to appoint an official to guard them (Van Hecken 1960, p. 301). The location and names of the boundary-marking ovoos then became the most important elements for the knowledge of banner boundaries.

• 51 This term is found in the Qianlong period: Heissig 1944, p. 130. Also: kiǰaγar neyilegsen γaǰar-tu (...)
• 52 Davaa-Ochir 2008, pp. 52-53, quoting Sühbat & Luvsan Darjaa 2004. According to B. Tseden-Ish: “Any (...)
• 53 These had to be periodically rebuilt because they collapsed easily (Heissig 1944, p. 130).
• 54 Chagdarsurung (1975, p. 365) translates an inscription on a “boundary stake” (Cl. Mo. qosiγun nutuγ(...)
• 55 The Iledkel šastir mentions the erection of marker pole/wooden post (temdeg modon) to mark the boun (...)

27Boundary-marking ovoos were called “temdeg/temdegtü oboγa” (marker-ovoo)51, or “nutuγ-un ǰiqa-yin oboγa” (ovoo [marking] a territory’s boundary). According to Davaa-Ochir, they “consisted with three ovoos. Two lateral oboos are called khyazgaariin oboo [Cl. Mo. kiǰaγar-un oboγa] (frontier oboo) marking the boundaries of two territories, while the central ovoo is called khaich oboo [Cl. Mo. qayiči oboγa] (junction oboo)”52. They usually were stone cairns but could occasionally also be made of wood, of small stones, sand or earth (fig. 7a, b)53. Their name was inscribed on a stele54 or on the central wooden pole of the stone ovoo (fig. 7c). A photo of a boundary-marking ovoo shows a high stone or earth mound with no poles or arrows; it is not called “ovoo” but “düise” (Mongolized transcription of duizi 堆子, “pile, heap”) (Van Hecken 1960, p. 277) (fig. 7a). The boundary could also be marked by an inscription of a stone stele or a wooden post (payiǰa-yin oboγa) with or without an ovoo55 (fig. 7b, c).

• 56 In addition to conflicts between banner princes, land conflicts with Chinese farmers multiplied in (...)
• 57 Bello (2015, chapter 3) studies cases of boundary modifications and reallocation of pastures follow (...)
• 58 The legislation of the Qing state progressively evolved from a right based on the person to a right (...)
• 59 In 1739, a meeting that gathered an imperial emissary, the head of the confederation and the banner (...)
• 60 In one report, some sections of a boundary have not been clarified at the time the document was iss (...)

28A main task of the banner administration was to adjudicate lawsuits caused by land disputes deriving from ambiguous banner boundaries56. Frédéric Constant argued that the banner maps were not just documents destined for the central government. Maps could serve or were sometimes made to solve a conflict on a banner’s boundaries, and “the drawing of the maps became a priority for the authorities of the central administration only when precise knowledge of boundaries was a public policy issue”, for instance when the Yellow River forming a boundary changed its course (Constant 2010, p. 78)57. When a conflict was reported to the central administration, the latter first checked if maps were produced earlier and preserved in the archives of the Lifanyuan (Constant 2010, p. 79)58. The banner maps were then a proof of the state of the boundary at a given moment. In other cases, the dispute was solved by the drawing of a map (or maps) and the laying of boundary-marking ovoos. For instance, to solve a conflict between the Tümed and Dalad Banners, the central administration asked each party to draw a map and then checked the correspondences of the toponyms of the two maps (Constant 2010, p. 78)59. The banner princes then affixed their seals upon the texts specifying the boundaries of the banner, signifying that they had validated this essential information of the map. Maps can include annotations indicating the existence of boundary conflicts (Kamimura 2005, p. 5, M004). Some reports also evoke conflicts about the boundaries with adjacent banners60.

• 61 “The rights conferred by a map could be called into question by a de facto situation. […] The right (...)

29For Charles Bawden (1979, pp. 580-581), the maps were not precise enough to allow the resolution of major boundary conflicts. Quoting a document that indicates pasture allocations, he concludes that “in itself the map and survey system would seem to have been insufficiently developed to cover all disputes which may arise”: the map actually represented an ideal, formal situation. Conversely, Constant argues that the map (and the report accompanying it) was one of the evidences used to settle disputes. The officials undertook “an investigation close to the criminal investigation” to check the adequacy of boundaries to their layout on the map and determine the membership of the populations concerned, but in the end, they usually had to acknowledge an existing situation61. The banner maps were therefore the result of a process of negotiation between the central authority, the reigning local prince, the chief of the league, and the Mongol herders.

• 62 The word nutuγ, “homeland”, designated both a family’s seasonal pasturelands and the banner to whic (...)
• 63 During the Mayidari/Maitreya festival in Mongolia, the procession of the Maitreya statue around the (...)
• 64 The banner monastery (qosiγun-u küriye) was the main and biggest monastery of the banner and had ju (...)

32The territory of the banners was therefore circumambulated and sanctified by these processions, which typically were undertaken in case of epidemics. Hence, the boundaries of the banners acquired a concrete status, delimiting a recognized entity in people’s mind. Anthropologist B. Lindskog (2016) recently observed in Mongolia (Arhangai Province) that the circumambulation of the boundary of the homeland (nutuγ) marked the end of the ovoo offering: a truck loaded with sutras drove around the homeland to purify it against natural catastrophes such as devastating weather phenomena (ǰud). In Qing times, a sentence for criminals consisted in circumambulating the territory of the banner with irons. The boundary-marking ovoos “closed” the territory, with the function of encompassing, containing (fortune) and protecting the homeland (nutuγ) against evil influences62; they formed a symbolic belt like the stupas, stone images and prayer-wheels on the circumambulation path that circled monasteries63. Hence, individually these boundary-marking ovoos may not have a religious function, but collectively they delimited a homogeneous territory that was recognized as such by people who inhabited it. The centralized administration of the banner with the banner’s office (yamen), prince’s residence, banner monastery64 and banner ovoo, as well as boundaries marked by ovoos fostered a sentiment of common belonging to one’s banner as one’s homeland (nutuγ). As Johan Elverskog argued in his book Our Great Qing, in the late-19^th century, the Mongol banners defined by the Qing became sources of local identity, and this “banner localization” supplanted the pan-Mongol identity based on the imperial heritage (2006, p. 127). Although such sentiment of common belonging was non-existent in some banners, as Samuel Bass argues (this volume), this does not appear on banner maps, that show an ideal situation from the viewpoint of the banner administration.

• 65 Banners were maintained in several parts of Inner Mongolia that were not turned into municipalities (...)
• 66 Dabaγa, “mountain-pass”, also means “difficulty, obstacle”. When crossing mountain-passes, Tibetans (...)
• 67 Quoting Dorémieux (2002), Hamayon (2020) contrasts “public” ovoos of mountain-passes, boundaries an (...)
• 68 On ovoos built to suppress bad influences, notably at the “mouth” of rivers and valley, on narrow m (...)

33We have little information on boundary-marking ovoos, which have theoretically disappeared with the administrative reorganization and the suppression of banners in 20^th century-Mongolia65. In mountainous territories, were they built on mountain summits (or slopes) or on mountain-passes? Both locations are depicted on banner maps, where boundary-marking ovoos are often called “peak ovoo” (oroi-yin oboγa) or “mountain-pass ovoo” (dabaγa-yin oboγa). On the one hand, it would be logical to build them on elevated spots visible from afar; on the other hand, if erected at mountain-passes, they were immediately visible for those who crossed the boundary. It is possible that some ovoos which nowadays are located on mountain-passes, at the entrance of a valley and on roads and tracks (dabaγan-u oboγa and ǰam-un oboγa) were former boundary-marking ovoos. When the herders seasonally move from one pasture to another, they usually circumambulate mountain-pass ovoos that delimit their seasonal pastures and mark the entrance to a different territory, and deposit offerings on them. This ritual is an act of declaring to the local deities that they are coming to a new pasture and asking protection for their family and livestock (Davaa-Ochir 2008, pp. 53-54). Because of climate and altitude, mountain-passes are dangerous to cross in winter66; it is also believed that they are places of circulation of angry souls and ghosts of people who died of violent death67. Like stupas, ovoos are built on these dangerous spots of mandatory passages to calm down angry spirits68. Mountain-pass ovoos thus mark places at the edge of different geographic and symbolic worlds. Their localisation pattern is also related to a discourse of regulating access to pastures and other land rights. It is therefore logical that the Qing chose this object to delimit banners’ boundaries.

• 69 Alaγ aγula-yin qoyitu degere oboγa (cat. 709); Baraγun emüneki tala-du ǰiqa-yin oboγa (cat. 732); A (...)
• 70 Chagdarsurung (1975, pp. 31-59) translated and published a report (čese); Soninbayar transcribed a (...)
• 71 This is very clear on maps of the banners of Sečen qan, for instance the same names of boundary-ovo (...)

35Boundary-marking ovoos are represented by very different kinds of symbols, but their name is always written near the symbol or attached to it by a dotted line. They sometimes have long descriptive names indicating elements of the boundary or explaining their location on a mountain, such as “summit obo located on the north side of a mountain”69. In addition, precise indications are written near the boundary line such as the names of the adjacent banners, “bordered by the territory of” (nutuγ-luγa ǰiqa neyilümüi), and the directions and distances to places of the neighboring banners. The accompanying report (čese) also details the names of the boundary-marking ovoos and other elements of the boundaries, and the distances and directions separating each of the ovoos70. Of course, the same boundary-marking ovoos are supposed to be depicted on the map of a banner and of those of its neighboring banners, which is confirmed by the comparison between maps of the same year71.

• 72 In Qing Inner and Outer Mongolia, adjacent to – or included in – Mongol princely (ǰasaγ) and monast (...)

36Boundary-marking ovoos not only marked the boundaries between banners’ territories but also the boundaries of the pastures of the relay-stations inside banners, as well as imperial hunting grounds and other “reserved” or “prohibited areas”72. Border-posts and ovoos delimited buffer zones to guard the external border of the empire with Russia, as well as some specific pastures. Let us examine how these different types of ovoos are represented on maps.

• 73 The 1890 mapping instructions required that the scale of maps was so small that it was necessary to (...)

37The main concern of the maps of Outer Mongolia is the clear delimitation of boundaries, to the point that some maps are reduced to a boundary line with symbols and names of ovoos, and are almost empty inside (see for example cat. 727)73. On many banner maps, mountain ranges, lakes and rivers form parts of the boundaries; these natural landmarks can be connected by a black dotted line or a continuous red or black line. Some maps do not show any boundary line but have isolated landmarks.

• 74 Sometimes with intermediary annotations; for instance, another symbol on the boundary is a Y-shaped (...)
• 75 According to the 1864 regulations, this system following the sexagenary cycle was supposed to repla (...)
• 76 With the exception of a map of the banner of Namsarai (Tüsiyetü qan, cat. 729), and a map of the ba (...)
• 77 On the astrological system of twenty-four or forty-eight directions combined with eight colors, of (...)
• 78 Cat. 722 dated 1907, in Chinese only, has twenty-five ovoos but on cat. 723 of the same banner [no (...)

38The most remarkable maps with boundary-marking ovoos are those of the banners of two of the four Qalqa ayimaγs, Tüsiyetü qan and Sečen qan, on which almost all the boundary-markers are ovoos (from twelve to one hundred and twenty)74. The great map of Sečen qan ayimaγ (cat. 733) has about three hundred boundary-marking ovoos. On the banner maps of this ayimaγ, the ovoos are all numbered, starting from the bottom or from the left, in the clockwise direction (or sun direction, nara ǰöb, like the way of circumambulation of temples and stupas), as are border-posts (on their maps, the Chinese number the border points in the other direction) (fig. 8)75. On the maps of the banners of Tüsiyetü qan ayimaγ, ovoos are not numbered76, but on six out of twenty-three maps, the banner is surrounded by a circle indicating the twenty-four directions77 (the directions were all identified by the names of the twelve animals of the zodiac to locate ovoos in relation to one another; the number and directions of the boundary-marking ovoos rarely correspond to that of the circle) (fig. 1b). On twelve other maps, there is no circle but the texts are radiating from the center, as if there was one. Maps of the same banner drawn at different dates usually have the same ovoos at the same location, but later maps can add new boundary-marking ovoos78.

• 79 On older maps of the Sečen qan aimaγ (dated between 1843 and 1860) preserved in the National Archiv (...)

39Boundary-marking ovoos are depicted as small abstract symbols on the boundary line: small dashes or black or red dots, red circles, small rectangles, or small triangles with their name written near the symbol79. On several maps of the banners of Sečen qan ayimaγ, their depictions are more pictorial: as small circles on top of a triangular shape symbolizing a mountain or as green mounds with a black outline resembling the mountains inside the banner as shown in fig. 9. Another of the abstract symbol for boundary-marking ovoos resembles a red flame or an almond-like shape located on the bordering mountains (fig. 20).

• 80 For instance, a map of the three banners of the New Torγuud (cat. 675) has seven boundary-marking o (...)
• 81 One map has no boundary-marking ovoo. When many papers with toponyms written in Chinese are pasted (...)
• 82 Two maps depict some ovoos as a round mound topped by two rectangles with two adjacent round summit (...)
• 83 On an old drawing of the town of Kiakhta facing Maimaicheng studied by Dear (2019), two stele-like (...)

40On the maps of banners of J̌asaγtu qan ayimaγ, Sayin noyan qan ayimaγ, Qobdo and Uriyangqai, as well as two maps of banners of Xinjiang, only a few boundary-marking ovoos are named and depicted, and large portions of the dotted lined boundary have no ovoos80. The boundary-marking ovoos of the maps of the banners of J̌asaγtu qan and Sayin noyan qan ayimaγs’ banners are located only on the borders with other ayimaγs or with the Qobdo frontier; often there are only two or three ovoos up to as much as thirteen81. An explanation of the small number of ovoos may be the mountainous character of these western regions, with several mountain ranges (the Altai, the Qangγai), which may have made boundary-ovoos less necessary than in the open steppe and semi-desert regions of the Sečen qan ayimaγ. These boundary-marking ovoos are depicted as two elongated rectangles of different sizes, drawn directly on the boundary line or on a mountain or a pass82 (figs 6, 10). These two rectangles possibly represent two steles (fig. 7c)83.

• 84 The eastern ovoo marking the boundary with Aru qorčin and J̌arud Banners is guarded by a border-pos (...)
• 85 Four of them are called in Chinese “northern ovoos” (Ch. Ch. Beifang ebo 北方鄂博) on the north-west bo (...)

41On maps of Inner Mongol banners, the banner often occupies the whole rectangular frame or the greatest part of it, as if the territory had a rectangular shape, and often, no boundary-marking ovoos nor lines are represented (fig. 1c). Mountain ranges, lakes and rivers, the Great Wall and palisades form parts of the boundaries. Maps of the banners of eastern Inner Mongolia have many place names marking their boundaries but few of them are names of ovoos. A map of Aru qorčin Banner (cat. 805) has two drawings of brown mounds with tufts named “oboγa” (fig. 11a) on its north-east, north and east boundaries which are similar to the drawings of cult ovoos with tufts inside the banner 84, as well as eight small brown mounds without a tuft, also named “oboγa”, on points of the boundary (fig. 11b)85. These depictions of tuft-less mounds might be abstract symbols of boundary-markers, or rather drawings of big earth mounds like the boundary-marking ovoo on a photograph of south Ordos (fig. 7a). Besides, trees also seem to function as boundary-markers on fig. 11b.

• 86 Fig. 12b shows each of the twelve boundary-marking ovoos as a mound topped with a flag located on a (...)
• 87 On cat. 831 (Maγu mingγan), mountains are also topped by boulders but the four cardinal directions (...)
• 88 On maps of the banners of Sili-yin γool and Ulaγančab, there are many landmarks but few of them are (...)
• 89 One mountain is called Pai-yin oboγa; the others have names of mountains and cliffs (aγula, qada).

42Most of the other maps of Inner Mongol banners make no visual distinction between boundary-marking and cult ovoos, and use various degrees of abstraction. Naturalist drawings and mimetic symbols are common (figs 12b, 13b)86. Big boulders on mountain summits may represent ovoos or main mountain peaks (fig. 12c)87. Sometimes there are only names of ovoos and no drawing88. Maps of the banners of J̌irim have no drawing or name of boundary-marking ovoos, but eight red circles or “flames” crowning mountains on cat. 782 and cat. 783 could symbolize boundary-marking ovoos (fig. 12a)89.

• 90 Tseden-Ish ([1997] 2003, pp. 303-309) provides reproductions of images showing examples of various (...)
• 91 On the map, čongyi(?) is translated in Chinese by tan 潭, “pool, depression”. The second anonymous r (...)

43A map of Eǰin γool (cat. 673) has two ovoos depicted as small tufts atop mountains on its border with Sayin noyan qan ayimaγ; between them are two mounds topped by two rectangles (possibly steles)90, named Ulaγan čongyi(?) and Barlarqai čongyi(?)91 (fig. 14a).

• 92 By 1765, a total of 73 frontier border-posts had been set up along the Qalqa-Russian border. Each b (...)
• 93 On maps of the Sečen qan ayimaγ of the first half of the 19^th century, the line of border-posts is (...)
• 94 A map reproduced by Chuluun (2014, pp. 116-117) represents the buffer zone between Tüsiyetü qan ayi (...)
• 95 This map represents nineteen princely (ǰasaγ) banners and three monastic banners; the southern bann (...)

44The system of border guard-posts (qaraγul)92 was established following the Treaty of Kiakhta with Russia, to control the Qing-Russian frontier. On maps of the northern frontier of the empire, border-posts are usually marked by an abstract symbol – a cross – with their name written next to it (figs 15b, 16)93. Some maps focus on the frontier buffer zone guarded by border-posts between the Qing and Russian empires94. The general map of the J̌asaγtu qan ayimaγ (cat. 688) shows the buffer zone between the Qing and Tangnu Uriyangqai, with seven crosses indicating border-posts linked by a dotted line (fig. 16)95. Four other crosses indicate border-posts on its southern part, and fifty-two-rectangle signs symbolize boundary-marking ovoo on its other borders.

• 96 For S. Chuluun, ovoos were also built at border-posts (2014, p. 117).
• 97 Šabis were lay families of serfs who belong to the estate of a reincarnated lama (qutuγtu) “with a (...)
• 98 See another map of Tuva: Chuluun 2014, p. 103.

45Devon Dear (2019) writes that there were no ovoos at border-posts but border-posts alternating with ovoos along the border96. This is obvious on the south border of cat. 685 (four banners of Tangnu) with Qobdo-Dörbed Left and Right Banners, where eighteen crosses written “qaraγul” (border-post) alternate with twenty-seven single-rectangle symbols (possibly representing a stele) called “ovoo”; this line of border-posts and ovoos (along with two relay-stations) crosses the southern part of the banner (fig. 15b). Its northern border with Russia is marked with thirteen two-rectangle ovoos; its eastern border with the Darqad monastic estate (Šabi)97 of the J̌ebčündamba qutuγtu has neither ovoo nor border-post98. On a map of the banner of Čedengdorǰi (J̌asaγtu qan, cat. 693), two crosses symbolize two border-posts that delimit a buffer zone with the Köbsgöl frontier (fig. 1a, top left of the map).

• 99 A map of the Dörbed left and right banners of Qobdo (cat. 683) has six two-rectangle white symbols (...)
• 100 On the border with Russia the inscriptions read: “Oros man-u Uriyangqai ene(?) ǰiqa kiǰaγar neyileg (...)
• 101 Five are called “qaraγul”, six have no name and one is called “obo”.

46Another abstract symbol can represent both boundary-marking ovoos and border-posts: two rectangles of different sizes (maps of Qobdo and Uriyangqai, and two maps of Xinjiang)99. Cat. 687 (Uriyangqai of Köbsgöl Lake) has thirteen two-rectangle boundary-marking ovoos circling its borders (including the border with Russia)100 and nine two-rectangle border-posts on two lines that cross the banner (fig. 15a). A banner map of Altai Uriyangqai (cat. 676, dated 1920) highlights new and older borders with Russia following two territorial transfers with Russia in 1869 and 1883 which are explained in the Mongolian and Chinese inscriptions. The new borders are marked by a red line (western border, 1869) and a red dotted line (1883 northern border); the twelve active border-posts and ovoos are marked by two-rectangle symbols painted in white101, while the eleven former border-posts of the older borders are painted in blue (Heissig & Sagaster 1961, p. 348) (fig. 10c).

• 102 Relay-stations were located about 30 km from each other. Travelling on post roads was easy and fast (...)
• 103 Yeke küriye, known to the Russians as Urga, Cl. Mo.  Örgöge, ‘residence’ [of the pontiff]), was the (...)
• 104 Relay-stations were manned by conscripts and their family; thirty-five were operated by the Qaračin (...)
• 105 This is not the case of a map of Vang Banner in Ordos (cat. 841): the territory of a relay-station (...)
• 106 On older maps, relay-stations are symbolized by rectangles (see map M002 dated 1892, kept in Japan, (...)

47Ovoos also demarcated special territories inside banners. First, they marked the boundaries of the territory of the relay-stations (örtege). Relay-stations were especially numerous on the road linking Qobdo to Uliyasutai and Yeke küriye/Urga; trade caravans followed the same roads102. Two maps of the Berlin collection represent a line of relay-stations inside banners; ovoos mark their boundaries (cat. 731: from Yeke küriye103 to Kiakhta, cat. 732: part of the route to Kalgan); each relay-station has eight to twelve ovoos around its territory104. But the scale of maps depicting banners and groups of banners was too small to depict the territory of relay-stations105, that were usually symbolized by triangular signs connected by a dotted line (similar to that of the banner’s boundaries) on maps of Qobdo-Uriyangqai and the Qalqa ayimaγs106.

• 107 Qoriγ designates territories where it was forbidden to hunt, cut trees, graze herds, cultivate the (...)

48A second kind of special territory inside banners is the imperial hunting ground. Imperial hunting grounds belong to the category of forbidden territories (qoriγ, čaγaǰalaγsan or čaγajilaγsan γaǰar)107. Čaγaǰalaγsan or čaγaǰilaγsan γaǰar, more specifically, designates a “territory of unique imperial importance, such as the imperial hunting grounds and the lands rich with sable on the border with Russia”; they were patrolled by imperial guards (High & Schlesinger 2010, p. 7). Henry Serruys translates an excerpt from the Lifanyuan zeli 理藩院則例 (Tuojin & Yue Xi (comp.), code published in 1817, cited in the 1895 edition, 10.43a):

• 109 Wang (2019, 2021) studies the Chinese colonization of Ordos: a first boundary was set up by the Lif (...)
• 110 Enclosed areas opened to Chinese cultivation (identified by an inscription) are visible on maps of (...)

50During the Zunghar (J̌egün γar) conflict (1687-1757), around 1689, the Lifanyuan ordered to erect border-posts to allocate new pastures to Qalqa refugees in the Urad banners (Ulaγančab, Inner Mongolia) (Bello 2015, p. 129). In the process of negotiation between Mongol herders, Han cultivators and the Qing government, border-posts were also erected to delimit land north of the Great Wall which was rented or allocated to Chinese migrants in Ordos (Otoγ, Üüsin, J̌asaγ, Vang, Jungγar banners)109. Agricultural lands are visible on a few banner maps but are not demarcated with ovoos110.

• 111 They are worshiped by people from the same area using the same water source (Davaa-Ochir 2008, p. 5 (...)

52The banner maps also depict and name a number of ovoos on mountain summits inside the banners. Ovoos to worship water spirits (luus-un takilγa-tai oboγa) near rivers (γool-un oboγa) and springs (rasiyan-u oboγa, bulaγ-un oboγa)111, near lone or oddly shaped trees, special rocks and so on, are much less often depicted than mountain ovoos. We may assume that the mountain ovoos were cult ovoos serving as altars or seats (saγudal) to worship master-spirits of the land: all over Mongolia in the modern period, communities were generally based around a sacred mountain dominating the valleys, and propitiated local deities with seasonal sacrifices at an ovoo erected on its summit or highest slopes. The ritual aimed at obtaining the favors of the ambivalent local deities – primarily, rain for pastures and fertility for flocks and herds –, and at asking them not to cause calamities. These territorial cults were central to the life and social and political organization of Mongol communities. However, it is not clear whether all these ovoos were actually worshiped, or if some of them just functioned as specific landmarks rather than religiously important monuments.

• 112 Two maps of the Qorčin Bingtü vang Banner have a curious ovoo (cat. 789: pagoda-like stack of rocks (...)

53Mountain cult ovoos are depicted with many different symbols upon mountains; they are often mimetic drawings that can be reduced to a cone or a tuft112; they can also be abstract symbols marking a mountain, such as dots, dashes, or poles, but the symbols of two rectangles or abstract symbols without a mountain are never used for cult ovoos (figs 2, 3, 5).

• 113 Such as the thirteen ovoos of the Altai and the thirteen ovoos of the Torγuud in western Mongolia, (...)

54Banner ovoos (the main, officially worshiped ovoo of a banner) and many other ovoos often consisted of thirteen ovoos with twelve cairns forming a cross according to the four cardinal points with the bigger ovoo in the middle, or twelve cairns in a line with a bigger one in the middle113. But they were not depicted or named as such: banner maps only depict one (sometimes three) ovoos per mountain (see also fig. 30 that depicts the many ovoos of Bandida gegen süme [fig. 3c] as one ovoo).

• 114 As mentioned above, in many cases “ovoo” can designate “mountain”.
• 115 In some cases, it is difficult to know if the alternately rounded or pointed shape of mountains is (...)
• 116 According to Qing regulations, all the mountains must be represented from the south (as in Chinese (...)
• 117 As expressed by Caroline Humphrey: “for Buryats like other Mongolic people have an acute awareness (...)

55Mountains (ovoos) are important elements of banner maps114. While on the same map, the majority of mountains seem to be depicted with the same conventional, mimetic drawing (sometimes with variations in height and color)115, some are depicted according to their real shape, sometimes with distinctive features116. Map painters were influenced by Buddhist painting and treatises of geomancy in the depiction of landscape (Pratte 2021, pp. 267-276). A curious mountain of the Qaračin Left Banner has a human face (J̌osoto, cat. 794)117.

• 118 Cyr. Mo. Hentii han Mountain in present-day Hentii Province, Mongolia.
• 119 Although sacred mountains were usually called “qayiraqan” because it was forbidden to pronounce the (...)
• 120 For instance, a mountain of Qaγučid Left Banner with a flat top is called “Dösi aγula”, “anvil moun (...)

56The mountains of a few maps of Sečen qan are distinguished by their size, the pointed or flat shape of their summit (cat. 740, cat. 741, cat. 753) and the presence or absence of snow (cat. 761, cat. 762). In the middle of two maps of the banner of Navangsikür (cat. 734 and cat. 735) is a remarkable white (snowy?) mountain topped with a circle resembling a boulder marking its top, while other mountains are green: it is the great Kengtei qan aγula118, the highest mountain of the banner (fig. 5a). Outstanding mountains are often surmounted by one or more ovoos, and are depicted as higher than the others; they are usually qualified by the terms boγda, “holy, sacred”, and qayiraqan, “sacred, merciful, beloved”119. Flat-topped mountains were particularly venerated in Mongolia because they represented the perfect altar; they were called “sirege”, “throne, altar” or “dösi”, “anvil”120 (fig. 18).

• 121 Three of these mountains are called “ovoo”: Čaγan oboγa, Delger oboγa and Tuγul oboγa.
• 122 Cat. 806 (J̌arud Right and Left Banners) and cat. 805 (Aru qorčin) have respectively eighteen and e (...)
• 123 Three maps of Otoγ Banner (Ordos: cat. 848, cat. 849 and cat. 850) show four cult ovoos, six monast (...)
• 124 While cat. 803 (Baγarin Right and Left Banners, 1907) depicts eleven ovoos, cat. 804 drawn a year l (...)

58According to the region, the number of cult ovoos that are depicted varies significantly from one map to the other. The maps of western Mongolia, J̌asaγtu qan and Sayin noyan qan do not depict ovoos inside banners, but many mountains are called “oboγa” (cat. 711 for instance). Conversely, mountain ovoos are especially numerous on the maps of Sečen qan ayimaγ: a fourth to half of their mountain summits are topped with an ovoo. On cat. 768 (banner of Dorǰiǰab; see also cat. 767), nineteen mountains have an ovoo on their top, that is, almost all the mountains (fig. 19)121. Cult ovoos are also particularly numerous in J̌oo-uda, Sili-yin γool and Ordos leagues122. Maps that also have a large number of monasteries superimpose the two sacred geographies, the number and size of ovoos and monasteries being comparable123 (fig. 32b). Different maps of the same banner can show very different numbers of cult ovoos124.

• 125 An exception is the depiction of ovoos and border-posts (qaraγul) as red circles that delimit an im (...)
• 126 On cat. 779 (same banner, in Chinese only, no date), the distinction between boundary-marking ovoos(...)
• 127 The mountains and cliffs of these maps are represented by a mimetic symbol: a central blue mountain (...)
• 128 See also cat. 777 and cat. 779.

60On most of the maps of Inner Mongol banners, the same symbol is used for both boundary marking and cult ovoos125, hence it is not always clear whether ovoos marked along a border are boundary-marking ovoos, cult ovoos, or both. Conversely, maps of the banners of Sečen qan ayimaγ make a distinction between boundary-marking ovoos and cult ovoos, either by using a different symbol or by a detail marking a difference. For instance, on cat. 773 (banner of Sangsarayidorǰi), the boundary-marking ovoos are depicted as small mountains or mounds topped with a red dash symbol while mountains inside the banner are crowned by a black dash or a triangle. On a map of the Čeringnima Banner in Sečen qan (cat. 778), a red almond crowned by a kind of tuft suggesting an ovoo is depicted on Qan aγula, the highest mountain inside the banner126 (fig. 20b). Along the boundary, boundary-marking ovoos numbered from one to thirty are depicted as red flames127. Another ovoo depicted as a red circle topped by a tuft crowns a bigger mountain located right on the border with Kölön Buir League. This mountain, Soyolǰi aγula, is clearly distinct from boundary-marking ovoos; its ovoo has no number, it certainly is a cult ovoo (fig. 20a). On its left slope are two uncolored almonds – the left one being the first numbered boundary-marking ovoo –, and on its right slope, a red flame is the thirtiest boundary-marking ovoo. Therefore, this mountain had its own cult ovoo but also marks the border with Kölön Buir: it is the starting point of the numbering of border-markers (fig. 20a)128. The big tuft differentiates cult ovoos from boundary-marking ovoos.

• 129 Cat. 686 belongs to the family of banner maps of Sečen qan ayimaγ (use of a grid, numbered ovoos an (...)

61On cat. 686 (Darqad monastic estate [šabi] of the J̌ebčündamba qutuγtu in Qobdo Uriyangqai), boundary-marking ovoos are symbolized by yellow pointed triangles129. On the north-west boundary, an outstanding blue mountain called in Chinese “Alagahada” has a yellow pointed triangle on its top, which is identified as the “twelfth Alagahada ebo”: it belongs to boundary-marking ovoos but also seems to be a sacred mountain (fig. 21a).

• 130 In Buryatia, after having worshiped their private ovoo, the men of all the related lineages would j (...)
• 131 An official Qing “state” cult to the four mountains circling Yeke küriye (Urga) started in the late (...)

63Mountain-cult ovoos can be classified according to their territorial and communal importance, from private ovoos of family groups and small communities (such as ovoos of patrilineal clans [otoγ oboγ-un oboγa] in Buryatia) located on their winter pastures, near the tombs of their ancestors, to ovoos worshiped by larger communities, such as banner ovoos in Qing Mongolia, ovoos of monasteries, Qan (“king”) ovoos in Buryatia130, and official ovoos of the sacred mountains of the Qalqa that received a state official cult131.

• 132 The banner’s subunits also worshiped their specific ovoo. Banner and banner’ subunits’ ovoos are co (...)
• 133 Vreeland 1962, p. 127 (Čaqar Taibas/Taipusi pastures); Evans & Humphrey 2003, p. 201 (Urad, Ulaγanč (...)

64In the Qing period, every banner had a main sacred mountain (qosiγun-u takilγa-tai aγula) with an ovoo called “banner ovoo” (qosiγun-u oboγa) on one of its highest peaks: the banner was an autonomous entity fixed on a particular territory with its own sacred landscape (nutuγ). This ovoo represented the banner as a whole. The ruling prince (ǰasaγ) presided over the banner ovoo ritual that gathered the banner’s central administration and its subunits132. This ritual was an important occasion for the reaffirmation of his political power133. It was generally held in conjunction with a banner assembly at which important affairs were resolved (Bulag 2010, p. 175). The monks of the “banner monastery” officiated during the ovoo ritual, which was generally followed by a festival (naγadum, “three manly games”) to entertain the local deities. The banner ovoo was built and repaired by members of the banner as part of their corvée civil obligations (Vreeland 1962, p. 180). It was called “the spirit-ovoo of the banner” (qosiγun-u sülde oboγa) (Davaa-Ochir 2008, p. 15):

• 135 Qosiγun oboγa near the banner monastery (cat. 807), Qosiu-yin oboγa (cat. 757, cat. 758); Qosiun ob (...)
• 136 In Ordos: cat. 842-845, cat. 847, 851-cat. 853.

65On the banner maps of the Berlin collection, Haltod (1966) recorded only three occurrences of ovoos called “banner ovoo”135, and eight occurrences of “Altan obo/oboγa”136.

• 137 Boro oboγa, Grey/Brown ovoo might be the name of the mountain, which is brown colored, thus produci (...)

66Some maps have one prominent mountain, topped with an ovoo: we can assume that it is the banner’s sacred mountain with the banner ovoo; sometimes it is the only mountain-cum-ovoo shown on a map. On a map of the Old Torγuud of Xinjiang (cat. 674) which has a disproportionally big ovoo, we may ask why this particular mountain has been chosen while other mountains are more imposing (fig. 2a). On other maps, the highest mountain with the biggest ovoo stands among many others mountains-cum-ovoos. In some cases, what seems to be the most prominent mountain-cum-ovoo is located in the center of the banner: the mountain named Ölǰeitu aγula crowned by a piled rock ovoo, south of the banner prince’s residence on cat. 759 and cat. 760 (fig. 22); the highest, blue mountain with a big ovoo of cat. 847 (Üüsin) (fig. 23) and the J̌ibqulangtu čaγan oboγa of cat. 837 (J̌üngγar, in the middle of the map) (fig. 24). On a map of Eǰin γool (cat. 673), there is an ovoo on an outstanding vertical central mountain, Bayan ǰirüke (fig. 26), and a second one, Boro oboγa north of Lake Sub naγur137 (fig. 14a). Cat. 810 and cat. 811 (Üǰümüčin Right Banner) have a big mountain (Ölǰitü [Ölǰeitü] uru oboγa) crowned by an ovoo depicted as a mound topped with a tuft and pole, located near the center of the map, just north of the banner prince’s encampment, while other mountains called “ovoo” (“oboγa” or “oboγatu”) are topped with a big boulder (fig. 25a). Evans and Humphrey write that the site of the banner oboo “was presumably chosen for being central in the Banner’s territory” (2003, p. 203). The main sacred mountain and its ovoo may also be depicted near the center on the map to construct a more auspicious narrative.

• 138 In the lower left part of the map there is another mountain with a smaller boulder on its peak, beh (...)
• 139 The same map has an ovoo depicted in a more naturalistic way.

67On several banner maps, prominent mountains located near the center of the banner are crowned by a big round boulder that may designate the main sacred peak or the main ovoo of the banner (or both). For instance, a map of the banner of Navangsikür, Sečen qan (cat. 735) with the Kengtei qan Mountain (fig. 5a); a map of Tümed Banner in J̌osoto (cat. 793138); a map of Ongniγud Left Banner (cat. 801: Bayan oboo139) and a map of Čoqor Qalqa Banner in J̌oo-uda (cat. 797: Dalu-yin aγula); a map of Sünid Right Banner (cat. 826 and cat. 828: Bayan oboγa; nearby is an inscription indicating the settlement [nutuγlal] of the banner prince [ǰasaγ vang]) (fig. 27).

• 140 A few Chinese-style palaces were preserved in both Mongolia and Inner Mongolia. For a description o (...)

68One of the most important element of banner maps is the palace or encampment and banner administration (yamen) of the banner prince, or, for monastic banners, of the reincarnated lama140. It is not clear which seasonal location of the encampment is depicted when the residence was mobile. On banner maps, the prince’s palace or encampment is often located in or near the center of the map, not far from the banner’s monastery, symbolizing the union of politics and religion (qoyar yosu, “two laws” or “two systems”). Kollmar-Paulenz (2006, pp. 371-372) partially attributes the map’s centrality of administrative and religious establishments to Chinese cultural influence onto Mongol cartography. Although some banners may have had a centrally located banner prince’s seat or monastery, in other cases, the political seat may have been moved to the map’s center for auspicious reasons.

• 141 See also a map in the banner of Qurča vang in Sečen qan ayimaγ dated 1913, with the residence of th (...)

69In some cases, when we can identify the main sacred mountain, the residence of the banner prince is located in its proximity (figs 22- 25, 27a). It can also be backed by a smaller mountain (geomantical rules prescribe that palaces and temples should be built at the foot or on the southern slope of a mountain). On a map of J̌arud (cat. 806, J̌oo-uda), three buildings of the residence of the Imperial prince of the second rank (ǰasaγ törö-yin giyün vang) stand at the foot of a mountain with a big ovoo, Erdeni čoγca aγula (fig. 28a). On cat. 814 (Qaγučid Right Banner), the name of the palace of the banner prince (not depicted) is written at the foot of a great mountain with an ovoo, in the center-right part of the map (fig. 28b). On cat. 830 (Qalqa Banner, Ulaγančab) a mountain topped with an oval-shaped ovoo towers above the inscription indicating the banner prince’s palace and temple141.

• 142 On cat. 816, the mountains-cum-ovoo are depicted but the monasteries are just named.
• 143 Cyr. Mo. Bogd han Mountain, south of Ulaanbaatar.
• 144 Cyr. Mo. Tsetse gün.
• 145 Boγda qan was the most important of the four mountains protecting Yeke küriye. On the map, south of (...)
• 146 Other examples include, in Qalqa Mongolia: cat. 775 and 776 (Dorǰipalmu, Sečen qan: three monasteri (...)

71Before building a monastery, lamas travelled over the mountains and through the valleys to find the ideal place according to geomantic prescriptions and pragmatic considerations (near a river, with southern exposure, sheltered from the northern winds, etc.). A typical auspicious location was the southern slope of a mountain or hill (Charleux 2006, pp.155-159). The monastery then erected an ovoo to worship the genii loci of the mountain or hill. Monasteries’ ovoos are depicted on many maps (figs 29-30). On cat. 816 and cat. 817 (Abaγa Left and Abaγanar Left Banners), the Sayin-i erkilegči süme (Ch. Chongshansi 崇善寺), i.e. the great Bandida gegen süme/Beizimiao 貝子廟 is backed by the Erdeni oboγa (fig. 30; see fig. 3c), and the Soyol-i badaraγuluγči süme, by the Delger oboγa142. On cat. 720 (banner of Pungčuγčering, Tüsiyetü qan ayimaγ), a red marked spot crowns Qan aγula, i.e. Mount Boγda qan143, south of J̌ibǰündamba (J̌ebčündamba) qutuγtu’s Yeke küriye/Urga (fig. 31). This red spot probably is the Čeče güng Peak144 with the Čeče güng-ün qural, a small temple dedicated to the worship of mountain gods and Chinggisid ancestors145 (Teleki 2011, pp. 258-261)146.

• 147 The present-day seats of the banners usually preserve archives, local gazetteers and, often, a smal (...)

72When comparing old banner maps with modern maps, it appears that the great majority of place names have changed, as well as many landmarks, especially in Inner Mongolia because of 20^th century Chinese immigration: many new settlements, villages and cities with Chinese names were built, the courses of rivers have been diverted because of irrigation, lakes and rivers have dried, even mountain summits have sometimes been levelled by mineral exploitation147. Yet, we still observe a great number of names ending with “ovoo” on modern maps, even in Inner Mongolia where many toponyms were sinicized or recently invented.

• 148 In the two examples below, the great majority of the monasteries has been destroyed.
• 149 Ch. Alukeerqin Banner 阿魯科爾沁旗, Chifeng Municipality 赤峰市.
• 150 Joči, alias Qabutu Qasar, Chinggis Khan’s elder brother, is the ancestor of the Qorčin aristocracy.
• 151 The ovoo was rebuilt in 2003 and every year, a day of the sixth month of the lunar calendar is chos (...)

74Because modern local gazetteers sometimes precisely locate destroyed monasteries, for some banners they appear to be the most reliable landmarks148. I could locate on modern maps seventeen of the twenty-three monasteries that are named on a map of Aru qorčin Banner149 (cat. 805) which also shows eleven ovoos inside the banner and ten boundary-marking ovoos (fig. 32). The banner stretches along a roughly south-east/north-west axis with most of the ovoos on cat. 805 concentrated in the north-western mountainous area (the highest peak on modern maps is Batai oboγa, 1 540 m above sea level). But except for Sira öndör (which is represented north of its present location) and Boγda aγula, the names of mountains and ovoos given by the 1908 banner map are not found on modern maps. Yet having precisely located Γanǰuur süme and Γabču(?)-yin süme, I could approximately locate the Yeke qayiraqan/Da Hai-le-qin Mountain and the other mountains-cum-ovoos north of the banner. The present-day most worshiped ovoo is that of Mount Γuγustai aγula, also called Qan aγula, a National Park and main tourist area. Another main ovoo worshiped in the banner, is the Qabutu Qasar150 oboγa on Boγda aγula, just north of Kundu Village (Ch. Kundu 坤都镇), the small city founded on the ruins of the Bayasqulang čiγulγuči süme and the banner prince’s palace (now in Xin Baolige gacha新包力格嘎查). This was probably the banner ovoo: cat. 805 depicts it just behind the monastery, in the center of the map (fig. 32)151.

• 152 Ch. Kesheketeng Banner克什克騰旗, Chifeng Municipality.
• 153 The official Chinese name of the Bayasqulang amuγulang süme, now located in the banner center (Jinp (...)
• 154 The banner is now well-known for its “Global Geopark”, a protected area covering some exceptional g (...)
• 155 It is worshiped on the thirteenth day of the fifth Chinese lunar month (Lonely Planet 2018, p. 216)

75A second example is a map of Kesigten Banner152 (cat. 802, J̌oo-uda league), with three ovoos in its center (Bayan ǰirüke oboγa, Bayan čaγan oboγa, and the ovoo of a monastery), and one on each of its south and north boundaries (on Sayiqan aγula and on Qongγor oboγa). The highest peak (Kesigten aγulan-u nutuγlaγsan(?) Kirbis qada) of the historical map, located “behind” the prince’s palace (now Biraγu qota, Ch. Jinpeng 金棚 City), has no ovoo (fig. 33: this peak may not be a steep massif but a minor mountain153). Again, except for the main monasteries that are precisely located in local gazetters, most of the landmarks have changed. The Qongγor oboγa of the banner map may be identified with Mount Qongγu-yin dabaγa, the highest summit of the banner, that culminates at 2 029 m154. The present-day banner ovoo, known as Bayan oboγa (Ch. Furaoshan 富饶山)155, it is located at 1 498 m above sea level on a plateau (Γongγor steppe 贡格尔草原) surrounded by forests.

• 156 On geomantic defects: Charleux 2006, pp.155-159. The same is observed for Chinese maps (Yee 1994, p (...)
• 157 We can also assume that in some cases, ruling princes have tried to depreciate their territory so n (...)

76Maps are cultural productions that construct an image of the visible world which favors specific interests, and “culturalize the natural through processes of denomination and identification, of categorization and inclusion” (Smith 1998, p. 57). As for Chinese maps, the attention paid to the high accuracy of representation on Mongol maps is remarkable, yet it contrasts with the tendency to deform the spatial distribution of objects in order to construct a more auspicious visual narrative: they mix concrete and idealized geographies. The banner maps sometimes depict places that one thinks they are or would like them to be – for instance, a hill can be depicted as a high mountain, landscape features can be arranged to correct a “geomantic defect” (such as the absence of watercourse or a mountain to the south)156, and the banner prince’s residence and the main sacred mountain and banner ovoo can be relocated in the center157.

• 158 More research needs to be done. Hopefully maps from the collections of Mongolia will be digitalized (...)

78This paper tried to shed new light on details of the various modes of the ovoos’ presence in Mongolian cartographic representations of space158. Ovoos were ancient human marking tools in the open steppe, which the Qing appropriated to mark the border with Russia and internal boundaries inside Mongolia. The maps inform us on how the borders were conceptualized, and it was precisely one of their main functions. But why depict cult ovoos on banner maps if this information was not specifically asked by the Lifanyuan? The indication of the most sacred mountains of a banner (such as Kengtei qan) may have been obvious information for map-makers. Cult ovoos were territorial indicators for the supernatural realm considered no less important than the physical one. The local administrations’ worship of mountain deities at banner ovoos was a major religious and political ritual. Main ovoos were probably also used as landmarks (including rivers’, wells’ and springs’ ovoos). Still today, some ovoos are major landmarks, markers visible from afar, and useful for orientation, such as the Binder ovoo in Hentii Province (Mongolia).

• 159 The Inner Mongol banners do not include the Čaqar and Kölön Buir, which saw important forced migrat (...)

80The fact that boundary-marking ovoos are not always represented on banner maps of Inner Mongolia does not mean that they were absent from the landscape itself. It is possible that ovoos were built to mark the boundaries but they are not depicted and their place names do not include the term oboγa; they were rather called by the name of a mountain or a particular landmark. It is also possible that the territories of some Inner Mongol banners were more stable over a long period than those in Outer Mongol banners, and also had more coherent geographical boundaries. In Sili-yin γool for instance, the banners had elongated shapes allowing access to the southern Chinese markets of the Great Wall. The number of banners in Inner Mongolia was fixed to forty-nine in the early Qing period and did not change, so the Inner Mongol groups were relatively stable on their territory since the beginning of the Qing period, while in Outer Mongolia, land was redistributed to the Qalqa after the Zunghar wars, the Oirad were relocated west of them, and a new ayimaγ was created (Sayin noyan qan, in 1725). This does not mean that there were less boundary conflicts in Inner Mongolia (see the above-mentioned conflict in the Ordos)159.