Uncovering Amoghavajra's legacy in the Hexi Corridor and Tibet

主编 霍巍 石硕 常务副主编 张长虹 编辑委员会 ( 以姓氏拼音为序 ) 主席 巴桑旺堆（西藏自治区社会科学院） 委员 才让太（中央民族大学） 霍 巍（四川大学） 石 硕（四川大学） 沈卫荣（清华大学） 熊文彬（四川大学） 张 云（中国藏学研究中心） 多吉旺秋（德国汉堡大学） 范德康（美国哈佛大学） 马休·凯普斯坦 ( 法国巴黎高等研究实践学院、 美国芝加哥大学 ) 滕华睿（美国纽约哥伦比亚大学） 谢 萧（法国巴黎高等研究实践学院） 编辑 陈 波 嘎尔让 华青道尔杰（张延清） 许渊钦 杨清凡 玉珠措姆（金红梅） 张长虹 编务 孙昭亮  Editors-in-chief Huo Wei, Shi Shuo Deputy Editor-in-chief Zhang Changhong Editorial Board Pasang Wangdu (Chair, Tibetan Academy of Social Science, China) Tsering Thar (Minzu University of China) Huo Wei (Sichuan University, China) Shi Shuo (Sichuan University, China) Shen Weirong (Tsinghua University, China) Xiong Wenbin (Sichuan University, China) Zhang Yun (China Tibetology Research Center) Dorji Wangchuk (Hamburg University, Germany) van der Kuijp, Leonard W. J. (Harvard University, U.S.A.) Kapstein, Matthew T. (École Pratique des Hautes Études, France; University of Chicago, U.S.A.) Tuttle, Gray (Columbia University, U.S.A) Scherrer-Schaub, Cristina A. (École Pratique des Hautes Études, France) Editors Chen Bo, Gaerrang, Pelchan Dorje (Zhang Yanqing), Xu Yuanqin, Yang Qingfan, Yudru Tsomu (Jin Hongmei), Zhang Changhong Editorial Assistant Sun Zhaoliang  目 录 尼泊尔上穆斯塘地区桑宗墓地所见本教施垛仪轨的考古学证据（英文） … ……………………………………………………………… 马克·奥登德弗 / 1 羊距骨 ：西藏最古老的博具和占卜工具的考古学和民族史初探（英文） … ……………………………………………………… 约翰·文森特·贝勒扎 / 13 唐初的中尼交通 ：资料的再审视… ……………………………………… 王邦维 / 43 在河西走廊和西藏发现的不空之影响（英文）… …………… 尤利·霍赫洛夫 / 48 关于藏文蓝黑纸（མཐིང་ཤོག）写经的初步考察 ：以文献为中心 … ……………………………………………………………………… 朱丽双 / 134 西藏夏鲁寺“介尊·喜饶琼乃”壁画释读…………………………… 贾玉平 / 151 夏鲁寺一层集会大殿回廊中的佛传壁画研究…………………………… 孟瑜 / 163 桑噶地区赤洛纳特寺观音崇拜的艺术史证据（英文）……………… 林瑞宾 / 177 明朝颁赐西藏印章的初步研究… …………………………………………… 李帅 / 205 清代四川藏区土司官印初探… ………………………………………………刘莎 / 227 1 乾隆御用嘎巴拉鼓研究………………………………………………………林欢 / 244 故宫博物院藏乾隆皇帝本生故事集——《法王新胜传》 ………………李若愚 / 273 西藏的胎藏曼荼罗传统及其彩沙曼荼罗制作（英文） ……………… 田中公明 / 295 摘要… ………………………………………………………………………………… / 305 2  Table of Contents Archaeological Evidence for the Bon Mdos rgyab Ritual at Samdzong, Upper 1 Mustang, Nepal Mark Aldenderfer Knucklebones, An Archaeological and Ethnohistorical Inquiry into one of Tibet's 13 Oldest Gaming and Divination Objects John Vincent Bellezza Re-examination of Materials about the Communications between Nepal and China 43 in the Early Tang Dynasty Wang Bangwei Uncovering Amoghavajra's Legacy in the Hexi Corridor and Tibet 48 Yury Khokhlov A Preliminary Survey on Mthing shog Manuscripts 134 Zhu Lishuang An Interpretation of the Mural of Lce btsun Shes rab 'byung gnas in Zhwa lu 151 Monastery Jia Yuping On the Mural Painting of Buddha's life in the Corridor of the Assembly Hall on the 163 First Floor of Zhwa lu Monastery Meng Yu 3 Art Historical Evidence for a Cult of the Triloknāth Lokeśvara in Zangskar 177 Rob Linrothe Research on the Seals Granted by the Ming Dynasty to Tibet 205 Li Shuai Research on the Official Seals of the Chieftain (Tusi) in the Tibetan Region of 227 Sichuan Province in the Qing Dynasty Liu Sha The Imperial Kapala Drums of the Qianlong Emperor 244 Lin Huan Fa Wang Xin Sheng Zhuan — Stories of Qianlong Emperor's Incarnations 273 Li Ruoyu The Tradition of the Garbha-maṇḍala in Tibet and its Sand Maṇḍala Creation 295 Kimiaki Tanaka Abstracts 305 4  Uncovering Amoghavajra's legacy in the Hexi Corridor and Tibet 1 Yury Khokhlov Abstract: This research analyses the images from Mogao and Yulin caves, which have been traditionally ascribed to Tibetan influence and believed to represent the early Tibetan style. Contrary to this common opinion, the author shows that these images are in fact influenced by Southern Indian art and in particular by the art of the Pallava kingdom in Tamil Nadu and closely affiliated with teachings introduced to China by Vajrabodhi ( 金 刚 智 , 671-741), a renowned Buddhist master from the Pallava Kingdom, in the first half of the 8th century and specifically propagated in Hexi by Amoghavajra ( 不 空 Bukong, 705-774) in the mid 8th century. He argues that these images belong to a distinctive artistic tradition, which was established by Vajrabodhi who was also known in China as a great painter. The presence of these images in artwork from the Tibetan period merely signifies the continuation of the 1 The paper expands on a talk given at the Seventh International Conference on Tibetan Archaeology & Arts in Chengdu (October, 2018). Many people helped me in various stages of preparing it. I would like to thank in particular Tristan Bruck, Analicia Busha, Geoffrey Goble, Shawo Khacham, Rory Lindsay and Darik Thokmay for their assistance with my studies on various topics discussed in this paper; and Emmanuel Francis, Yuko Fukuroi, Amy Heller, Deepanjana Klein and Arno Klein for sharing with me their photographic archives. I am immensely grateful to Claudine Bautze-Picron, Rolf Giebel, Yannick Laurent, Robert Mayer, Neil Schmid, Per Sørensen, Jeffrey Sundberg, Joie Szu-Chiao Chen, Sam van Schaik and Zhang Changhong who read my manuscript, corrected my errors and offered me their extremely valuable feedback. All remaining errors are of course my own. 48 artistic and religious trends that existed in Hexi before the Tibetan conquest. Furthermore, the article argues that the Tibetan conquest of the Hexi corridor enabled the Tibetan appropriation of Chinese religious and artistic trends, which were popular in the region. This conclusion well supports Christopher Beckwith's theory according to which the Tibetan adoption of Buddhism as a state religion was inspired by the Chinese model of state-protective Buddhism established by Amoghavajra. Introduction As a starting point, this research analyses the imagery in a distinctive Indian artistic mode2 found in various paintings from Mogao and Yulin caves.3 The material under review belongs to the 8th to 11th century and often features Esoteric Buddhist imagery.4 Since some of the early examples within this group belong to the period of the Tibetan occupation of the Hexi Corridor,5 their artistic mode has often been ascribed — erroneously, in my view — to the cultural influence of Tibet. Several recent publications on this topic provoked me to address the issues that I see in such attributions.6 Contrary to this common opinion, I will argue that this artistic mode is in fact influenced by Southern Indian7 art and, in particular, certain stylistic and iconographic features can 2 The term "artistic mode" is used as a broader category than artistic style and refers to a general way of representation. Additionally, "imagery in an Indian artistic mode" refers to Indianized images that were made to look as if they were Indian as opposed to images in a traditional local artistic mode, which for the area discussed in this essay is Chinese. 3 A list of these paintings can be found in Appendix 1. 4 The term "Esoteric Buddhism" is notoriously complex and problematic; in this article it is used mostly as a reference to teachings which were introduced to China during the Tang and possessed instructions in the use of mantra, dhāraṇī, and maṇḍala."Esoteric Buddhist imagery" refers here to representations of deities associated with these teachings. An extensive discussion on the term "Esoteric Buddhism" can be found in Orzech, Sørensen, and Payne 2011: 3-18. 5 In this article, I adopted the dates of the Tibetan conquest of the Hexi corridor from van Schaik and Galambos 2012: 62-63. Guazhou ( 瓜州 ), the prefecture where the Yulin caves are located, was captured by the Tibetans in 776, while Shazhou ( 沙州 ), Dunhuang, fell in 786. After the collapse of the Tibetan empire, the whole Hexi Corridor was once again under Chinese rule by 851. 6 Recent publications that argue for the Tibetan origin of the discussed imagery include Kapstein 2004, 2009, 2014; Sha 2013; McCoy 2016; M. Wang 2018; Debreczeny 2019. See footnotes 53, 134 and 144 for details. 7 Geographically, the Indian subcontinent has been divided into the Northern and Southern parts. While the Northern part includes the northern mountains and the Indo-Gangetic plains, the Southern part is comprised of the Indian Peninsula. An imaginary division line between the two parts goes along the Vindhya Mountains and the Satpura Range, which separate the Gangetic plains and the Deccan plateau (see Map 1, where it is shown as a black dashed line). The two parts of the subcontinent, being interconnected on many levels, are, nevertheless, characterized by different histories, economies and cultures, and often by distinctive artistic styles. See Thapar 2003: 37-68. There is yet another distinction, which is particularly important for this research. While the cultural influence of Northern India mostly spread along land-bound trade routes, Southern Indian influence is mostly noticeable in territories with which Southern Indian polities engaged in sea trade. From this perspective, landlocked Tibet belongs to the Northern Indian sphere of influence, while China was naturally predisposed to receive influence from both Northern and Southern parts of India. 49 be traced exclusively to the art of the Pallava kingdom in Tamil Nadu and the Pallava influenced art of the Western Deccan, thereby ruling out alleged artistic influences from Tibet. Moreover, the discussed images are closely affiliated with teachings introduced to China by Vajrabodhi ( 金刚智 , 671-741) in the first half of the 8th century and specifically propagated in Hexi by Amoghavajra ( 不 空 , 705-774) in the mid 8th century. Thus, the presence of such depictions in artwork from the Tibetan period merely signifies the continuation and Tibetan appropriation of the artistic and religious trends that existed in Hexi before the Tibetan conquest. Furthermore, it will be shown that this artistic mode was transmitted to Central Tibet during the Tibetan occupation of the Hexi Corridor and can be recognized in sculptures from a number of early Central Tibetan temples. I argue that this artistic transmission was a result of significant religious transmission, which included, but was not limited to, the cult of Vairocana and the eight great bodhisattvas, as well as the teachings of the Yoga tradition (i.e. Yoga Tantras according to the Tibetan classification). This article is divided into four major parts. Part 1 presents a brief historical overview of Buddhism and artistic trends in Hexi before the Tibetan conquest. Part 2 discusses the most representative depictions within this group of materials and their connection to Southern Indian art and Amoghavajra's textual heritage. In Part 3, I will provide an overview of Southern Indian influence on Buddhist art in China in general and situate this Hexi group of images within a larger picture. In Part 4, I will demonstrate that the artistic mode under review and its corresponding teachings were transmitted from Hexi to Tibet during the period of Tibetan occupation. Sites mentioned in the article are shown in Map 1, which also illustrates major land and maritime routes between South and East Asia in the first millennium CE.8 Part 1. Historical Setting By the Han dynasty (206 BC-220 CE), the Hexi Corridor was already a strategic part of the Silk Road and a gateway to China for those travelling from India and Central Asia. This area was an early recipient of Buddhist teachings and artistic styles arriving to China from the West. Frequent contacts with the West led to a highly cosmopolitan Buddhist culture that encouraged a close imitation of foreign artistic models. Imperial Buddhist patronage and expansion of the Tang dynasty to the West in the Tarim Basin created a steady traffic of monks travelling between China and Northern India through Hexi. Indian monks travelling to China and Chinese monks returning to China from India carried with them Buddhist texts and imagery as tribute to the court. This process resulted in a rapid transmission of the newest — and increasingly Esoteric — teachings and in the popularity of Indian icons and Indianized styles in China in general and in Hexi in 8 Routes on the Tibetan plateau are not shown in the map. 50 Map. 1 Major land and maritime routes between South and East Asia in the first Millennium CE and sites mentioned in the article. Routes on the Tibetan plateau are not shown (based on a map from https:// en.unesco.org/silkroad/about-silk-road). particular.9 Whereas the cultural influence of Northern India mostly spread along land-bound trade routes, including those known as the Silk Road, the impact of Southern Indian culture is mostly noticeable in territories with which Southern Indian polities engaged in sea trade. Recent scholarship has recognized that maritime relations played an equally important role in the spread of Buddhist ideas and artistic styles.10 Perhaps one of the most important maritime Buddhist imports to Tang China was the 9 The most recent and comprehensive research on travelling monks as agents of religious and artistic transmission from India to China is in Wong 2018. On Esoteric Buddhism in Dunhuang and depictions of Esoteric deities in Dunhuang art during the Tang, see Schmid 2011. 10 On the spread of Buddhism through maritime networks, see Acri 2016, 2018. 51 Vajroṣṇīṣa Yoga ( 金 刚 顶 瑜 伽 Jin gang ding yu qie; lit. the Yoga of Adamantine Crown)11 teachings from Southern India. This tantric system includes eighteen structural texts, out of which only the Sarvatathāgatatattvasaṅgraha Tantra (STTS),12 also known as the Vajraśekhara Sūtra, was transmitted to China during the Tang dynasty. The tradition based on the STTS (hereafter the Yoga tradition) became the foundation for a Buddhism of national protection in China in the second half of the 8th century.13 The propagation of the Yoga tradition in China is attributed to the activities of the Indian monk Vajrabodhi (or Vajrabuddhi)14 and his pupil Amoghavajra.15 Vajrabodhi studied the Vajroṣṇīṣa Yoga under Nāgajñāna (Nāgabodhi or Nāgabuddhi) in Southern India, most likely at Kāñcīpuram, 16 the capital of the Pallava Kingdom and a great center of religious and Sanskrit learning equal almost to Nālandā.17 Nāgajñāna in turn was a disciple of Nāgārjuna, who is also believed to be from Kāñcīpuram,18 and who according to legend obtained the Vajroṣṇīṣa Yoga teachings from an “iron stūpa”, which was probably located in Tamil Nadu or Andra Pradesh.19 Vajrabodhi was famed as a great Esoteric master and performer of miraculous deeds, and before traveling to China he was active at the Pallava court in Kāñcīpuram. 20 The core territory of the Pallava Kingdom roughly corresponds to the modern-day state of Tamil Nadu. The Pallavas were a major military and cultural power in the Peninsula and Tamil merchants historically played a key role in maritime commerce with Southeast Asia and China.21 In the late 7th and early 8th centuries, the Pallavas' relations with China in particular 11 As was shown by Rolf Giebel, Vajroṣṇīṣa is a correct Sanskrit reconstruction of the Chinese title 金刚顶瑜伽 , which is traditionally known as Vajraśekhara, see Giebel 1995: 109. The Vajroṣṇīṣa Yoga is understood as "the Esoteric tradition that encompasses the doctrines and practices contained in the eighteen texts of the Vajroṣṇīṣa Yoga cycle and the lineage concerning the legitimate transmission of them." See Yang 2018: 10. 12 On the STTS and its English translation, see Giebel 2001. 13 Yang 2018: 120-158. On the notion of National Protection Buddhism in Tang China, see Orzech 1998: 1-9; 135-209. 14 On Vajrabodhi, see Sundberg and Giebel 2011. 15 On Amoghavajra and his teachings, see Chou 1945: 284-307; Goble 2012; Orzech, Sørensen and Payne 2011: 263- 285; Yang 2018. 16 Sundberg and Giebel 2011: 151 17 On Kāñcīpuram, see Thapar 2003: 344. Lokesh Chandra even suggested Kāñcīpuram as the location of Uḍḍiyāna. Although this issue was recently solved by Sanderson, who has sufficiently proved that Uḍḍiyāna is located near Kashmir, Chandra's article provides a useful discussion on Kāñcīpuram as a centre of Esoteric Buddhism, see Chandra 1980. On the location of Uḍḍiyāna, see Sanderson 2007: 265. 18 Sundberg and Giebel 2011: 195, n.110. 19 The story of the "iron stūpa" is found in Amoghavajra's "Instructions on the Gate to the Teaching of the Secret Heart of the Great Yoga of the Scripture of the Diamond Summit." ( 金刚顶经大瑜伽秘密心地法门义诀 Jin gang ding jing da yu qie mi mi xin di fa men yi jue, T.1798.39:808a19-b28). See Orzech, Sørensen, and Payne 2011: 152. Amoghavajra's account is translated integrally in Orzech 1995. See also Tanaka 2018: 309-311. 20 Sundberg and Giebel 2011: 7-9. 21 On the history of maritime relations between these regions, see Guy 2014: 3-13. 52 strengthened and it is recorded that the two states exchanged several diplomatic missions.22 The Pallava king Narasiṁhavarman II (r. 700-728) is credited with building a Buddhist temple in the Chinese style at the port city of Nāgapaṭṭinam for Chinese Buddhists arriving by sea.23 On the Southern Chinese shore, three temples (Hindu or Buddhist) for Tamil merchants were erected in the port city of Guangzhou ( 广 州 ) in the 8th century or even earlier.24 Xuanzang ( 玄奘 , c. 602-664) visited Kāñcīpuram by land routes in the mid 7th century and reported the presence of about 100 Buddhist temples and 10,000 Buddhist monks.25 Intent on propagating Buddhism in China, Vajrabodhi joined a Pallava diplomatic mission and travelled to the Tang court as an envoy of the Pallava king. Navigating along established sea routes through Sri Lanka and Sumatra, the mission arrived at Guangzhou, from whence they reached Luoyang ( 洛阳 ) in the year 720. Vajrabodhi brought to China various Buddhists texts, including a short version of the Vajroṣṇīṣa Yoga scripture (a complete one hundred thousand verse text of the scripture was purportedly lost on the way during a storm) and a copy of the Mahāprajñāpāramitā Sūtra offered to the Chinese emperor by the Pallava king Narasiṁhavarman II.26 The master spent the remaining 21 years of his life in China propagating the Yoga tradition and working on translation projects.27 It is especially relevant for this study that Vajrabodhi was known in China as a great painter, who was particularly good at painting Indian Buddhist images. He was included in the Record of Famous Painters of Successive Dynasties ( 历代名画记 Li dai ming hua ji; 9.16b-17a), compiled by Zhang Yanyuan ( 张彦 远 ) in 847.28 The master himself painted numerous maṇḍalas as well as designed wooden sculptures for the Luoyang Guangfu Temple and a stūpa of Vairocana, which was considered as an excellent piece of work. None of Vajrabodhi's works are known to survive, but it could be surmised that they were most likely executed in the Pallava style. Vajrabodhi's disciple Amoghavajra, whom he met during his journey to China, is regarded as a towering figure in the history of Chinese Buddhism. Amoghavajra studied under Vajrabodhi for more than 20 years and became a prolific translator, second only to Xuanzang, and an imperial preceptor for three successive Tang emperors: Xuanzong ( 玄宗 , r. 712-756), Suzong ( 肃 宗 , r. 756-762), and Daizong ( 代 宗 , r. 762-779).29 In 742-746, Amoghavajra travelled to Sri Lanka and Southern India, where he received Esoteric initiations and acquired more than five hundred Buddhist texts, including a complete scripture of the eighteen texts of 22 Sen 2016: 26; Sundberg and Giebel 2011: 144-145. 23 Sundberg and Giebel 2011: 145. 24 Guy 1993: 295. 25 Sundberg and Giebel 2011: 154. 26 Sundberg and Giebel 2011: 138, 147. 27 Yang 2018: 27-28. 28 Chou 1945: 276 n. 30, 280; Sundberg and Giebel 2011: 4 n.10. 29 Goble 2012: 62. 53 the Vajroṣṇīṣa Yoga.30 Amoghavajra made his first significant translations of the texts obtained during his journey, including the first part of the STTS, in the Hexi Corridor, in the regional capital city of Liangzhou ( 凉州 , modern Wuwei) between 754 and 756. He was sent to Hexi upon request of the military governor Geshu Han ( 哥舒翰 , 700-757) to assist the troops with ritual support in the battlefield against the Tibetans who constantly harassed the northwestern frontier.31 In Hexi, Amoghavajra initiated thousands of people into the secret practices of the Yoga and gained many followers amongst the local nobility, government, and military. The Yoga teachings were propagated in Hexi on an unprecedented scale. This can be illustrated by an excerpt from Amoghavajra's official biography, composed upon his death in 774, which describes the events in 754: In the thirteenth year [of Tianbao], [Bukong] arrived in Wuwei, where he resided at the Kaiyuan Monastery. From the military commissioner to officials of the lowest rank, [Bukong] conducted the service of abhiṣeka for all; as many as several thousand people, both the gentry and commoners, ascended the altar [to receive the service of abhiṣeka]. [Bukong] imparted the teachings of the Five Divisions to his clerical disciple, Hanguang. Next, [he] administered the abhiṣeka ritual of the Five Divisions for Li Yuancong, now the Commissioner of Merit, [with the prestigious rank of] Opening Bureau, and instructed him in the teachings of the Great Maṇḍala of the Diamond Realm.32 It is important to note that Amoghavajra gained his renown as an Esoteric master in Hexi, and from there it spread to the capitals via Hexi officials and military commanders who were his disciples.33 In addition, Hexi came to be regarded as the master's ancestral place; towards the end of his life, he was awarded the title of Duke of Su and given a nominal fief of 3000 households in Suzhou ( 肃州 ), a district in the northwestern part of the Hexi Corridor.34 During the turbulent events of the An Lushan rebellion and the war with the Tibetans, Amoghavajra significantly strengthened his religious and political influence at court. He performed countless Esoteric initiations and protective rituals for the imperial troops, the 30 Yang 2018: 33-35, 256-257. It is not known what regions Amoghavajra visited while in India; the sources are extremely vague about the Indian part of his trip and give detailed information only on his stay in Sri Lanka. Bearing in mind the shortness of the trip and that Amoghavajra spent most of his time in Sri Lanka, it is unlikely that he travelled beyond Southern India. 31 Yang 2018: 41-42, 264-265. 32 Yang 2018: 265. "The teachings of the five divisions" refers here to the Yoga tradition, while "the Great Maṇḍala of the Diamond Realm" stands for the Vajradhātu maṇḍala. See Yang 2018: 30. 33 Yang 2018: 41-45. 34 Yang 2018: 170-171, 362. 54 state, and the emperor. The institutionalization of a close group of his disciples as a form of governmental organization for the purpose of serving the state by performing apotropaic rituals can be regarded as one of Amoghavajra's major achievements.35 At the same time and as Robert Sharf has argued, the Esoteric practices propagated by Vajrabodhi, Amoghavajra and alike should not be regarded as an independent tradition/sect/school. What these masters promoted was essentially a new ritual technology intended for members of the ruling elite who could afford spending considerable resources on elaborate rituals conducted for the simultaneous achievement of mundane and supermundane goals. Sharf has emphasized that in China in the 8th century these new teachings were transmitted to monastics from a variety of backgrounds and beyond sectarian lines imposed by later traditions. 36 Amoghavajra and his Buddhist institution reached the peak of their power during the reign of Daizong (762-779). The master had a vast network of supporters and disciples, including the most powerful people of the empire; all the court officials and commanders of the imperial guards were required to be initiated into the Vajradhātu maṇḍala and to participate in rituals orchestrated by Amoghavajra.37 Among the master's other deeds are the translation of a great number of Buddhist texts and the promotion of the cult of Mañjuśrī and Mount Wutai.38 Liangzhou, where Amoghavajra translated the STTS and initiated thousands of people into secret practices of the Yoga tradition in 754-756, fell to the Tibetans in 758 and became a major centre of the Tibetan administration over the eastern areas.39 It is worth remembering that Tibetan Buddhism and the Tibetan Buddhist order were at this time in their early formative stage. Only in the 760s Buddhism started to be promoted by the Tibetan emperor Trisong Detsen (Tib. Khri srong lde btsan, r. ca. 742-797); the first Tibetan monastery of Samye (Bsam yas) was consecrated in 779.40 In general, the Tibetan ruling elite was highly suspicious towards Esoteric teachings, however the STTS and closely related texts (in general correspond to Yoga Tantras according to the Tibetan classification) were not only permitted, but, in parallel with the contemporaneous developments in Chinese state Buddhism, became a foundations for official Buddhist ceremonies in the late 8th century (this will be discussed in detail in Part 4).41 It is also clear that Tibetan interest in the Yoga tradition emerged in close temporal proximity to Amoghavajra's propagation of this tantric system in the Hexi region, which soon after became a part of the Tibetan Empire. 35 Yang 2018: 194-195. 36 Sharf 2017: 85-86, 114. 37 Yang 2018:188-200. 38 Yang 2018: 147-160. 39 van Schaik and Galambos 2012: 46. 40 Kapstein 2000: 60. 41 Weinberger 2010: 140-150. 55  Amoghavajra died in 774, two years before the Tibetan conquest of Guazhou ( 瓜州 , the prefecture containing Yulin caves). With the conquest of Shazhou ( 沙 州 , the prefecture containing Dunhuang) in 786, the whole Hexi Corridor came under Tibetan rule. Tibetans continued the patronage of existing Buddhist institutions in Hexi, which were numerous and many were large, and sponsored building of several new monasteries and temples.42 However, Tibetan impact on Buddhist thought and Buddhist art in Hexi during this period of occupation remains largely unclear, since major developments with regard to the Tibetan appropriation of Buddhism took place during or after the conquest of the Hexi corridor. One undisputable sign of the Tibetan input on Dunhuang art, however, is depictions of Tibetans and a Tibetan emperor in several caves at Mogao and Yulin.43 The Chinese emperor continued to be depicted as well and his superior status was not diminished. For instance, in a painting in Mogao Cave 231 dated to circa 839, which shows the debate between Mañjuśrī and Vimalakīrti, a Chinese emperor retains his exalted position as the Buddhist authority and is shown in the entourage of Mañjuśrī, while a Tibetan emperor leads a procession of foreign kings on the side of Vimalakīrti.44 The Dunhuang manuscripts mention some mixed Sino-Tibetan Buddhist communities and show no signs of opposition between Chinese and Tibetan Buddhists.45 Given the richness of Buddhist culture in Hexi and written records describing Tibetans eagerly seeking to learn the Dharma from the Chinese, the impact of Hexi Buddhist culture on what was to become known as Tibetan Buddhism and Tibetan Buddhist art must have been significant.46 The Tibetan period in Hexi lasted until the mid 9th century, when, following the collapse of the Tibetan Empire, most of the region was again under Tang control. Despite this political upheaval, many Tibetans remained in the region and the Tibetan language continued to be widely used. 47 To summarize, when Tibetans arrived in Hexi, they found themselves in a highly sophisticated Buddhist environment, where Esoteric teachings, especially those of the Yoga tradition, enjoyed great prestige. The Buddhist art in Hexi was characterized by an increasing popularity of Indianized styles and Esoteric imagery. 42 Sha 2013: 152-156; van Schaik and Galambos 2012: 62. On Buddhist institutions in Hexi during the Tibetan period, see Horlemann 2012. 43 Just a few examples: Yulin Cave 25 and Mogao Caves 158, 231, 359. See Sha 2013: 53-81. 44 For a discussion on the symbolism of representations of the Mañjuśrī-Vimalakīrti debate in Dunhuang, see Schmid 2006: 183-185. For a reproduction of the scene from Cave 231, see Debreczeny 2019: 50, fig.1.2. 45 H. Sørensen 2000: 41. 46 For example, according to the Testament of Ba, two Tibetan delegations were sent to the Chinese capital to learn Buddhist teachings in the second half of the 8th century. It is also stated that one thousand Buddhist texts were granted by the Chinese emperor and brought to Tibet by one of these delegations. See Wangdu and Diemberger 2000: 16-17, 21, 42 n.91, 49 n. 121-122, 52 n. 136, 138, and 70 n. 245. Moreover, this source shows that Buddhism in 8th century Tibet was strongly associated with China and sometimes even referred to as the doctrine of China. See Wangdu and Diemberger 2000: 37 n.71. 47 van Schaik and Galambos 2012: 62-69. 56 Part 2. Definitions and Origins A list of paintings executed in the artistic mode discussed in this article is presented in Appendix 1. It is by no means exhaustive and includes only the most significant published examples. The majority of the paintings depict maṇḍala-like compositions. Common central figures include Vairocanābhisaṁbodhi (see section 2.2),48 Vajrasattva, a Buddha, four-armed Amoghapāśa (see section 2.3), and two-armed Avalokiteśvara. As will be shown these images mostly draw from repertoire of the Yoga tradition. Other subjects include Vairocanābhisaṁbodhi with the eight great bodhisattvas (see section 2.1) as well as Kubera with attendants (see section 2.4). In addition, the list includes portable paintings with single figures of bodhisattvas (see section 2.5) and paintings with stylistically mixed imagery (see section 2.6). Before proceeding to a detailed discussion on the paintings under review, it is necessary to mention that for almost all of them potential textual sources can be found in both the Chinese and Tibetan traditions. Since it has been historically maintained that Tibetan Esoteric Buddhism stems directly from the Indian tradition, possible relationships between the parallel Chinese and Tibetan texts have been rarely seriously considered. Moreover, it is commonly believed that developed Esoteric Buddhism did not exist in Dunhuang/Hexi before the Tibetan conquest. Thus, a decisive factor for the attribution of the discussed images to the Tibetan tradition has often been a combination of their Indianized style (as opposed to the Chinese style), which was understood as Tibetan, with Esoteric subject matter. In the following, I will challenge the validity of this approach. 2.1 Representations of Vairocanābhisaṁbodhi with the Eight Great Bodhisattvas in Hexi and their spread to Eastern Tibet 2.1.1 Yulin Cave 25 2.1.1.1 Description of the Cave and the State of Scholarship The scholarship on Yulin Cave 25 perfectly illustrates issues surrounding the attribution of the imagery under review to Tibetan influence and therefore I discuss it first and at greatest length. The cave was excavated during the period of Tibetan rule in Guazhou and is currently dated by 48 Vairocanābhisaṁbodhi refers here to bejewelled Vairocana displaying dhyāna mudrā, as this form was initially described in the Vairocanābhisaṁbodhi Tantra (VAT). See Wayman and Tajima 1998: 8. However, images of Vairocana in this form do not necessarily signify the context of the VAT. 57 Fig. 1 Vairocana and the eight great bodhisattvas, the east wall of the main chamber, late 8th C., Tibetan period, Yulin Cave 25, China (extant section (in color): Dunhuang Academy 2014: pic. 10; the lost section (in b/w): Kapstein 2009: fig. 9). the Dunhuang Academy to the early years of the Tibetan period (i.e. 776-781).49 The centre of the cave is occupied by a sculptural representation of Vairocana perhaps dating to the Qing dynasty, which could be a replica of the original sculpture. The wall paintings are regarded as originals. The central wall of the main chamber consists of Indianized images of Vairocanābhisaṁbodhi and the eight great bodhisattvas (Fig. 1). 50 The bodhisattvas are identified by Chinese inscriptions as Ākāśagarbha, Kṣitigarbha, Maitreya, Mañjuśrī, Vajrapāṇi, Avalokiteśvara, Sarvanivāraṇaviṣkambhin, and Samantabhadra, who together are known as the eight great bodhisattvas.51 This group of nine — Vairocanābhisaṁbodhi and the eight great bodhisattvas — will henceforth be abbreviated as VAEGB. The rest of the paintings in the cave are executed in the Chinese mode. They include the Pure Lands of Maitreya and Amitābha on the northern and southern walls respectively; the bodhisattvas Mañjuśrī and Samantabhadra with attendants and Sudhana on the western wall (Fig. 2); and six standing bodhisattvas, whose identities are not 49 A review of scholarship on the attribution of the cave to the Tibetan period can be found in Kapstein 2009: 56-57. The date period 776-781 is provided by the Dunhuang Academy in https://www.edunhuang.com/ cave/10.0001/0001.0002.0025 (accessed 22.04.2019). 50 Only four bodhisattvas on the left side have survived to the present day. The group on the right side is shown in the old photographs taken in 1940s that are currently preserved in the Lo Archive, Princeton. 51 Wang 2018: 108-109. The most comprehensive study on the eight great bodhisattvas as a group is in Bautze- Picron 1997. 58 Fig. 2 Bodhisattvas Samantabhadra (left) and Mañjuśrī (right) with Sudhana and retinue, late 8th C., Tibetan period, the west wall of the main chamber, Yulin Cave 25, China (Dunhuang Academy 2014: pics.32, 33). Fig. 3 Left: Face of Vairocana, detail of fig. 1; Right: Face of attending bodhisattva from Samantabhadra retinue, detail of fig. 2 ( https://www.e-dunhuang.com/cave/10.0001/0001.0002.0025). 59 entirely clear, and who are depicted on the corners of the cave. These Indianized and Chinese images show identical line structures and shading, and the faces of some figures between the two groups bear a striking resemblance (Fig. 3). From this it is clear that, differences in style notwithstanding, all the paintings in the cave were likely executed by the same group of artists. A publication produced in 2000 by the Dunhuang Academy for a centennial commemoration of the discovery of the library in Mogao cave 17 briefly mentions that the central composition (VAEGB) of Yulin cave 25 "indicates close ties to Master Amoghavajra's spread of Tantrism in the Hexi region."52 This idea has been ignored in later scholarship. Instead, the coexistence of the two artistic modes, as well as the coexistence of Esoteric and traditional Pure Land subjects within the cave have been ascribed to the cultural influence of Tibet. The style of the composition of VAEGB has been understood either as the Tibetan style or as the Pāla style of northeastern India being transmitted through Tibet.53 In the following, I 52 Unfortunately no arguments supporting this attribution were given. See Zhang 2000: 31. 53 The publications include Kapstein 2004, 2009, 2014; Sha 2013; Wang 2018; Debreczeny 2019. In a series of articles, Matthew Kapstein has developed a theory that Yulin Cave 25 should be identified as a temple built by Tibetans to commemorate a peace treaty between the Tibetan and Tang empires made in 821-822. Although the iconographic program of Yulin Cave 25 only partially matches a description of the temple found in the Dunhuang manuscripts, the author develops his idea by pointing to the popularity of Vairocana cult in the Tibetan Empire and to existence of roughly contemporaneous rock carvings of VAEGB in Eastern Tibet (Kapstein 2009: 53, 58). As stylistic evidence, the figure of Vairocana from Cave 25 was compared with an undated sculptural representation of the same deity that was ascribed to the Imperial period. The sculpture, however, is executed in a considerably different style, which makes it difficult to accept the author's point. Summarizing his argument, Kapstein emphasizes the coexistence of the two styles in the cave and concludes: "In both iconography and style, Anxi Yulin 25 thus expresses the coexistence of the Tibetan and Chinese Buddhist worlds. On the surface at least, it seems an especially fitting memorial to a peace accord between the two powers." (Kapstein 2009: 54). Recently, the author revised his initial notion and suggested instead that Yulin Cave 25 could be either an earlier Tibetan model for the treaty temple or its copy (Kapstein 2014). I will discuss issues concerning the Tibetan rock carvings in section 2.1.2.3. As for the coexistence of Indianized and Chinese depictions, as well as Esoteric and Pure Land subjects, such mixed imagery can be already found in Tang dynasty caves predating the Tibetan period. I will discuss some examples in the next chapter.    Sha Wutian's research on Cave 25 was published in his monograph devoted to Dunhuang art of the Tibetan period. The author postulates that the figures of VAEGB in Cave 25 represent the Tibetan style and more specifically the northeastern Indian Pāla style as transmitted through Tibet. These statements seem to ignore the fact that no comparable depictions that can be dated to the period before the Tibetan occupation of Hexi are known from Tibet. Moreover, the discussed material certainly does not belong to the Pāla artistic tradition. See Sha 2013: 360-382, 470- 492. Sha 2016: 119-122.    Michelle Wang in her recent publication Maṇḍalas in the Making came to the conclusion that the depiction of VAEGB in Cave 25 belongs to the Tibetan Buddhist tradition and is executed in the Tibetan style, which she defines as a "Nepalese-derived Tibetan style." The author compared the images from the cave to Tibetan and Nepalese sculpture and noted various similarities. Significant differences were, however, overlooked. The Eastern Tibetan rock carvings and the coexistence of the two artistic modes in Cave 25, named as "stylistic bilingualism," were regarded as evidence for the Tibetan origin of the cult of VAEGB in Hexi. See Wang 2018: 91-121.    In the latest contribution to this subject, Karl Debreczeny sees the group of VAEGB in Yulin Cave 25 as a manifestation of the new aesthetic and religious interests introduced under Tibetan rule. The author's argument is essentially based on the works by Kapstein, Sha and M. Wang and therefore has the same shortcomings. See Debreczeny 2019: 24-25. 60 Fig. 4 Two bodhisattvas (detail of the wall painting), Early Tang dynasty (618-712), the east wall of the main chamber, Mogao Cave 329, China (https://www.e-dunhuang.com/ cave/10.0001/0001.0001.0329). will argue against the current scholarly consensus and demonstrate that neither such thematic and stylistic diversity nor the style of the group of VAEGB can be ascribed to the Tibetan or Pāla-Indian influence. 2.1.1.2 Stylistic Considerations Indian and Chinese Modes in Hexi Art before the Tibetan Conquest Indianized imagery started to appear early on in Tang dynasty caves.54 For example, Fig. 4 shows two bodhisattvas from Mogao Cave 329 (datable to the early Tang period, 618-712).55 A bodhisattva on the right is depicted in the Chinese mode, characterised by a round halo, long scarves wrapped around the body, and a voluptuous lower garment. The bodhisattva on the left, on the other hand, is shown in the Indian mode. It is similar to what is found in Yulin Cave 54 For example, Mogao Caves 202, 321, 329, and 335. See Sha 2013: 133-134. 55 For a description of this cave and additional images, see https://www.e-dunhuang.com/cave/10.0001/0001.0001.0329 (accessed 22.04.2019). 61 Fig. 5 Bodhisattva Bhaiṣajyarāja with eight bodhisattvas and two lokapalas, 766-776, Tang dynasty, slopped ceiling inside the niche on the north wall of the main chamber, Mogao Cave 148, China (Dunhuang Academy 2013: fig.29). 25 (Fig.1), where the bodhisattvas also have inverted U-shaped halos and wear similar lower garments resembling trousers. Another example of such coexistence of the two modes can be found in Mogao Cave 148, which contains depictions of the Buddha's Parinirvāṇa and various Pure lands painted in the Chinese mode and Esoteric imagery (e.g. various multi-armed forms of Avalokiteśvara) painted in the Indian mode. According to the inscription engraved on the stele at the entrance, the cave was completed sometime between 766 and 776 — that is within the decade leading up to the Tibetan conquest of Guazhou in 776.56 Particularly interesting are six rectangular panels painted inside the niches. They contain compositions akin to that of VAEGB in Yulin Cave 25 (Fig. 1). Fig. 5 is one of the panels, and it depicts the bodhisattva Bhaiṣajyarāja surrounded by eight bodhisattvas and two Lokapālas. It has a similar layout to the image in Yulin Cave 25 (Fig.1), with a large central figure placed under a canopy and flanked by two groups of smaller figures shown symmetrically on either side. The lotus seats of the central figure and the two chief bodhisattvas are almost identical to the lotus seats in the Yulin composition. In both paintings, the stems of the lotus seats of the bodhisattvas grow from a seat of the central divinity. The small lotus flowers and leaves in the intermediate space are similar in both compositions as well. It is thus evident that the coexistence of the Indian and Chinese artistic modes as well as of Pure Land and Esoteric imagery within Yulin Cave 25 is a continuation of the trend that started in early Tang dynasty caves. Similarly, the depiction of the group of VAEGB itself is based on a template that was popular in Hexi before the Tibetan conquest. 56 Mogao Cave 148 is discussed in Lee 2010: 181-201. Esoteric imagery in the cave is discussed in Liu 2008. 62 Southern Indian Artistic Mode Esoteric teachings were transmitted not only through texts, but also through their accompanying images.57 Although none of the original Indian illustrations survive in China, the fact that they existed is evident from Chinese translations made by Indians. For example, Śubhākarasiṁha ( 善 无 畏 , 637-735), an Indian translator-monk living in Tang China, frequently made references to iconographic illustrations in his texts.58 In addition, the importance of imagery in the transmission of Esoteric teachings was stressed by Kūkai ( 空 海 , 774-835), who studied in China under Amoghavajra's successor Huiguo ( 惠果 , 746-806). In his inventory (dated 806), Kūkai states: "The secret wisdom of the true words is hidden and implicit in scriptural texts; only through imagery can it be properly transmitted."59 In China, original Indian illustrations were copied by local artists, and, as in the case of Vajrabodhi, many Indian monks themselves were able to produce such images. In any case, works with Indian background were considered extremely valuable; they were carefully recorded and praised in monastic inventories. Furthermore, accuracy in copying of these images was regarded crucial for the authenticity and efficacy of replicas in Esoteric rituals.60 Another important feature of Chinese practice was the standardization of religious imagery. Standardized models and templates for replication were known as yang ( 样 ). In her recent book, Hsueh-man Shen ( 沈 雪 曼 ) describes many Tang dynasty scrolls with yang of iconic images of various deities, including a group of the eight great bodhisattvas.61 It is through these processes of copying, standardization, and replication that elements of Indian costumes and decorations were adopted in Chinese art. In this regard, tracing the origins of distinctive decorative elements in early Chinese Esoteric depictions yields an important trace for determining the provenance of the artistic tradition, to which Indian prototypes belonged, and, by extension, the provenance of the corresponding teachings. The group of VAEGB in Cave 25 was likely based on standardized models. The figures of the bodhisattvas are essentially of the same type. They are shown either frontally or in a three-quarter profile and have almost identical decorations. Diversity is achieved through different hand gestures, leg positions, and attributes. Although the figures are overall similar to the imagery in Mogao Cave 148 (Fig. 5), they show certain stylistic peculiarities (see an enlarged image of Vairocana in Fig. 6) such as: 57 This was recently discussed by Kimiaki Tanaka, see Tanaka 2018: 125-126. 58 Tanaka 2018: 125-126. 59 Shen 2019: 154. 60 Hsueh-man Shen shows that the inventories of objects taken from China to Japan during the Tang specifically comment on images produced by Indian masters or originated from India, see Shen 2019: 154. The earliest extant example of iconographical drawings is an early 9th century Chinese scroll, known as the Gobu shinkan ( 五部心观 ), see Tanaka 2018: 293-294. 61 Shen 2019: 149-155. 63 Fig. 6 Vairocana and close-ups of the crown and armband, details of fig. 1 (https://www.e-dunhuang.com/ cave/10.0001/0001.0002.0025). - elongated Y-shaped torsos; - crowns and armbands with spikes of pearls or little spears (henceforth spiked crowns and armbands), see close-ups in Fig. 6; - broad necklaces with swags of pendants (henceforth swagged necklaces); - and earrings with large pendants. These elements are the defining characteristics of the artistic mode that this article is concerned with, though it is not always the case that all of them are present in a given image. Of the four characteristics, the spiked decorations are the most distinctive and persistent feature (see close-ups in Fig. 6). It is a matter of fact that such decorations are almost exclusively associated with Southern Indian art (see examples in Figs. 7, 8, 9, 10, 19, 23, 24). In this regard, 64 they are neither found in Tibetan art prior to the occupation of Hexi, nor in the art of Northern India, Nepal, or Central Asia, artistic traditions which, along with that of China, contributed to the formation of Buddhist art in Tibet. As such, this indeed eliminates the currently assumed possibility of the alleged Tibetan background of the discussed imagery. In contrast, this Southern Indian stylistic trace corresponds well to the Southern Indian current in Chinese Esoteric Buddhism represented by Vajrabodhi and Amoghavajra. In the 7th and 8th centuries, there were two regions in Southern India where Esoteric Buddhism was prominent and these exact kinds of spiked jewelry were common in artistic representations. They are Tamil Nadu in the southeastern part of the Indian Peninsula and the Western Deccan. With regard to Sri Lanka, which was an important destination for both Vajrabodhi and Amoghavajra, it is unlikely that the discussed artistic mode in Hexi art has a Sri Lankan origin. Although Sri Lankan Mahāyāna and Vajrayāna Buddhism and associated art were greatly influenced from the Pallava Kingdom of Southern India,62 the spiked decorations are not particularly common in Sri Lankan Buddhist depictions. 63 Connections with the art of Tamil Nadu are not surprising as this region was the home of Esoteric teachings propagated by Vajrabodhi and Amoghavajra.64 Moreover, Vajrabodhi as a painter most likely belonged to the Pallava artistic tradition flourished in Tamil Nadu. Having in mind the records about imagery produced by Vajrabodhi in China and the Chinese practices of copying, standardization, and replication, we can assume that a Chinese artistic tradition based on Pallava art and closely associated with Vajrabodhi emerged in the Tang capitals by the mid 8th century and spread from there to Hexi/Dunhuang. The existence of religious and artistic interexchange between Dunhuang and the Tang capitals is well recorded65 and thus Vajrabodhi's teachings and related imagery likely reached Hexi already during the master's lifetime. Nevertheless, the tradition undoubtedly owes its greater popularity in the region to Amoghavajra. As discussed, the master not only taught in Hexi, but also made his first major translations there. Regarding the Western Deccan, the existence of Esoteric Buddhism in the 8th century there is evident from Ellora cave sculpture. It has been also suggested that the region could have 62 On the Pallava Buddhist and artistic influence in Sri Lanka, see Indrapala 2005: 189-192; Dohanian 1983. 63 On Sri Lankan Mahāyāna imagery, see Dohanian 1977. The most comprehensive research on Sri Lankan Buddhist sculptures can be found in von Schroeder 1990. It has to be noted, however, that remaining Esoteric Buddhist material is only lately being studied and the chronology of Sri Lankan Buddhist art of that period is debatable because of a lack of firmly datable material. 64 On Esoteric Buddhism in Tamil Nadu, see Yang 2018: 179-180, n. 370. In particular, on Kāñcīpuram as a centre of Esoteric Buddhism, see footnote 17. Esoteric Buddhism in Tamil Nadu was probably part of the larger tradition encompassing the neighboring Andhra region, the motherland of the Pallavas. On the development of Esoteric Buddhism in Andhra, see Barber 2008. 65 Fraser 2004: 52-53. 65 been the origin place of such fundamental scripture as the VAT.66 At the same time, numerous cross-regional military raids of the Pallavas (ruling in Tamil Nadu), and the Cālukyas (ruling in the Deccan) in the 7th and 8th centuries intensified cultural exchange across the peninsula and it is not unlikely that the respective Buddhist communities were in conversation, influencing each other and sharing major scriptures.67 For example, the earliest surviving images displaying iconography that could be related to the STTS can be found in the Western Deccan, at Ellora Cave 12 (early 8th century).68 In general, Ellora sculptures from the 8th century onwards reflect increasing influence from the South of the Peninsula, and in particular the Pallava influence.69 Regrettably, we do not know the specific regions Amoghavajra visited while in India; the sources are extremely vague about the Indian part of his trip and give detailed information only on his stay in Sri Lanka.70 It is fair to assume, however, that Tamil Nadu, a place where Vajrabodhi studied the Yoga tradition before going to China, would have been Amoghavalra's prime destination, though a small possibility that he visited other regions, including the Western Deccan, cannot be entirely excluded. In addition, it is possible that materials collected by Amoghavajra in Sri Lanka were of an Indian origin, as the introduction of Esoteric Buddhism to Sri Lanka is associated with Tamil Buddhists migrated to the island in the 7th and 8th centuries fleeing from the antagonistic movements in Śaivism and Vaiṣṇavism that were specifically directed against Buddhists and Jains.71 As a result of the anti-Buddhist campaigns, only a handful of Pallava Buddhist images survived. Most of them have been vandalised and were found in the vicinity of Hindu temples which suggests a violent expropriation and conversion of Buddhist establishments.72 For our purposes, the Pallava stylistic features can be reconstructed from abundant examples of Brahmanic imagery. For instance, Pallava sculptures in Figs. 7 and 8 possess all major characteristic of the discussed artistic mode in Hexi art and show remarkable similarities to the aesthetics of the Yulin images (Figs. 1, 6): Y-shaped torsos, tall headdresses, spiked crowns and armbands, swagged necklaces, and large earrings. In the Western Deccan, Buddhist material heritage largely survived. Ellora Caves 11 and 12, which are traditionally dated to the early 8th century, represent the last large-scale Buddhist 66 On the development of Esoteric imagery in the Western Deccan, see Huntington 1981, Malandra 1993, Bautze-Picron 2000. On possible geographical origins of the VAT, see Wayman and Tajima 1998: 11-12. 67 On artistic and architectural connections between these regions see D. Dayalan. 68 Tanaka 2018: 265, fig. 4.9. 69 Malandra 1993: 116-117. The most striking example of the Pallava influence is a famous Kailāsanātha monument, which was built in the mid-8th century and partially patterned on Kailāsanātha temple in Kāñcīpuram. 70 Yang 2018: 256-257. 71 Indrapala 2005: 181, 192-193; Yang 2018: 180-181; Davidson 2002: 192. 72 Rao 1915; Verardi 2018: 296ff, fig. 4. 66 Fig. 7 Śiva as Gaṅgādhara (detail), stone relief, 7 th C., Fig. 8 Viṣṇu, copper alloy, 8 th -9 th C., Pallava period, Lalitankura cave, Tiruchirapalli, Pallava period, 47 cm, Thyagaraja Tamil Nadu, India (photograph courtesy of temple, Thiruvarur, Tamil Nadu, Emmanuel Francis). India (Asha Rani 1988: 147, fig. 7). undertaking in the region.73 The eight great bodhisattvas as a group assume prominent positions in both caves. While Cave 11 contains two groups of the standing bodhisattvas, Cave 12 has one group of the seated and two groups of the standing bodhisattvas. Fig. 9, from the first floor of Cave 12, shows two groups of the four bodhisattvas flanking a statue of the Buddha with dharmacakra mudrā. Their identities were reconstructed by Malandra and Bautze-Picron according to their attributes and they appear to be the same bodhisattvas as the bodhisattvas in Yulin Cave 25.74 The Ellora and Yulin bodhisattvas are visually similar as well and convey 73 For a detailed analysis of Buddhist art at Ellora in general and these caves in particular, see Malandra 1993. 74 Malandra 1996: 197; Bautze-Picron 1997:10-11. 67 Fig.9 Eight Bodhisattvas, first floor, early 8th C., Ellora Cave 12, Maharastra, India (www.elloracaves.org, photographs courtesy of Deepanjana and Arno Klein). Fig. 10 Left: Detail of fig. 9; Right: Detail of fig. 1 (https://www.e-dunhuang.com/cave/10.0001/0001.0002.0025). late Pallava aesthetics (see comparison in Fig. 10). They have elongated and Y-shaped torsos, crowns with spikes, and swagged necklaces. The armbands display similar leaf-like elements decorated with spikes. Spikes in the armbands of the sculptures are not clearly visible in the 68 pictures as the decorations were finished with plaster which is largely lost now. It is evident that stylistically the images in Yulin Cave 25 depended on Southern Indian prototypes and that the defining characteristics of the discussed artistic mode have a Southern Indian origin. Therefore, from here onwards this mode in Hexi art will be labelled as Southern Indian. 2.1.1.3 Iconographical Considerations Ellora Caves 11 and 12 and their Iconographic Innovations According to Geri Malandra, Ellora caves 11 and 12 reflect Esoteric teachings which were new for the Deccan, and this new iconography was executed in a new style that shows a noticeable influence from the South of the Peninsula.75 She suggested that the teachings reflected in the caves were imported to the Deccan and that they might have had the same origin as the sculptures' new style. In support for her theory, Malandra noted that contemporaneous or slightly later representations of the eight great bodhisattvas are found in Odisha and Java that points to the Southern (as opposed to Northern) source of the innovations.76 Indeed, Northern Indian depictions of the eight great bodhisattvas mostly date from the 9th century and later.77 In addition, the popularity of the cult of the eight great bodhisattvas in the 8th century is evident in the Malay Peninsular,78 as well as in China (discussed in this Part and in Part 4.). This distribution implies that the cult of the eight great bodhisattvas was popular in the eastern coastal areas of the Indian Peninsular that historically played a key role in maritime relations of Southern Indian polities with Southeast Asia, and China. Moreover, the Pallava influenced style of the Dunhuang depictions points specifically to Tamil Nadu as a source of the cult in China. Similarly, the Pallava influenced style of the Ellora sculptures should be regarded as evidence for a Tamil Nadu origin of the respective teachings in the Western Deccan as well. As was already discussed, travelling monks and pilgrims were instrumental in the transmission of new teachings and their art forms. Ellora was an important holy site (tīrtha), a place of pilgrimage that attracted people from distant lands and different religious communities; its sacred status encouraged commissioning religious imagery at the site. Lisa Owen has recently shown, for instance, that Ellora Jain caves reflect activities of the Jain communities associated with Tamil Nadu and Karnataka (see discussion in section 2.4).79 Although donors of the discussed 75 Malandra 1993: 116-117; Malandra 1996: 206. 76 Malandra 1996: 205-206. 77 Bautze-Picron 1997: 14-27. 78 Skilling 2011: 375-377; Sharrock and Bunker: 2016: 244; Acri 2018. 79 Owen 2012: 16, 20, 160-163. 69 Buddhist caves remain unknown, there is a possibility that they hailed from the South as well.80 It is also possible that some Buddhist communities migrated from Tamil Nadu to the Western Deccan during the period of interreligious antagonism in their homeland. The Aṣṭamaṇḍalakasūtra In addition to the sculptural groups of the eight bodhisattvas, Ellora Cave 12 contains five wall reliefs of maṇḍalas of the eight great bodhisattvas centred around a dhyāna-mudrā Buddha, identified by scholars as Vairocana (Fig. 11). The bodhisattvas in the maṇḍalas were identified according to their attributes and they also happen to be the same bodhisattvas as those depicted in Yulin Cave 25.81 According to Kimiaki Tanaka, the group of Vairocana and the eight great bodhisattvas ultimately originates from the Aṣṭamaṇḍalakasūtra (The Sūtra of the Mandala of the Eight Great Bodhisattvas), which was also the most probable source of the Ellora maṇḍalas.82 The central figure of the Tathāgata is only briefly described in the text83 and this perhaps explains an iconographic variety of his representations. He can be depicted unadorned, like in the Ellora maṇḍalas (Fig.11), or bejewelled, like in the Yulin composition (Fig.1), and with the bhūmisparśa-, dharmacakra- or dhyāna-mudrā. Examples of the group with the central figure displaying the dhyāna-mudrā are relatively rare in India and found only in the Southern part, namely in the Western Deccan (Ellora) and Odisha (Ratnagiri).84 Thus, it is evident that the iconography of the group depicted in Yulin Cave 25 was particularly common in Southern India. The Aṣṭamaṇḍalakasūtra was translated by Amoghavajra (T1167)85 and the central deity of the maṇḍala in his text is described as "Tathāgata in His True Golden Form Body" ( 如 来 真 金 色 身 ).86 This epithet perhaps refers to Vairocanābhisaṃbodhi, a form of 80 Buddhist monks from Tamil Nadu and Andhra were active even in Northern India. Bautze-Picron has recently discussed a number of stone and bronze images from Kurkihar dating from the 8th to 11th century which bear inscriptions mentioning donors who hailed from Kāñcīpuram and Andhra. Notable among these are two circa 9th century stone slabs which show a set of the eight great bodhisattvas around the Buddha. See Bautze-Picron 2015: fig. 91, 92. In a private communication on 09.08.2019, Bautze-Picron remarked that these slabs would seem to suggest that the cult of the eight great bodhisattvas might have been introduced to Bihar from Tamil Nadu and Andhra. 81 On the identification of the figures in these maṇḍalas, see Malandra 1996: 197, 201; Tanaka 2018: 141-144. 82 Tanaka 2018: 92-93. This sutra is also known as the Aṣṭamahābodhisattvamaṇḍalakasūtra. 83 For the description of the eight great bodhisattvas and the central figure in the text, see de Viser 1915: 15-17. 84 Northern Indian examples show Buddha/Vairocana either in the bhūmisparśa- or dharmacakra-mudrā. See Bautze- Picron 1997; Tanaka 2018: 81-82. 85 The sūtra was initially translated into Chinese by Puṇyodaya in the 7th century (T486), but it became particularly popular in the translation by Amoghavajra, who assigned to this maṇḍala a much greater importance, see Li-Kuang Lin 1935. There are two Tibetan versions of the sūtra, both are believed to be translated around 800, see Skilling 2011: 376, n. 9, 10, 11. 86 I am grateful to Geoffrey Goble for sharing this information. Private communication 25.07.2018. 70 Fig. 11 Maṇḍala of the eight great bodhisattvas, Ellora Cave 12, early 8th C., Maharastra, India (www.elloracaves.org, photograph courtesy of Deepanjana and Arno Klein). Vairocana, which is initially appeared in the VAT and described there as having a golden body (in contrast, for example, with Vajradhātu-Vairocana whose colour is white), bejewelled and displaying the dhyāna-mudrā.87 This is precisely how Vairocana is depicted in Yulin Cave 25. To summarise, in both style and iconography the Yulin group of VAEGB is closely related to Southern Indian examples. 87 For a description of Vairocana in the VAT, see Wayman and Tajima 1998: 101. The relation between the maṇḍala of the eight great bodhisattvas and the VAT was earlier suggested by Malandra and Tanaka. See Malandra 1996: 76; Tanaka 2018: 77-78. 71 The Bhadracarī The popularity of the eight great bodhisattvas in Southern Indian Buddhism is well reflected in Amoghavajra's textual heritage. In addition to the translation of the Aṣṭamaṇḍalakasūtra, Amoghavajra integrated the eight great bodhisattvas into practices of such popular prayers as the Bhadracarī and the Uṣṇīṣavijayā dhāraṇī.88 These combinations were proabably meant to employ the power of the boddhisattvas to magnify the efficiency of these prayers. According to the Aṣṭamaṇḍalakasūtra, the mantras and maṇḍala of the eight great bodhisattvas are effective in fulfillment of one's wishes, eradicating one's sins, setting one on the path to swift enlightenment and protection from harm.89 I believe that although the Aṣṭamaṇḍalakasūtra is the ultimate source for the VAEGB group, the depiction of VAEGB in Yulin Cave 25 is not a maṇḍala, as the bodhisattvas do not encircle the central figure, and is actually based on another Amoghavajra's work entitled Puxian pusa xing yuan zan ( 普贤菩萨行愿赞 ) (T297).90 The text was composed between 756 and 774 and consists of four parts, where the first part is a translation of the Samantabhadrācāryap- ranidhānarāja ( 普贤行愿品 ), also known as the Bhadracarī;91 the second part is a eulogy to the eight great bodhisattvas, which is a part of the Aṣṭamaṇḍalakasūtra (T1167); and the third and final parts are a dhāraṇī and a praise of this dhāraṇī. In short, the work includes the cult of the eight great bodhisattvas of the Aṣṭamaṇḍalakasūtra into the practice of the Bhadracarī, an immensely popular aspirational prayer in which one vows to worship the buddhas and become enlightened for the sake of all beings. The Bhadracarī is known as an independent work and as the final part of the Gaṇḍavyūha Sūtra ( 入法界品 ),92 which itself is a part of the Avataṁsaka Sūtra ( 华严经 ), a core scripture of the Avataṁsaka ( 华 严 Huayan) Buddhist tradition.93 According to modern scholarship, the Vairocana of Esoteric Buddhism developed from the Vairocana of the Avataṁsaka Sūtra, 88 See Amoghavajra's works T297 and T972. On Amoghavajra's ritual manual for the Uṣṇīṣavijayādhāraṇī sutra (T972), see Wang 2018: 47-50, Lin 2014: 139-43. Amoghavajra produced the manual in 764 and it became the most influential redaction in Tang China (several earlier redaction are known, including the one by Śubhākarasiṃha). The manual introduces a maṇḍala which is known in Japanese Buddhism as Sonshō Maṇḍala in the Amoghavajra tradition. The maṇḍala is similar to the maṇḍala of the eight great bodhisattvas of the Aṣṭamaṇḍalakasūtra, but, in addition, it includes two wrathful deities, Acala and Trailokyavijaya. See Tanaka 2018: 88, 95. 89 Granoff 1968/1969: 2; Geoffrey Goble, private communication 25.07.2018. 90 On this text, see Dessein 2003. 91 The Bhadracarī is discussed in Osto 2010. 92 The Bhadracarī is included in the translation of the Gaṇḍavyūha Sūtra made by Prajña 般 若 , which was finished in 798 and was based on the Sanskrit copy that had been presented to the Chinese emperor by the King of Odisha. The extant Tibetan translation of the Gaṇḍavyūha was made during the imperial period and also concludes with the Bhadracarī. See Osto 2010: 2-3. 93 See Cleary 1993 for a translation of the sūtra. 72 who is featured in the sūtra under the special name of Rocana ( 卢舍那 Locana). The Huayan tradition integrated various Esoteric teachings, and the Huayan Vairocana and Vairocana of Esoteric Buddhism were sometimes "fused iconographically."94 The image of Vairocana in Yulin Cave 25 is an example of such fusion. A cartouche next to it contains the Chinese inscription 清净法身卢那舍佛 (the Pure Dharmakāya Locana Buddha) which refers to Vairocana of the Avataṁsaka Sūtra.95 Thus, we have Locana Buddha accompanied by the eight great bodhisattvas of the Aṣṭamaṇḍalakasūtra. This combination corresponds to the Puxian pusa xing yuan zan ( 普贤菩萨行愿赞 ) (T297), where the eight great bodhisattvas of the Aṣṭamaṇḍalakasūtra are understood as guardians of Locana of the Avataṁsaka Sūtra.96 Therefore, the Yulin painting (Fig. 1) represents the deities of the Aṣṭamaṇḍalakasūtra in the Huayan context. However, in contrast with the Aṣṭamaṇḍalakasūtra, which prescribes a strictly circular layout for the bodhisattvas, the Puxian pusa xing yuan zan (T297) doesn't provide any guidance in this regard. I believe, this explains a variety of spatial arrangements found in the representations of the group (discussed below). Furthermore, not only the central composition, but also the whole iconographic program of Yulin Cave 25 can be understood in the Huayan context. First of all, the characteristic Huayan depictions of Mañjuśrī, Samantabhadra, and Sudhana occupy the western wall (Fig. 2).97 The depictions of Amitābha’s and Maitreya's Pure Lands as a pair are common in Dunhuang art and their compositions are based on respective sūtras.98 However, their presence in Cave 25 might reflect an additional level of symbolism. For example, aspiration for rebirth in Amitābha's Pure land is emphasized in the Bhadracarī,99 while Maitreya is one of the most important spiritual friends of Sudhana who visits Maitreya's abode in the narrative of the Gaṇḍavyūha Sūtra. Therefore, Yulin Cave 25 can be understood as a manifestation of the donor's practice of the Bhadracarī and of their aspiration for enlightenment. As a matter of fact, references to the Bhadracarī and/or the Gaṇḍavyūha Sūtra can be found in almost all representations of VAEGB in Hexi and Tibet.100 94 Orzech, Sørensen, and Payne 2011: 91-92. 95 The characters na 那 and she 舍 have been transposed. See Wang 2018: 109-110. 96 Dessein 2003: 332. As was noted by Wang, Dessein mistakenly replaces bodhisattva Sarvanivāraṇaviṣkambhin with Acalanātha, see M. Wang 2018: 258, n. 59. 97 Paired images of Mañjuśrī and Samantabhadra, the major bodhisattvas of the Huayan tradition, were often depicted in Dunhuang caves starting from the early Tang dynasty. According to M. Wang, this pair reflects the interest in the Gaṇḍavyūha Sūtra narrative and the Bhadracarī. See Wang 2018: 247. 98 For example, Mogao Caves 329 and 321. See Wang 2005: 382-386. 99 Osto 2010: 5-6. 100 This also finds parallels at Borobudur, Java. Panel IV-3 from the Gaṇḍavyūha reliefs contains images of Sudhana and the eight great bodhisattvas. See Bautze-Picron 1997: 28, 54 schema 11; Fontein 2012, p. 173; Long 2009: 165-168. I am very grateful to Claudine Bautze-Picron for drawing my attention to the Borobudur depiction and providing these references. 73 2.1.2 Other Representations of VAEGB in Hexi and Eastern Tibet 2.1.2.1 A Wooden Shrine One striking example is a wooden shrine from the Nelson-Atkins Museum of Arts (Fig. 12). The shrine has attracted considerable scholarly attention and was variously ascribed to Central Asia, Kashmir, or Tibet.101 On stylistic grounds, I argue that it most likely originates from Hexi, as is evident from the Chinese facial features of the figures, the combination of the Indian and Chinese artistic modes, and the spiked armbands worn by Vairocana, which is a characteristic of the Southern Indian mode in Hexi art (see a close-up of the armbands in Fig. 12). The central section of the shrine shows Vairocana flanked by two vertical rows of four bodhisattvas who alltogether most likely represent the eight great bodhisattvas.102 Such a layout can be found in Southern Indian art from the 6th century onwards and this particular group is depicted in the same manner in a 9th or 10th century stone stele from Odisha.103 In the shrine, three more figures are placed below Vairocana. While Granoff and M. Wang identify this group as a representation of an abhiṣeka ceremony, Linrothe argues that it depicts a donor of the shrine being blessed by the bodhisattvas.104 I would argue that this group likely symbolises Sudhana's spiritual path towards enlightenment. The two bodhisattvas thus represent Mañjuśrī (on the left) and Samantabhadra (on the right), who together with a central figure of Vairocana (here Locana of the Avataṁsaka Sūtra) form the typical trinity of Huayan Buddhism. Therefore, a figure seated on a rug represents Sudhana, who is customarily depicted in Chinese art as a little boy with a shaven head.105 Accordingly, the depiction of Sudhana together with Mañjuśrī and Samantabhadra symbolises his spiritual journey as described in the Gaṇḍavyūha Sūtra; in his search for enlightenment, Sudhana meets fifty-three spiritual fiends (kalyānamitra) starting from Mañjuśrī and ending with Samantabhadra.106 At the same time, these three are part of the final stage of Sudhana's pilgrimage which culminates in the pronouncement of the Bhadracarī. Before his final meeting with Samantabhadra, the boy sees Mañjuśrī again and 101 The most notable publications include Granoff 1968/69; Linrothe 2014; M. Wang 2018. While Granoff ascribes the shrine to Central Asia, Linrothe suggests a Kashmiri provenance, and M. Wang argues for a Tibetan origin. See Granoff 1968/69: 81; Linrothe 2014: 31; Wang 2018: 83. 102 They were identified as such by Granoff. See Granoff 1968/69: 90-91. 103 For a 6th century example in Cave 23 at Nasik, see Bautze-Picron 2000: fig.9-10. For the Odisha stele, see Tanaka 2018: fig. 1.15. 104 Granoff 1968/69: 90; Linrothe 2014: 31; Wang 2018: 72. 105 On representations of Sudhana in Chinese art, see Fontein 1968: 23-110. 106 The representations of Mañjuśrī, Samantabhadra and Sūdhana in Yulin Cave 25 likely have the same meaning. Fig. 2 shows Sudhana depicted in the right corner of the Mañjuśrī painting and in the left corner of the Samantabhadra painting. 74 receives from him various kinds of teachings.107 In our composition, Mañjuśrīis shown offering Sudhana a round object. A depiction of this meeting can be also found in a wall painting in Mogao Cave 85 (late 9th century), where it is identified by an accompanying cartouche.108 There Mañjuśrī is also depicted offering Sudhana a round object. This object probably represents the jewel of the teaching, references to which can be found in the sūtra. For example, during their first meeting, Sudhana Fig. 12 Shrine and a close-up of the Vairocana's armband, wood with traces of of polychrome and gold, 9th-10th C., probably from the Hexi corridor, 31.1 × 35.6 cm., the Nelson-Atkins Museum of Arts, Kansas 107 Cleary 1993: 1502. City, USA. Object number 44-18 (Linrothe 2014: fig. 1. 108 Wang 2018: 250, fig. 119. 1). 75 supplicates to Mañjuśrī: "Instruct me, O King, with the jewel of the wheel of teaching."109 Properly prepared by Mañjuśrī, Sudhana sees Samantabhadra (depicted in the shrine as a bodhisattva with his right hand placed upon Sudhana's head).110 The crucial moment of their meeting, which takes place at the feet of Vairocana/Locana (exactly as shown in the shrine), culminates in the pronouncement of the Bhadracarī and is described in the Gaṇḍavyūha Sūtra as follows: "Just as in this world, in the presence of Vairocana Buddha, the enlightening being Universally Good (bodhisattva Samantabhadra) extended his right hand and laid it on Sudhana's head … And just as facets of truth entered Sudhana as he was touched by the hand of Universally Good at the feet of Vairocana Buddha, likewise facets of truth entered Sudhana in various ways as he was touched by the multitudes of hands extended from the bodies of all the Universally Goods."111 I believe that the image of an elephant, the vehicle of Samantabhadra, placed in the centre of the Vairocana's throne is a reference to the Bhadracarī. Douglas Osto remarked that Samantabhadra and the Bhadracarī were "associated with each other in the sense that Samantabhadra is thought to embody the aspirations contained in the verses."112 In addition and as was pointed out to me by Claudine Bautze-Picron, lion(s) and elephant(s) form a pair which is often encountered in the ornamentation of the pedestal of throne in India from an early period. The meaning imparted to the two animals is rich, but, in short, an elephant refers to enlightenment and a lion to the teaching.113 The Nelson-Atkins shrine carries an unobtrusive and mostly illegible Tibetan inscription on the back. It opens with an initial Tibetan mark (yig mgo) followed by the syllables byang chub. As was suggested by Robert Linrothe, these syllables may have originally stood for the words "awakening" (byang chub), "bodhicitta" (byang chub sems), "bodhisattva" (byang chub sems dpa'), which would correspond well to the proposed symbolism of the central scene. According to Yannick Laurent, the syllables may have alternatively stood for a personal Tibetan 109 Cleary 1993: 1175. 110 A similar representation of the scene where Samantabhadra lays his hand on Sudhana's head can be found in the Gaṇḍavyūha narrative panels located on the east wall on the fourth level at Borobudur, panel IVB-82. See https:// www.photodharma.net/Indonesia/11-Gandavyuha-Level-4/11-Gandavyuha-Level-4.htm, picture 82; Fontein 2012: 147 fig. 33. 111 Cleary 1993: 1508. 112 Osto 2010: 5. 113 Bautze-Picron, private communication 09.08.19. This symbolism is also evident in Huayan art, where a lion is the vehicle of Mañjuśrī, an embodiment of the teaching, while an elephant is the vehicle of Samantabhadra, an embodiment of the enlightenment. On the symbolism of these two animals in Indian art, see Bautze-Picron 2008. 76 name (i.e. Jangchub).114 It has to be also noted that there is no way to know when and where the inscription was added and its presence per se does not signify a Tibetan provenance of the shrine. It merely suggests that the shrine was at some point used by Tibetan-speaking people.115 2.1.2.2 A Portable Painting from Mogao Cave 17 Another version of the group of VAEGB can be found in a portable painting from Mogao Cave 17, which could be dated to the late 9th century on stylistic grounds (Fig. 13).116 The bodhisattvas are partially identified by Tibetan inscriptions and partially by iconographic attributes. As with the Yulin painting (Fig. 1), the figures display all defining characteristics of the Southern Indian mode. The layout of the composition is similar to that in the Nelson-Atkins shrine (Fig. 12) and the same as in the already mentioned stele from Odisha.117 The bottom foreground of the painting contains representations of sand hills and birds which are normally associated with depictions of Amitābha's Pure Land. As mentioned earlier, the Bhadracarī contains a reference to Amitābha's Pure Land that perhaps explains why its features are depicted in the painting. 118 2.1.2.3 Eastern Tibetan Rock Carvings There are three rock carvings in Eastern Tibet which depict a group of VAEGB (Fig. 52- 54).119 They have often been offered as evidence for the Tibetan provenance of the composition 114 Linrothe 2014: 31-34; Yannick Laurent, personal communication, August 2019. Laurent further remarked that the foreign provenance of this Buddhist shrine and its now effaced epigraph make it difficult to determine the nature of this Tibetan inscription. In comparison with other non-Tibetan portable images bearing Tibetan inscriptions, the epigraph of the shrine may have acted as a laudatory inscription, a descriptive caption, or even as an expression of ownership, as shown by Laurent during his presentation at the 15th Seminar of the International Association for Tibetan Studies in Paris, July 2019. 115 On a wide use of Tibetan language in Hexi and even in correspondence between non-Tibetan parties, see van Schaik and Galambos 2012: 67-68; Takata 2019. 116 For an overview of the scholarship and the latest discussion on the painting, see Wang 2018: 111-113. 117 Bautze-Picron; 1997: 27, 53 schema 10; Tanaka 2018: fig. 1.15. 118 This was noted by Michelle Wang as well. See Wang 2018: 258 n. 59. 119 For an overview of the carvings, see Huo 2018; 1) Renda high relief rock carving, dated 804/816, Brag gyab County, Chamdo Prefecture of TAR. On this carving see Heller 1994; Shanxi Province Academy of Archaeology and TAR Relics Preservation Institute 2014; 2) Sgar thog high relief rock carving, near the town of Sgar thog in Smar khams county, Chamdo Prefecture of TAR. On this carving, see Yang Qingfan, Lu Suwen and Zhang Yanqing 2017; 3) 'Bis khog high relief rock carving (also known as Bimda), dated 806, Yushu Tibetan Autonomous Prefecture, Qinghai province. On the most recent research on this carving and for an overview of the previous scholarship, see Huo 2017. 77 Fig. 13 Vairocana and the eight great bodhisattvas, ink and color on silk, 10th C., from Mogao Cave 17, 63.5 × 17.95 cm, the British Museum, London. Museum number 1919,0101,0. 50 (Ch. 0074) (Wang 2018: fig. 2). 78 in Yulin Cave 25.120 The carvings are concentrated in the areas bordering Tang empire and date to the early 9th century. At least two, if not all, of them were commissioned by the same person — Ye shes dbyangs, who was identified as a high-profile Tibetan monk listed in a Chan lineage in the later Dunhuang texts, and who, according to an inscription at Renda, commissioned many other carvings in the area.121 Altogether this strongly suggests that the carvings were created under the cultural influence of the Hexi Corridor.122 I argue that these rock carvings are associated with the Bhadracarī and, like the examples from Hexi, reflect the practice established by Amoghavajra in the Puxian pusa xing yuan zan (T297). A complete text of the Bhadracarī is carved about 100 meters away from the group of VAEGB at 'Bis khog.123 The inscription is considered part of the monument and covers a large portion of the rock that gives it an exceptional prominence at the site. In addition, a prayer to Vairocana and the eight great bodhisattvas is carved in 23 lines to the right of the group of VAEGB.124 This prayer corresponds to the eulogy to the eight great bodhisattvas in the Amoghavajra's text. A single religious text accompanies the rock carving at Renda. It was initially identified as an excerpt from the Bhadracarī, but this was questioned by Amy Heller.125 At present, the text is understood to be a short prayer, probably a local creation, which focuses on the bodhisattva Samantabhadra and is inspired by the Bhadracarī.126 In addition, the historical inscriptions at Renda and 'Bis khog have clear references to the main qualities of the Bhadracarī. They praise the path to enlightenment and dedicate the monuments to the wellbeing of the Tibetan Emperor, tsenpo (btsan po).127 Sam van Schaik and Lewis Doney have shown that the Bhadracarī was one of the most popular Buddhist prayers in Imperial Tibet and that it was understood as having the power to give the tsenpo longevity. Since many Tibetan copies of the prayer were found among the Dunhuang manuscripts, they have remarked that its popularity among Tibetans might have been due to Chinese influence.128 Indeed, even the link between the Bhadracarī and the tsenpo finds parallels in Chinese practice, as Amoghavajra's translation of the Bhadracarī was intended as a eulogy to the Chinese emperor as well.129 Furthermore, out of three extant Chinese translations, the one by 120 See footnote 53. 121 Heller 1994; On Ye shes dbyangs see also, van Schaik 2015: 163-174. 122 Huo Wei also argues that these carvings should be understood in the context of the Sino-Tibetan relations, see Huo 2018. 123 Huo 2017: 6-7. 124 Huo 2017: 5. 125 Heller 1994: n 6. 126 Shanxi Province Academy of Archaeology and TAR Relics Preservation Institute 2014: 14. 127 Heller 1994; Huo 2017: 6. 128 van Schaik and Doney 2007: 185 n. 19. 129 Dessein 2003: 331-332. 79 Amoghavajra most closely corresponds to the Tibetan version.130 Regarding artistic features, the groups of VAEGB in the carvings were depicted either in the Tibetan mode, with the figures clad in Tibetan-style robes (Figs. 53 & 54), or in the Indian mode (Fig. 52).131 The iconography of the group of VAEGB in the carvings corresponds to the Southern Indian and Hexi examples (Vairocana is shown with the dhyāna-mudrā). Compositionally, the Renda group is organized in the same way as the group in a portable painting from Mogao Cave 17 (Fig. 13) and in the stele from Odisha.132 It displays typical elements of Chinese design such as a canopy above Vairocana, which is almost identical to that in the Hexi examples (Fig. 1 & 5), and two apsarasas flying on Chinese-style clouds. The two other carvings follow the layout of the composition in Yulin Cave 25, though the bodhisattvas are depicted standing, as at the Ellora caves. Discussion on the transmission of the cult of VAEGB from Hexi to Tibet will continue in Part 4. Summary of section 2.1 The depictions of VAEGB in Hexi and Eastern Tibet reflect the popularity of the Bhadracarī. The central figure in these depictions represents Vairocana of the Avataṁsaka Sūtra, whose iconography was fused with that of Vairocana of the Aṣṭamaṇḍalakasūtra. The connection between the Bhadracarī and the Aṣṭamaṇḍalakasūtra was established in Amoghavajra's work the Puxian pusa xing yuan zan (T297), which integrated the eight great bodhisattvas of the Aṣṭamaṇḍalakasūtra into the practice of the Bhadracarī. The Southern Indian stylistic elements detected in the Hexi depictions confirm the context of the Amoghavajra teachings and, by extension, the Southern Indian origin of the cult of VAEGB in Hexi. The Tibetan occupation of the Hexi Corridor in the mid 8th century enabled religious and artistic transmission from this region to Tibet. In this regard, the discussed eastern Tibetan rock carvings bear material witness of that process. 2.2 Maṇḍala-like compositions featuring Vairocanābhisaṃbodhi At both Mogao and Yulin, Vairocanābhisaṁbodhi and the retinue figures are consistently depicted in the Southern Indian mode. Figures 14, 16, and 17 show maṇḍala-like 130 Osto 2010: 2. 131 It is difficult to go beyond this general observation, as in their current state 'Bis khog and Renda figures are freshly plastered and painted, while the Sgar thog carving is significantly damaged. Initially, the reliefs were likely covered with plaster, in which smaller details were made. Such details could have been lost over time and/or altered during subsequent renovations. 132 See footnote 117. 80 Fig. 14 Maṇḍala-like composition centered around Vairocana, wall painting, second half of the 9th C., Guiyijun period (848-1036), Mogao Cave 14, Gansu, China (Wang 2018: fig. 73). Circles indicate offering goddesses. compositions133 from Mogao Cave 14, Yulin Cave 20, and Yulin Cave 38 dating from the 9th to 11th centuries.134 The compositions feature Vairocanābhisaṁbodhi surrounded by bodhisattvas, Offering Goddesses, gatekeepers and/or guardians of the directions. Each of these maṇḍalas is accompanied by a second maṇḍala in the same style depicted on the same or opposite wall and centred around Vajrasattva in Mogao Cave 14 (Fig. 15), and a Buddha in the two other caves. 133 Although these compositions look like maṇḍalas, it is not clear if they were created and used in a ritual context that established their function as maṇḍalas. For a definition of "maṇḍala," see Orzech, Sørensen, and Payne 2011: 81-82. 134 For a review of the scholarship and a discussion on these compositions as well as for additional images, see Wang 2018: 156-168, fig. 28, 29, 72, 73, 74, 80, 81. On the ground of their style, M. Wang ascribes the maṇḍalas to the Tibetan tradition. She designates them as "Maṇḍala of eight great bodhisattvas" and groups them together with the previously discussed depictions of VAEGB. In my view, this can be misleading, as the maṇḍala of the eight great bodhisattvas contains only nine deities and is based on the Aṣṭamaṇḍalakasūtra. Compositions with more deities are based on different texts and can feature different eight bodhisattvas. Moreover, the eight great bodhisattvas in the maṇḍala of the Aṣṭamaṇḍalakasūtra are not the same eight bodhisattvas as those found in maṇḍalas of the Yoga cycle, to which the discussed examples most likely belong. See, for example, Tanaka 2018: 194-195. 81 Fig. 15 Maṇḍala-like composition centered around Vajrasattva, wall painting, second half of the 9th C., Guiyijun period (848-1036), Mogao Cave 14, Gansu, China (Wang 2018: fig. 75). Textual sources behind these twin-maṇḍalas have not been identified.135 What is clear, however, is that they display iconography similar to that of basic maṇḍalas of the Yoga tradition, which, for instance, can be recognized by presence of such distinctive deities as four or eight Offering Goddesses, who also sometimes are depicted in male forms as Offering Bodhisattvas (marked by circles in the images).136 For example, Cave 14 has the maṇḍala-like compositions centred around Vairocanābhisaṁbodhi (Fig. 14) and Vajrasattva (Fig. 15). Somewhat similar maṇḍalas with the same central deities are described in the Amoghavajra’s translation of the Prajñāpāramitānaya Sūtra ( 般 若 理 趣 经 ) (T243). According to Kimiaki Tanaka, this sūtra is the oldest in the Yoga cycle and it greatly influenced the development of the other texts, 135 The central deities in the maṇḍalas in Yulin Cave 20 are identified by accompanying inscriptions, which give the same name "the pure dharmakāya Vairocana Buddha" (qingjing fashen Piluzhena fo 清净法身毘卢遮那佛 ), see Wang 2018: 165. The name Piluzhena fo corresponds to the name of Vairocana in the eight fascicle translation of the Avatamsaka sūtra (T279: 39a 17-18), see Wang 2018: 109. Similarly with the case of Yulin cave 25, this likely reflects adaptation of Esoteric teachings for the Huayan context. 136 On the Offering Goddesses as a feature of the basic maṇḍalas of the Yoga cycle, see Tanaka 2018: 279-288. In fig.15, it is difficult to distinguish Offering Goddesses/Bodhisattvas from other bodhisattvas. 82 Fig. 16 Maṇḍala-like composition centered around Vairocana, wall painting, second half of the 9th C., Guiyijun period (848-1036), Yulin Cave 20, Gansu, China (Wang 2018: fig. 28). Circles indicate Offering Goddesses/Bodhisattvas. including the STTS.137 Thus, it is clear that there is a strong affiliation between the Southern Indian artistic mode and the Yoga tradition. 2.3 Seated Amoghapāśa with four arms There are three portable paintings from Mogao cave 17, which depict maṇḍalas centred on 137 On the maṇḍalas of the Prajñāpāramitānaya Sūtra, see Tanaka 2018: 248-280. The first and second maṇḍalas described in the text are centred around Vairocanābhisaṁbodhi (21 deities) and Vajrasattva (17 deities), respectively. There are versions, where the second maṇḍala is centred on the Buddha instead of Vajrasattva (see Tanaka 2018: 204) that corresponds to examples in Yulin caves 20 and 38. In Cave 14, the central deities of the maṇḍalas are swapped, with the Vairocanābhisaṁbodhi maṇḍala containing 17 deities and the Vajrasattva maṇḍala containing 21 deities. 83 Fig. 17 Maṇḍala-like composition centered around Vairocana, 10th/11th C., Guiyijun period (848-1036), Yulin Cave 38, Gansu, China (Wang 2018: fig. 80). Circles indicate offering Goddesses/Bodhisattvas. seated Amoghapāśa with four arms (Figs. 18, 20, 21).138 The painting in Fig. 18 is traditionally dated to the 8th or 9th century.139 While four-armed Amoghapāśa in the centre is identified by a noose held in the upper left hand, the presence of four Offering Goddesses (marked by circles in Fig. 18) points the context of the Yoga teachings.140 Amoghapāśa is an Esoteric variant of Avalokiteśvara; he can manifest in different forms, including peaceful and wrathful, standing and seated, with four, six, eight or more arms, and so on. There are nine Chinese translations of several versions of the Amoghapāśadhāraṇīsūtra dating from the 7 th to 10 th centuries, including one by Amoghavajra. 141 Iconographic 138 The paintings were discussed by Kimiaki Tanaka, who relates two of them (Figs. 18 and 20) to the Yoga tradition, see Tanaka 2000: 39-71. 139 It was discussed by Karl Debreczeny in Brauen 1997: 80-81, where its Indianized style was ascribed to Tibetan influence. 140 See footnote 136. 141 Wong 2007a: 152. 84 Fig. 18 Amoghapāśa Maṇḍala (shown without mount), 8th-9th C., from Mogao Cave 17, 204.5 × 107.5 cm with mount, the Musée Guimet, Paris. Museum number ART 371790 (Brauen 2009: plate 1). Circles indicate Offering Goddesses. 85 Fig. 19 Left: Viṣāpaharaṇa Śiva, 9th C., Pallava period, copper alloy, H 62 cm, Government Museum, Chennai, Tamil Nadu, India (American Institute of Indian Studies photo archive, https://vmis.in/ ArchiveCategories/gallery?search=visapaharana); Right: Amoghapāśa, detail of fig. 18. descriptions are rarely given in the texts and it is often simply mentioned that Amoghapāśa looks like Śiva.142 This probably explains the great variety of his iconographic forms, since Śiva has many iconographic forms as well. At the same time, at least three translations mention Amoghapāśa with one head and four arms.143 Amoghapāśa and his retinue in our painting (see close-up in Fig. 19 on the right) are depicted in the Southern Indian mode. Although no comparable Southern Indian images of Amoghapāśa are known, there are many contemporaneous images of four-armed Śiva seated with one leg pendant. These images can represent different forms of Śiva depending on the presence or absence of accompanying figures, hand gestures, and attributes. The late Pallava sculpture of Viṣāpaharaṇa Śiva in Fig. 19 (on the left) is very close to Amoghapāśa in the 142 Reis-Habito 1999: 48. 143 These are the translations by Maṇicintana ( 宝思惟 ; 693), Li Wuchan ( 李无谄 ; 670), and Bodhiruci ( 菩提流支 ; 707-9). See Reis-Habito 1999: 52. 86 Fig. 20 Amoghapāśa Maṇḍala, ink and color on silk, from Mogao Cave 17, 10th C., 115 × 65 cm, the Musée Guimet, Paris, France. Museum number EO. 3579 (Giès 1994-96: plate 99). Circles indicate Offering goddesses. painting. Both figures have three eyes and are adorned with large armbands decorated with spikes. The similarities in their proportions and in the position of the second pair of the arms are striking. The two other maṇḍalas are shown in Figs. 20 and 21.144 They date to around the 10th century and display some iconographic and stylistic variations in comparison to the earlier example. Nevertheless, the deities in the paintings show distinctive elements of the Southern Indian mode (spiked decorations, swagged necklaces) as well. The eight Offering Goddesses (marked by circles) are clearly recognisable, see Fig. 20. The composition in the Fig. 21 is organized similarly to that in Fig. 18, though only eight deities are depicted and the identities of some of them remain debatable. The colours of the four figures marked by circles match 144 Recent discussions and a review of the scholarship on the painting in Fig. 20 can be found in McCoy 2016 and Wang 2018: 184-187. The authors argue for a Tibetan background of the painting. 87 Fig. 21 Amoghapāśa Maṇḍala, ink and color on silk, from Mogao Cave 17, 10th C., 89.6 × 60 cm, the Musée Guimet, Paris, France. Museum number EO. 1131 (Giès 1994-96: plate 80). the colours of the Offering Goddesses in Fig. 19 and at least two of them are clearly female.145 Overall, the painting gives an impression that it was either unfinished or reworked.146 To summarise, all three paintings (Figs. 18, 20, 21) depict the deities in the Southern Indian mode, and at least two of them are clearly related to the Yoga tradition. 2.4 Yulin Cave 15 Yulin Cave 15 also contains a composition executed in the Southern Indian mode (Fig. 22).147 It features a god known under various names, including Kubera and Vaiśravaṇa. While the 145 Tanaka identifies them as deities of the Lotus family of the VAT. He dates the painting to the second half of the 8th century. Such an early date cannot be accepted from a stylistic point of view. Tanaka 2000: 39-44. 146 It looks like one figure is missing in the top of the painting; also the figure in the upper left corner curiously has three arms. 147 Sha Wutian argues that the painting is executed in the Tibetan style. See Sha 2013: 117-118. 88 etymology of the former is debatable, the latter derives from the name of his father Viśravis.148 In India, Kubera is the guardian of the North in a group of the eight dikpālas and the king of yakṣas and other demigods.149 In China, the god is mostly known as Vaiśravaṇa ( 多闻天王 ), the guardian of the North in a group of the four heavenly kings.150 Vaiśravaṇa is traditionally depicted in an armoured form, which probably has a Central Asian origin.151 In Cave 15, the god is represented as a yakṣa holding a club and a mongoose, which corresponds to the iconography of Kubera. Close parallels can be found amongst sculptures of yakṣas at Ellora, two of which are shown in figures 23 and 24. These and many other similar examples belong to Ellora's Jain caves.152 Lisa Owen has recently discovered that these caves reflect the patronage of Jains from the regions of Tamil Nadu and Karnataka.153 This observation corresponds well to the Pallava influenced style of the sculptures. Owen has remarked that few comparable depictions survived in Tamil Nadu and Karnataka,154 and thus the Ellora sculptures are especially valuable with regard to reconstructing artistic traditions of these regions. This observation provides evidence for a probable Tamil Nadu background of the Yulin painting. In terms of iconography, images of Brahmanic, Buddhist and Jain yakṣas represent a shared concept of wealth and protection and display only minor differences.155 Like the Yulin Kubera, the Ellora yakṣas are placed under an arched canopy formed by a tree with luxuriant foliage and flowers. Such design with a tree as part of a throne back and luxuriant foliage depicted above a divinity is unusual for the art of the Deccan, but is common in the examples from Tamil Nadu.156 The figures are seated with one leg pendant and flanked by two attendants, one of which (on the right) carries a purse made of mongoose skin. In contrast to the Ellora yakṣas, who are mounted atop elephants (a typical vehicle of the Jain yakṣas), the 148 On yakṣa cult and iconography, see Misra 1981; Bautze-Picron 2002. In particular, on Kubera/Vaiśravaṇa, see Misra 1981: 59-71. 149 On the iconography and symbolism of the guardians of the directions, dikpālas, in India, see Wessels-Mevissen: 2001. 150 On the development of iconography and symbolism of the four heavenly kings in Central and East Asia, see Shim 2013. 151 Shim 2013: 82-112. 152 On Ellora's Jain caves, see Owen 2012. In particular, on depictions of yakṣas at Ellora, see Owen 2012: 81-127. 153 Owen 2012: 16, 20, 160-163. 154 Owen 2012: 127-129. 155 Owen 2012: 127-129; Bautze-Picron: 2002. 156 I am grateful to Claudine Bautze-Picron who has drawn my attention to this. Comparable examples can be found in the Jain rock carvings at Kalugumalai, the 9th or early 10th century, Tamil Nadu, see Owen 2010: figs 3, 11. Also, see a 9th century Buddha from Tamil Nadu, now in the Museum of Fine Arts, Boston (accession numbers 1970.3 and 67.4a-b); and a 10th century Buddha with two bodhisattvas, from Nāgapaṭṭinam, Tamil Nadu, now in the Government Museum Chennai (accession number 81212). 89  Fig. 22 Kubera with attendants, wall painting, Tibetan period in the Hexi corridor, Yulin cave 15, China (Dunhuang Academy 2014: pic. 40). Yulin Kubera is seated on a throne.157 Nevertheless, both figures are similarly decorated with swagged necklaces and spiked crowns and armbands. Additionally, the images in Figs. 22 and 24 have almost identical square backrests and halos of an inverted U-shape. The appearance of such Kubera images in Hexi could be related to protective rituals centred on Vaiśravaṇa, rituals which Amoghavajra regularly performed to provide imperial 157 Kubera seated on a throne is depicted in the early 8th-century Ellora Cave 25, but the image is significantly damaged. See Owen 2012: fig. 57. 90 Fig. 23 Fig. 24 Yakṣas with attendants, early 9th C., Ellora Cave 32, Maharastra, India (American Institute of Indian Studies photo archive, https://vmis.in/ArchiveCategories/collection_gallery_zoom?id=923& search=1&index=57339&searchstring=ellora,%20cave,32). troops with supernatural assistance. The earliest recorded instance of Amoghavajra performing the Vaiśravaṇa rituals is specifically linked to the Tibetan siege of Liangzhou.158 There are five texts in the Taishō Canon attributed to him that specifically invoke Vaiśravaṇa for military magic.159 In these texts, the god is consistently referred to as a king of yakṣas. Frustratingly, his physical appearance is not always described and where it is given, the description is of his armoured form.160 However, Yulin Kubera is shown with the same two attendants as those 158 Yang 2018: 86-87. 159 These rituals and texts have been discussed in Goble 2013. 160 Geoffrey Goble, private communication 23.07.2018. 91 Fig. 25 Bodhisattva Samantabhadra, ink and color Fig. 26 Bodhisattva, ink and color on silk, from on silk, from Mogao Cave 17, 9th-10th C., Mogao Cave 17, 41 × 18.5 cm, 9 th C., 59.8 × 17.8 cm, the Musée Guimet, Paris, the British Museum, London, UK. Museum France. Museum number MG. 1770 number Ch. Iv. 007 (Giès 1994-96: plate 4). (Whitfield 1982-85: plate 53). 92 that are normally depicted with Vaiśravaṇa in Dunhuang art.161 This suggests that, despite the differences in iconography, he was understood to be the same god as Vaiśravaṇa.162 It can be surmised that the aforementioned protective rituals were of an Indian origin and therefore they must have been initially centred on Kubera, whose images likely accompanied the original Indian texts, but who was replaced with his well known local equivalent, Vaiśravaṇa, in the Chinese translations. 2.5 Portable Paintings with Single Figures of Bodhisattvas Figs. 25 and 26 show images of solitary standing bodhisattvas in the Southern Indian mode. Their vertical format implies that the paintings were displayed on banners and possibly produced in sets representing a group of eight bodhisattvas. Parallels can be found at Ellora as well, where Caves 11 and 12 contain several groups of the eight great bodhisattvas depicted standing.163 An outstanding painting in Fig. 27 was likely conceived of as an individual icon. The bodhisattva sports a tubular and extremely tall hairstyle (jaṭāmukuṭa), a typical feature of Pallava imagery and Sri Lankan imagery in the Pallava style. A figure attributed as Tārā from the British Museum in Fig. 28 is one striking example of this. The sculpture was discovered in Sri Lanka, but it possibly has a Pallava provenance.164 The bodhisattva and Tārā are clad in lower garments of a similar design and their headdresses are secured by nearly-identical leaf-like crown elements. Unlike the other discussed images, the crown and armbands of the bodhisattva do not have spiked elements. This, however, does not rule out a possible Pallava background as decorations without spikes were also common in Pallava art. Alternatively, the painting could have been based on a Sri Lankan original, as spiked decorations were not prominent in Sri Lankan art.165 Although the painting is currently attributed to the 9th century, 161 See, for example, a woodblock print dated to 947 in the British Museum collection (1919.0101.0.245). The attendants are a female attendant with a tray of jewels (identified by the museum as Goddess Śrī) and a gandharva (a class of demigods) wearing a tiger-head helmet that covers his head and shoulders. On depictions of gandharva in Chinese art, see Wong 2019: 73 fig 2.14, 120 fig. 3.8. Another example of Vaishravana attended by gandharva is a 9th century painting 1919,0101,0.38, in the British Museum collection. 162 The conflation between the dikpālas and the heavenly kings is also evident in the compositions in Figs. 14 and 15, which were discussed in the previous section. While the North, South, and West directions are guarded by the dikpālas (Kubera is depicted in the right top corner in Fig.14 and in the left top corner in Fig.15; Yama and Varuṇa are in the bottom corners), the East is guarded by the heavenly king of the East, Dhṛtarāṣṭra, who is clearly recognized by his attribute, a lute. 163 Malandra 1993: figs. 222, 223, 251, 252. 164 Srinivasan 1999: 110. 165 See footnote 64. 93 Fig. 27 Bodhisattva, ink and color on silk, 8th C., from Fig. 28 Tara ?, copper alloy, 8th C., Sri Lanka Mogao Cave 17, 57.5 x 38.1 cm, the British or Pallava Kingdom, H 143 cm, the British Museum, London, UK. Museum number Ch. 18003 Museum, London, UK. Museum number (Whitfield 1982-85: plate 4). 1830,0612.4 (https://www.britishmuseum.org/research/ collection_onlinecollection_object_ details.aspx?objectId=251954&partId=1). its almost pure Pallava/Sri Lankan aesthetics suggest the 8th-century date. 2.6 Examples of Mixed Imagery As discussed in section 2.1, the bodhisattvas in Yulin Cave 25 (Fig. 1) were likely painted in accordance with standardized models known in Chinese art as yang. Several portable paintings discovered in Mogao Cave 17 suggest that yang models of bodhisattvas in the Southern Indian mode were perhaps also used for depicting attending bodhisattvas in heavenly assemblies where the rest of the figures are in the Chinese mode. Perhaps the earliest example of such imagery is a portable painting shown in Fig. 29, 94 Fig. 29 Pure Land of Bhaiṣajyaguru, ink and color on silk., from Mogao Cave 17, dated 836, Tibetan period, 152.3 × 177.8 cm, the British Museum, London, UK, Museum number Ch.xxxvii.004 (Whitfield 1982-85: plate 16). which belongs to the Tibetan period and is dated to 836. The painting contains Chinese and Tibetan inscriptions and displays a mixture of the Chinese and Indian artistic modes. It has been suggested that the Indianized images in the painting represent the Tibetan style and their appearance along with Chinese images is a sign of Tibetan influence, which materialized in Dunhuang art during the period of the Tibetan occupation.166 I would argue that neither in terms of iconography nor in terms of style can this painting be ascribed to Tibetan influence. The painting was likely centred on Cintāmaṇicakra, Thousand-armed Avalokiteśvara, and Amoghapāśa, which was a popular grouping in Dunhuang. Their images are largely lost, but the deities are mentioned in the accompanying Chinese inscription. Cults of these Esoteric forms of Avalokiteśvara gained popularity during the reign of Empress Wu ( 武 则 天 , r. 684­-705) and were promoted by the court along with the Huayan doctrine.167 As a result, images of these deities are often found along with those belonging to the Huayan context.168 In this painting too, characteristic figures of two main bodhisattvas of the Huayan tradition, Samantabhadra riding an elephant and Mañjuśrī riding a lion, are depicted right above the Esoteric versions of Avalokiteśvara. Finally, an assembly of Buddha Bhaiṣajyaguru occupies the top of the painting. 166 Karmay 1975: 8-14; Wang 2018: 98-107. 167 On cults of Esoteric deities during the Tang, see Wong 2018: 184-200. 168 On Esoteric imagery in Huayan context, see Wong 2012: 223-260. 95 Fig .30 Ten stages of the bodhisattvahood according to the Avataṁsaka Sūtra, ink and color on silk, from Mogao Cave 17, 10th C., 286 × 189 cm, the Musée Guimet, Paris, France. Museum number MG 26465 (Giès 1994-96: plate II-1). All these subjects were popular in Dunhuang before the Tibetan period and altogether they can be found at already discussed Mogao cave 148, which was excavated between 766-776 (i.e. in the decade before the Tibetan conquest).169 As with Mogao Cave 148 and Yulin Cave 25, the painting shows the coexistence of the two artistic modes. The bodhisattvas flanking the figure of Bhaiṣajyaguru are shown in the Indian mode, while the rest of the figures are depicted in the Chinese mode. These bodhisattvas represent the same type as that in the VAEGB composition of Yulin Cave 25 (Fig. 1) and display the characteristics of Southern Indian art such as spiked decorations and swagged necklaces (see a close-up of the bodhisattva in Fig. 29). A large rectangular cartouche in the centre of the painting contains Chinese and Tibetan inscriptions. The Chinese inscription is almost completely faded and illegible to the naked eye, while the 169 The iconographic program of Mogao cave 148 includes sculptural representations of Cintāmaṇicakra and Amoghapāśa (no longer extant) and a painted image of Thousand-armed Avalokiteśvara along with distinctively Huayan images of Samantabhadra riding an elephant and Mañjuśrī riding a lion as well as several heavenly assemblies including that of Buddha Bhaiṣajyaguru. See Lee 2010: 91. 96 Tibetan inscription is relatively clear.170 The Chinese inscription includes the date and was read with the help of infrared photography. Although the inscriptions are similar in their content, it is clear that they are not translations of each other. The Tibetan one mentions an image of Pariṇatacakra instead of Amoghapāśa in the Chinese inscription, and furthermore dedicates the painting to all living beings, whereas the Chinese inscription dedicates it to the donor's deceased parents. Bearing in mind that the Chinese inscription is faded and the majority of the painting is lost, it could be that the painting was restored at some point and then rededicated with the Tibetan inscription. It is also worth remembering that during the Tibetan period the Tibetan language was the language of the ruler and was used in Hexi in parallel with the Chinese and not only by Tibetan people.171 In this regard, a Tibetan inscription per se cannot be regarded as evidence for the Tibetan religious and/or artistic influence. Another example of a similar synthesis is a 10th century painting from the Musée Guimet (Fig. 30) that depicts the ten stages of bodhisattvahood according to the Daśabhūmika Sūtra, which itself is a part of the Avataṁsaka Sūtra. The bodhisattvas in the Southern Indian mode are included in one of the assemblies whereas the other figures are depicted in the Chinese mode.172 The examples discussed in this chapter likely reflect the general popularity of Indian icons in Tang China as well as the creativity of Dunhuang artists.173 Conclusion for Part 2 The images from Mogao and Yulin discussed above share a great deal of iconographic and stylistic similarities with Pallava and Pallava-influenced examples from Southern India. Most of them can be related to the Yoga and other teachings propagated by Vajrabodhi and Amoghavajra. Their stylistic consistency, which is already evident in the earliest examples (late 8th century), confirms that this artistic tradition emerged earlier and thus it should be associated with Vajrabodhi as a painter, who was active in the Tang capitals from 720 to 741. The populiarity of such imagery in Dunhuang art in the second half of the 8th century is in all 170 For the photograph of the inscriptions, see Wang 2018: 102, fig. 44. 171 On the use of Tibetan language in Hexi during and after the Tibetan period, see van Schaik and Galambos 2012: 67- 68; Takata 2019. 172 The iconography of the painting is discussed in Wong 2007b: 343. According to Wong, the painting is only known illustration of this subject. 173 In addition, in a British Museum painting from Mogao cave 17 (Ch.XXVI. a. 007), which represents Huayan Mañjuśrī riding a lion, the bodhisattva is depicted in the Southern Indian mode as well. Another example from the British Museum collection where attending bodhisattvas are shown in the Southern Indian mode is a painting Ch.ivi.0034. 97 likelihood related to Amoghavajra's activities in the region. Since the dates of the discussed imagery span from the 8th to the 11th century, it can be surmised that Amoghavajra's teachings were adopted by the Tibetans in Hexi and remained popular in the region even after the collapse of both the Tibetan and Tang empires. They were, nevertheless, suppressed by new tantric teachings which started arriving to Hexi from northeastern India in the late 10th century.174 This new wave of Esoteric transmission brought along new aesthetics characterised by the Pāla artistic style. Part 3. A Brief Overview of the Southern Indian Influence on Buddhist Art in China This part aims to demonstrate that the existence of a Southern Indian mode in Hexi art should not be perceived as disturbing or exceptional. There are other examples in Tang Buddhist art that complement the Hexi material and reflect the centuries-long interactions between the polities of Southern India, Sri Lanka, Southeast Asia, and China. Historical records from the 3rd century CE onwards tell us about a great number of Buddhist monks travelling between India and China by sea.175 Sri Lanka, whose Buddhist culture was intimately linked to that of Southern India, was an important transit stop for many of them. Early examples of these travelling monks include Faxian ( 法显 , 337/342-422), who spent two years in Sri Lanka on his way from Northern India to China; and Bodhidharma ( 菩 提达摩 ), the founder of Chinese Chan Buddhism, who travelled from the Pallava Kingdom to China in the late 5th or early 6th century. Indianized polities of Southeast Asia played an important intermediate role in Buddhist and artistic transmission from Southern India to China. The Indianization of the region started in the early historical period and was facilitated by Tamil seafaring merchants. The process included the adoption of Southern Indian religious cults, models of statehood, languages, scripts, and artistic styles. Early Buddhist and Brahmanic icons found in Southeast Asia display a dominating Southern Indian influence.176 According to Chinese sources, the Southeast Asian 174 In the late 10th century, a large number of Chinese monks were sent to northeastern India through the Hexi corridor in search of new Esoteric teachings. At the same time, many Indian monks arrived in China attracted by the lavish Buddhist patronage of the early Song emperors. It can be assumed that travelling through the Hexi Corridor, these monks shared teachings and imagery with local Buddhist communities. On these movements, see van Schaik and Galambos 2012: 35-61; Sen 2016: 102, 110-125. The earliest examples of Pāla style imagery in Hexi are found in the early 11th-century Mogao Cave 76, see Toyka-Fuong 1998. On Pāla-style imagery in Hexi art, see Khokhlov, forthcoming. 175 On Buddhist monks travelling between China and India by sea, see Acri 2018. 176 Guy 2014: 3-21. 98 Fig. 31 Buddha, painted stone, Northern Qi dynasty (550-577), H 156 cm, excavated in Longxing monastery, Qingzhou, Shandong province, China (Okada 2004: fig. 21-4). Fig. 32 Buddha, stone, Ikshvaku period (3rd-4th C.), H 293 cm, from Nagarjunakonda valley, Archeological survey of India, Nagarjunakonda, Guntur district, Andra Pradesh, India ( http://www. museumsofindia.gov.in/repository/record/nkm_ hyd-mar-scu-0008-7). polities of Funan (Mekong delta, territory of Cambodia), Lynyi (territory of Central Vietnam) and Panpan (Malay peninsula, territory of Thailand) often sent Buddhist monks, texts, and images to the dynasties of Southern China.177 Although Buddhist art of the Southern Chinese dynasties before the Tang is almost completely lost, in his pioneering research on textual sources Alexander Soper argued that in the late 5th and 6th centuries it was greatly influenced by the Southern Indian artistic tradition transmitted by maritime routes as well as through the Indianized polities of Southeast Asia.178 Soper states that artists from Southern China were famed as great specialists in rendering "Indian Buddhist icons" and their legacy can be traced in Buddhist art from the Northern Qi (550-557) to the Northern Song dynasties (960-1127).179 Soper supported his argument by pointing to Southern Indian features found in some Northern Qi examples.180 His theory can be further confirmed thanks to sculptures recently excavated in the eastern coastal area of 177 Wade 2014: 25-31. 178 Soper 1960. 179 Soper 1960: 88-89. 180 Soper 1960: 95-96, fig. 20. 99 Shandong province.181 A Northern Qi Buddha sculpture in Fig. 31 shows typical features of the late Amaravati tradition of Southern India and it is comparable to a Buddha sculpture from Nāgārjunakoṇḍa in Fig. 32. Both figures have a low uṣṇīṣa, elongated and columnar bodies clothed in the distinctive Amaravati way with a garment cloth pulled diagonally across the body that leaves the left shoulder bare. As was discussed in Part 1, interactions between Southern India and China increased dramatically during the Tang dynasty. However, the resulting artistic current in Tang art has rarely been highlighted in art historical literature. The problem partially lies in the lack of surviving material from Chang'an and Luoyang where foreign influences on art were especially strong. The capitals were prime destinations and places of residence for the majority of foreign monks and this was also true for Vajrabodhi and Amoghavajra. These metropolises suffered badly from the calamities of Chinese history and none of their hundreds of Buddhist temples have survived.182 Nevertheless, in addition to the discussed Dunhuang images, the Southern Indian facet of Tang Buddhist art can be further illustrated by a few more examples. 3.1 Longmen Buddha Sculptures About a hundred of sculptures depicting the Buddha seated in bhadrāsana are distributed in several caves at Longmen ( 龙 门 ) and Gongxian ( 巩 县 ) (Fig. 33). Each one of them is about one meter high and is dated by inscriptions to the period between 655 and 680. These inscriptions name the sculptures as "Buddha of the King of Udayana."183 Amy McNair argues for a Northern Indian stylistic origin of these images and compares them with a Buddha figure from Sarnath. This attribution ignores clear features of Southern Indian art present in the sculptures, such as low uṣṇīṣa, bare right shoulder, and ankle-sweeping robes, a vestige of the Amaravati tradition. The closest comparable examples are metal sculptures excavated in Andhra Pradesh and Tamil Nadu and datable to around the 6th century (Fig. 34).184 Several sculptures with similar stylistic characteristics are also known from Southeast Asian polities that suggests a possibility of the transmission through Southeast Asia as well.185 181 For a discussion on this group of sculptures, see Okada 2004. 182 Wong and Field 2008: 132-135. 183 On these sculptures, see McNair 2007: 102-103. 184 In addition to the image in Fig. 34, see Buddha figures from the Buddhapad hoard, now in the British Museum collection (1905.1218.2; 1905.1218.3): https://www.britishmuseum.org/research/collection_online/collection_object_ details.aspx?objectId=225203&partId=1&searchText=buddhapad&page=1 https://www.britishmuseum.org/research/collection_online/collection_object_details.aspx?objectId=225204&partId= 1&searchText=buddhapad&page=1 185 See Guy 2014: cat. 20-23, 50, 51. 100 Fig. 33 The "King Udayana Buddha" images, stone, ca. 655-680, Tang dynasty, H ca. 100 cm, Longmen, Henan province, China (Revire 2012: fig. 11. 8). Fig. 34 Buddha, copper alloy, 5th-6th C., from Nagapattinam, Tamil Nadu, India, H 40.6 cm, the Metropolitan Museum, New York, USA. Accession number 1998.414 (https://www.metmuseum.org/art/ collection/search/45449). 3.2 Two Metal Sculptures The Metropolitan Museum of Art has a metal figure of a bodhisattva (Fig. 36) that was reportedly found in Vietnam and dated by the museum to the 7th century. Another similar figure attributed to the Sui Dynasty (581-618) of China has been recently published (Fig. 35).186 These figures show a synthesis of Chinese and Southern Indian aesthetics, similar to what is found in the Mogao painting in Fig. 25. Both display distinctive swagged necklaces, while the crown of the bodhisattva in Fig. 35 is decorated with spikes. The date of their production is unlikely to be earlier than the 8th century, as the figures show closer connections with late Pallava material, such as the sculpture in Fig. 37. The fact that the territory of Northern Vietnam (Da Viet, Red river delta) belonged to Tang China suggests a probable southern Chinese origin for these sculptures. The occurrence of such imagery in southern China is logical, as Guangzhou was a major entry point to China for seafaring merchants and monks arriving from 186 Schulenburg, Hessel, Schmidt, and Wagner 2015: 150-151. 101 Fig. 35 Bodhisattva, copper alloy, 8th C., Tang dynasty, China, 43 cm, private collection (Schulenburg, Hessel, Schmidt, Wagner K., Brauen 2015: fig. 24). Fig. 36 Bodhisattva Avalokitesvara, copper alloy, 8th C., Tang dynasty, China, 36.5 cm, the Metropolitan Museum, New York, USA. Accession number 1993.387. 14 (https://www.metmuseum.org/art/ collection/search/51170). Fig. 37 Bodhisattva Maitreya, copper alloy, 8th-9th C., Pallava period, 39.5 cm, from Tanjor, Tamil Nadu, formerly in the Government Museum, Chennai, India, stolen, present whereabouts unknown (Snellgrove 1978: fig .90). South Asia. As was previously mentioned, three Tamil temples existed in Guangzhou in the 8th century that clearly demonstrates significant religious transmission. Moreover, Amoghavajra’s teachings were popular in the area as well. The master preached in Guangzhou right before his trip to South Asia. He attracted a great number of local followers, converted thousands of people to Buddhism, and conferred Esoteric initiations on members of the ruling elite.187 In 187 Yang 2018: 33. Chou 1945: 288-289. 102 Fig. 38 Buddha Shakyamuni preaching, silk embroidery, probably Chinese work, 8th C., Tang Dynasty, 207×157 cm, Nara National Museum, Japan (https://www.narahaku.go.jp/english/collection/647-0.html). addition, upon his return to China, Amoghavajra was temporarily banished from the court and spent four years (749-753) in the Guangzhou area.188 It is worth noting that the sculptures under discussion are visually close to a famous icon of the Nanzhao and Dali kingdoms, the Acuoye Guanyin, which can also ultimately be aligned with a Pallava model.189 Taken together, these examples reflect the considerable Southern Indian influence in the region. 188 Yang 2018: 41. 189 On South-East Asian stylistic connections of this icon, see Guy 1995: 86. 103 Fig. 39 Left: Buddha, stone, 7th C., Pallava period, 240 cm, excavated on the territory of Kamakshidevi Temple, Kancipuram, Tamil Nadu, Government Museum, Chennai, Tamil Nadu, India (Photograph courtesy Yuko Fukuroi). Right: Buddha, Detail of Fig.38. 3.3 Nara Embroidery and Hōryūji murals An embroidery in Fig. 38 belongs to the Nara National Museum in Japan. It has traditionally been dated to the early 8th century and is believed to have been produced in China, likely in one of the Tang capitals.190 The work depicts a heavenly assembly with the Buddha attended by a retinue of bodhisattvas, monks, and lay people. I regard images of the Buddha with two leading bodhisattvas as heavily influenced by Southern Indian and Sri Lankan prototypes. 190 Wong 2018: 153-154. 104 Fig. 40 Left: Bodhisattva, copper alloy, 8th-9th C., H 17.9 cm, Sri Lanka, Los Angeles County Museum of Art, USA (https://collections.lacma.org/node/247426). Fig. 41 Middle: Bodhisattva, detail of fig. 38. Fig. 42 Right: Bodhisattva, copper alloy, 8th C., Sri Lanka, National museum, Colombo, Sri Lanka (Boisselier 1979: fig. 52). The Buddha exhibits the same iconography as that of the Longmen sculptures (Fig. 33). At the same time, his bulky proportions and a robe drape in the Amaravati style are strikingly similar to his counterpart in a circa 7th-century Pallava sculpture in Fig. 39.191 The folds of the robes in both figures are arranged in a similar and distinctive fashion. They are concentrated along the upper edge and below the waist, creating a flat area on the stomach. The two chief bodhisattvas with their wide shoulders and almost naked torsos stand out visually from the group (see a close-up in Fig. 41). They represent a particular type of 191 On this sculpture, see Rao 1915; Aiyappan and Srinivasan 2000: 70-72. 105 Fig. 43 Buddha Amida Pure Land, wall painting, panel 6, West wall, Kondo, Hōryūji, 8th C., Ikaruga, Nara Prefecture, Japan. Not extant anymore (Wong 2008: plate 12). bodhisattva which was popular in Pallava and Sri Lankan art (Fig. 40 & 42).192 Bodhisattvas of this type are characterized by an upright posture, and by the abhaya or vitarka mudrā (right hand) and kaṭaka mudrā (left hand), which are performed distinctively at elbow-level. These figures can be bejeweled or unadorned, but always wear a distinctive flat and wide sash, which 192 A number of metal sculptures representing bodhisattvas of this type were recovered in Indonesia and Thailand and it is assumed that they were imported from the Pallava Kingdom or Sri Lanka. See, Guy 2014: cat. 6, a Pallava example is cat. 7. 106 probably represents the skin of a black deer (kṛṣṇājina) worn in the manner of the sacred thread (yajñopavīta).193 It is loosely wrapped across the torso forming a loop on the left shoulder and drapes down to the right hip. The bodhisattvas in the embroidery follow this model closely. As has been recently shown by Osmund Bopearachchi, this model likely represents bodhisattva Avalokiteśvara as a protector of mariners.194 Such images were often found along river and sea trade routes that points to maritime transmission of this iconographic form to China. It has long been noted that the two bodhisattvas in this piece of embroidery are visually close to the bodhisattvas in the murals that used to adorn the main hall of Hōryūji Temple in Japan (Fig. 43).195 These murals depicted heavenly assemblies of four buddhas and eight unidentified bodhisattvas. It is commonly thought that the murals date to the early 8th century and represent the metropolitan Tang style that has largely been lost in China. Moreover, it is thought that the figures in the murals were most likely executed with the help of full-sized stencils, which are believed to have been brought to Japan from one of the Tang capitals.196 Fig. 44 compares the bodhisattva images from Yulin Cave 25, the embroidery, and the Hōryūjimurals. Other than certain variations in decorations, the figures look strikingly similar in their linear structures, proportions, postures, gestures, and facial features. What unite them the most are their elongated and almost naked torsos, remarkably broad shoulders, and contrasting narrow waists. Though located thousands of miles apart, the Hōr yūji and Yulin bodhisattvas, with their almost identical swagged necklaces of the Southern Indian type, and similarly drawn sash and kṛṣṇājina, look as if they are from the same painting. Such similitude suggests that the images were based on the standardized models. As was already mentioned, many of these yang — standardized models and templates for the replication of images — recorded in official inventories of objects transferred from Tang China to Japan have been described by Hsueh-man Shen.197 The Hōryūji example also shows that artistic models were transmitted from China to Japan even with the help of full-sized stencils. As mentioned, significant artistic interexchange existed between Dunhuang and the Tang capitals as well.198 In my view, taken as a group the imagery from Yulin Cave 25, the Nara embroidery, and the Hōryūji murals confirm the existence of the distinctive tradition in Tang Buddhist art, 193 It was attributed as such by Dohanian. See Dohanian 1977: 41. 194 Bopearachchi 2014: 161-187. 195 On the Hōryūji and its murals, see Wong and Field 2008. 196 Wong and Field 2008: 147-152. 197 Shen 2019: 149-155. 198 Fraser 2004: 52-53. 107 108 Fig. 44 Left: Bodhisattva, detail of fig. 43; Middle: Bodhisattva, detail of fig.1(https://www.e-dunhuang.com/cave/10.0001/0001.0002.0025); Right: Bodhisattva, detail of fig. 38. which was most likely established by Vajrabodhi as a painter.199 It should be mentioned that the Southern Indian influence on the Hōryūji murals has previously been suggested by at least two other scholars. Jeannine Auboyer noted similarities in clothing between the Hōryūji figures and Ellora statues, while Marilyn Rhie argued that the facial features of the bodhisattvas reflect the influence of Southern Indian art, an influence she saw as resulting from the maritime exchange of Tang China with Southern Indian polities.200 Part 4. From the Hexi Corridor to Central Tibet Tibetan rule in the Hexi Corridor lasted roughly from the mid-8th till the mid-9th century and coincided with a period of the most intense promotion of Buddhism in the Tibetan Empire. As was discussed in Parts 1 and 2, the region was one of the major centres for the teaching and practice associated with the Southern Indian tradition of the Yoga, which became dominant in the official Buddhist ceremony and doctrine of both Tang China and Tibetan Empire in the latter half of the 8th century. The Tibetan religious tradition ascribes introduction of the Yoga teachings in Tibet to Northern Indian master Buddhaguhya, who was supposedly active in the 8th century. However, his identity and life story cannot be historically verified and modern scholarship cast doubt on his influence in Tibet.201 On the other hand, the Testament of Ba (Dba' bzhed / Sba bzhed), a historical account on the origin of Buddhism in Tibet that was probably compiled in the 11th or 12th century, attests to significant activities of Chinese monks from Hexi in Central Tibet in the second half of the 8th century. It says, for example, that the Buddhist master Moheyan from Dunhuang was patronized by the Tibetan court and had the majority of followers in Tibet.202 According to the Testament of Ba, tensions developed between the Chinese and newly arrived Indian teachers, who allegedly taught different paths towards enlightenment. The Testament of Ba further states that these tensions were resolved in a formal debate, where the Chinese side was defeated and expelled from Tibet. As was discussed by Sam van Schaik, this episode has no documentary value and was likely designed to validate new Indian lineages, 199 Remarkable similarities between the images of bodhisattvas from Yulin Cave 25, the Nara embroidery, and the Hōryūji murals would also seem to suggest that they were created in a relatively close temporal proximity. Out of the three, only the Yulin image is securely datable. It was created between 776 and 851, but most likely in the late 8th century. In contrast, the Hōryūji murals and the Nara embroidery are traditionally dated to the beginning of the same century that, in the light of the presented analysis, should perhaps be reconsidered. 200 Auboyer 1941: 71; Rhie 1988: 27-28. 201 Weinberger 2003: 30, 83-90; Nagasawa: 2017. 202 Wangdu and Diemberger 2000: 76 n 287; van Schaik 2015: 14-17. 109 which started arriving in Tibet in the 11th century, as the only genuine source of Buddhism.203 It has been traditionally understood that the Chinese party represented the Chan tradition in its modern understanding, which emphasized a concept free meditation and the immediate presence of enlightenment in the mind as opposed to the scholastic Indian tradition advocating the graduated path and the multitude of methods including Esoteric teachings. In the light of recent scholarship, however, this picture looks not only simplified, but also anachronistic. As was argued by Robert Sharf, the very notion that, Chan and Esoteric teachings constituted independent traditions in China in the 8th century can be misleading. Moreover, these teachings had a few areas of congruence, and were likely practiced in the same Buddhist circles.204 With regard to Hexi, Sam van Schaik has demonstrated that monks in Dunhuang in the 8th and 9th centuries, who were identified as members of Chan lineages in the later texts, practiced various Esoteric rituals, including those associated with the Yoga tradition in general and the Vajradhātu maṇḍala in particular.205 Furthermore, several Chinese and Tibetan sources suggest that Moheyan himself was proficient in Esoteric rituals as well.206 These traits find striking parallels in the teachings of the Nyigmapa (Rnying ma pa), the oldest school of Tibetan Buddhism, which traces its origin back to the Imperial period. According to David Snellgrove, there was a considerable vogue of Yoga Tantras in Imperial Tibet and many Mahāyoga Tantras of the Nyigmapa, which were largely rejected by the later Tibetan schools "as unorthodox and probably Chinese," conform to the content and style of "orthodox Yoga Tantras" such as the STTS.207 In addition, the Nyigmapa tradition of Dzogchen (Rdzogs chen, the Great Perfection), which includes elements of both the Mahāyoga and Chan teachings, was viewed by some 13th and 14th century authoritative Tibetan scholars as the Chinese doctrine of Moheyan.208 Altogether this would seem to suggest that Buddhist masters from Hexi, who have been perhaps misleadingly labelled as Chan monks in modern scholarship,209 were instrumental in the transmission of the Yoga tradition and possibly many other "not Chan" teachings to the Tibetans in the 8th and 9th centuries. This can be also well supported by an example of Ye shes dbyangs, a Tibetan monk, who has been identified as a member of the Chan lineage in the later Dunhuang texts and, who propagated the cult of VAEGB as part of Amoghavajra's practice of the Bhadracarī (discussed in section 2.1.2.3). According to Weinberger and Dalton, it is nearly certain that the STTS, the root Tantra of 203 van Schaik 2015: 15. 204 Sharf 2017: 53. 205 van Schaik 2015: 168. 206 van Schaik 2015: 132-133. 207 Snellgrove 1987: 457-460. 208 S. Karmay 2007: 11, 140-142. 209 For example, see Kapstein 2000: 43. 110 the Yoga tradition, was transmitted to Tibet and probably even translated into Tibetan during the Imperial period, though no details of the translation are known and the text is missing from the Imperial catalogues.210 In contrast, the Chinese translations are much better documented and historically attested, and it is not impossible that the early Tibetan translation could have been based on the Chinese texts. The earliest work in Chinese summarizing the principles of the STTS dates to 723 and is authored by Vajrabodhi. It was followed by Amoghavajra's translation of the first part of the STTS produced in Liangzhou around 755.211 As a matter of fact, the Chinese legacy in Tibetan Buddhism was deliberately minimized by later Tibetan historiographies and by a practice of editing the colophons and texts of Tibetan translations from non-Sanskrit sources in order to connect them directly to the Indian tradition. This practice started in as early as the 9th century.212 As James Gentry put it, "the general unreliability of translation colophons and centuries of scribal and editorial activity, which often included efforts during the Imperial period and thereafter to standardize Tibetan scriptural translations according to an imperially-decreed common lexicon, have obscured the process by which these collections and their translations came into being and were transmitted. As a result, the role of Chinese and other language translations of Indic texts in Tibetan translation activity and the formation of Tibetan scriptural language remains little understood."213 Regarding the Tibetan lineages of transmission, it is clear that for such important religious traditions like the Yoga, they were of great political consequence, and, therefore, there is no a priori reason to assume that they are much more reliable than the colophons. 4.1 The Main Temple of Samye As was noted by Steven Weinberger, one of the strongest pieces of evidence for the importance of the Yoga tradition in early Tibetan Buddhism is the first Tibetan monastery of Samye.214 The iconographic program of its main temple, which was seriously damaged during the Cultural Revolution, can be reconstructed from various historical sources including the Testament of Ba and a Tibetan Chronicle the Mirror Illuminating the Royal Genealogies (Rgyal rabs gsal ba'i me long).215 In the following, I will try to make sense of it with the help of Kimiaki Tanaka's 210 Weinberger 2010: 146-150; Dalton 2017: 329-330. 211 Weinberger 2003: 9; Yang 2018: 41. 212 Issues concerning the Chinese legacy in Tibetan Buddhism were discussed in Li 2016: 95. Specifically on doubtful colophons in tantric material, see van der Kuijp 1992, Almogi 2008. 213 Gentry 2019. 214 Weinberger 2010: 140. 215 English translations of these sources can be found in Wangdu and Diemberger 2000 and Sørensen1994. 111 recent interpretation.216 The main temple had three floors and a Chinese-type roof.217 On the first floor, the central icon was Buddha Śākyamuni, which was surrounded by sculptures of the eight great bodhisattvas, two additional bodhisattvas, Acala and Trailokyavijaya. As Tanaka remarked, the combination of the Buddha/Vairocana with the eight great bodhisattvas, Acala, and Trailokyavijaya corresponds to a mandala introduced by Amoghavajra in a ritual manual for recitation of the Uṣṇīṣavijayā dhāraṇī (T972).218 Amoghavajra was the most enthusiastic propagator of this dhāraṇī; he elevated its status from being a scripture with salvific and protective properties for individuals, to a scripture of the national defence and protection.219 Bearing in mind the nationwide significance of Samye, the emphasis on apotropaic qualities of Buddhist teachings in its iconographic program would have made perfect sense. It is not unlikely that the enormous power ascribed to the Uṣṇīṣavijayā dhāraṇī in Tang China led to the appropriation of the related rituals in Tibet. A textual fragment in the Tibetan language found in Dunhuang, which describes a ritual involving Vairocana, the eight great bodhisattvas, Acala, and Trailokyavijaya, could be regarded as possible evidence for such appropriation.220 The second floor was centred on Vairocana of an unknown affiliation flanked by the eight great bodhisattvas and two additional bodhisattvas, the same as those on the first floor. Buddhas of the past, present and future; Medicine Buddha; and Buddha Amitābha were installed in front of the Vairocana and bodhisattvas. A pair of Chinese gatekeepers holding a vajra, known as King and Kang, was also included in the composition. It is tempting to suggest that the sculptures of Vairocana and the eight great bodhisattvas on this floor might have represented the practice of the Bhadracarī, which, as we discussed, was an extremely popular aspirational prayer in Imperial Tibet and was believed to prolong the tsenpo's life (see section 2.1.2.3). In this regard, it is worth mentioning that other depictions related to the Avataṁsaka Sūtra subjects were present at Samye as well. While Sudhana's spiritual journey from the Gaṇḍavyūha Sūtra was illustrated in a wall frieze in the great courtyard, the Daśabhūmika Sūtra, which describes the ten stages of bodhisattvahood, was illustrated in murals of the main temple's third floor.221 The only known illustration for the Daśabhūmika Sūtra comes from Dunhuang (Fig. 30) and a similar composition could have been depicted at Samye (for a discussion on the 216 Tanaka 2018: 83-85. 217 Sørensen1994: 376-381. 218 See footnote 88. 219 Lin 214: 139-140. For example, Emperor Daizong, under the thread from advancing Uighurs and Tibetans in 765, ordered Amoghavajra to install a group of monks at court to chant the dhāraṇī for the sake of the nation. In 776, Daizong ordered all nuns and monks of the country to chant the dhāraṇī, see Yang 2018: 159. 220 The text title is Rnam par snang mdzad 'khor dang bcas pa la bstod pa. It is found in P.T. 7a and in IOL TIB J 366/3. See Tanaka 2018: 81; Tanaka 2000; Heller 1997; Dalton and van Schaik 2006: 97-98. 221 van Schaik and Doney 2007: 187-188; Sørensen1994: 379. 112 Dunhuang painting (fig.30), see section 2.6; for a discussion on depictions of Vairocana and the eight great bodhisattvas in the Avataṁsaka context, see section 2.1). A principal image of the third floor represented Vairocana with four heads (facing four directions), surrounded by eight unidentified bodhisattvas, two for each head, and many other deities. This group has traditionally been understood as representing the Sarvavid-Vairocana maṇḍala of the Sarvadurgatipariśodhana Tantra (hereafter the SDP).222 However, as Kimiaki Tanaka has pointed out, Vairocana with four heads could just as well stand for the Vajradhātu- Vairocana, the central deity of the STTS.223 Tanaka further notes that images of the four Buddhas of the Vajradhātu maṇḍala depicted back-to-back were installed at Samye during the modern reconstruction.224 Moreover, a Tibetan historical source Lde'u Chos 'byung explicitly states that all images at the third floor were erected as a depiction of the god-assembly of the Vajradhātu.225 The SDP was specifically associated with mortuary rites and thus the alleged installation of the Sarvavid-Vairocana maṇḍala as the central composition of the principal temple requires further explanation. In contrast, a version of the Vajradhātu maṇḍala fits perfectly with the official status of Samye. As discussed, this maṇḍala was the cornerstone of the Yoga tradition and a central element of state protection rituals in contemporaneous China.226 In addition, the main temple at Samye can be compared to the Golden Pavilion of Jin'ge monastery ( 金阁寺 ) constructed in 767 on Mount Wutai.227 The construction of the monastery was initiated by Amoghavajra, who started promoting the cult of Mount Wutai towards the end of his life. Built with lavish imperial sponsorship, Jin'ge became one of the principal monasteries responsible to carry out apotropaic rituals for sake of nation. Similarly to Samye, the Golden Pavilion, a main temple of the monastery, had a three-storey design with a vertically oriented iconographic program centred around Vairocana, which, as was suggested by Wei- 222 Weinberger 2010: 140. The SDP is regarded as an explanatory Tantra on part two of the STTS, see Tanaka 2018: 321. It is centred on Buddha's Uṣṇīṣa deities and mirrors the narrative found in various versions of the Sūtra of the Revered and Victorious Dhāraṇī of the Buddha's Uṣṇīṣa (Uṣṇīṣavijayā dhāraṇī). Both the SDP and the Uṣṇīṣavijayā dhāraṇī were specifically employed for mortuary rituals, see Wang 2018: 25, 42-50. Weinberger 2010: 146. The history of the early translation of the SDP into Tibetan is far from clear. The text is listed in the Imperial catalogue and the colophon, which dates to the 8th century, gives the names of Indian and Tibetan translators, see Weinberger 2010: 148-149. However, the colophon is proven to be unreliable and the identity of the translators has been questioned, see van der Kuijp 1992. 223 Vajradhātu-Vairocana can be depicted with four heads as well, but in contrast with like Sarvavid-Vairocana, who performs the dhyāna mudrā, Vajradhātu-Vairocana performs the bodhyaṅgī mudrā. See Tanaka 2018: 83; On four- faced Vairocana in the STTS, see Weinberger 2003: 187. 224 Tanaka 2018: 84. 225 Sørensen 1994: 379 n 1260. 226 There is strong evidence showing that the Vajradhātu maṇḍala was known in Tibet in the 8th century. See Weinberger 2010: 144. 227 On the Golden Pavilion and Jin'ge monastery, see Lin 2014: 131-154. 113 Cheng Lin, was likely conceived to represent a three-dimensional version of the Vajradhātu maṇḍala whose five-fold system correlated with the five monasteries and five peaks of Mount Wutai.228 The Tibetans must have been aware of these developments in China as the popularity of Wutai Shan in Tibetan Empire is well attested in historical sources and according to the Old Dynastic History of the Tang, the Tibetan tsenpo once even requested a picture of Mount Wutai from the Tang court.229 Remarkably, the first representations of Mount Wutai appeared in Dunhuag art during the Tibetan period.230 The discussed parallels are striking and suggest that the decision to built Samye might have been motivated by apotropaic qualities of certain Buddhist teachings. I have later come across a 1983 article by Christopher Beckwith, where he put forward a theory suggesting that the Tibetan adoption of Buddhism as a state religion in general and the building of Samye in particular were inspired by the Chinese model of National Protection Buddhism maintained by Amoghavajra, and were, at least partially, dictated by a need to resolve the crisis of legitimacy faced by the Tibetan emperor Trisong Detsen.231 I was delighted to learn that I came to the same conclusion independently, but from a different angle. Regrettably, no original images survived from Samye, as they could provide art historical evidence for sources of their corresponding teachings. Nevertheless, I argue that Samye’s aesthetics can be reconstructed from artwork created at the earliest stage of the Buddhist revival in Tibet in the 11th century. 11th century Clay Sculptures from Central Tibetan Temples After the collapse of the empire in the mid 9th century, Tibet lost its colonies and went to a so- called period of fragmentation. The monastic tradition perished to be reintroduced only in the late 10th century with a religious movement known in Tibetan history as "the later spread of the Dharma" (phyi dar). Although most of Hexi was recaptured by the Tang forces, its capital and a former Amoghavajra's seat, Liangzhou, remained under local Tibetan rule up to the 11th century, when it was absorbed into the Xi Xia Empire.232 The region undoubtedly continued to hold a position of Buddhist authority in Central Tibet, as in the late 10th century a group of men from there were sent to the areas of Liangzhou and its neighbor, the Tibetan kingdom of Tsongka, in order to bring the monastic tradition back to Central Tibet.233 Being ordained in these areas, the new monks, sometimes labeled as the Eastern Vinaya monks, returned to 228 Lin 2014: 149. 229 Debreczeny 2011: 8-9; Wong 1993: 39. 230 Wong 1993: 41. 231 Beckwith 1983. On the notion of National Protection Buddhism in Tang China, see Orzech 1998. In particular, on the term "National Protection Buddhism", see Orzech 1998: 3 n. 9. 232 van Schaik and Galambos 2012: 64-65. 233 van Schaik and Galambos 2012: 68-74. 114 Central Tibet and initiated the formation of Buddhist communities, the reconstruction of old temples, and the building of new ones. Although Samye and other royal temples were not in use and in a poor state, they still contained sculptures and paintings.234 In addition, a network of small temples continued to function under the patronage of local princes. According to Tibetan historical sources, hundreds of Buddhist establishments were renovated and built in Central Tibet during the 11th century. Renovation of old temples, symbols of Tibet's ancient spirituality and the Imperial past, was a top priority during this period235 and it can, therefore, be surmised that survived specimens of Imperial art were renovated and replicated in the new temples. For example, Samye monastery was not only renovated, but also was taken as a model for Grwa thang monastery (founded in 1081).236 Thus, it is the artistic styles of the late Imperial period that would have been prevalent in Buddhist art of central Tibet in the late 10th and 11th century. In this regard, I would like to draw attention to a group of clay sculptures from the Central Tibetan temples/monasteries of G.ye dmar, Rtis gnas gsar, Zhwa lu Gser khang, Rkyang bu, Grwa thang, Zho nang, and Snye thang.237 Although most of them were lost during the Cultural Revolution, their images can still be seen in pre-1950 photographs taken by Li Gotami Govinda and Giuseppe Tucci's expedition.238 Luckily, the sculptures of Snye thang monastery have survived.239 These temples were founded or renovated at the earliest stage of the Buddhist revival in the 11th century. Despite the fact that some of these temples are located at a considerable distance from the others, their sculptures demonstrate remarkable stylistic consistency. This, I believe, confirms that the sculptures reflect Buddhist aesthetics of the Imperial period, which must have been a common source of inspiration for artists throughout Central Tibet in the 10th and 11th centuries.240 With regard to the iconographic programs, much of the subject matter depicted in the temples was popular during the Imperial period as well. For example, four- faced Vairocana with retinue and Vairocana grouped with eight bodhisattvas can be found in the iconographic programs of Samye and the temples under discussion. Most remarkably, the 234 Davidson 2005: 94-95. 235 On the activities of the eastern Vinaya monks and the general situation in Central Tibet at the time, see Davidson 2005: 84-115. 236 Vitali 1990: 49; Heller 2002: 40-41; Stoddard 2018: 96-97. 237 The temples were discussed in Vitali 1990: 37-69; Lo Bue: 2000; von Schroeder 2001: 836-851. 238 Many photographs are reproduced in publications listed in the previous footnote. Also, see Govinda 1979: 40-50. 239 For description and images of the sculptures, see von Schroeder 2001: 860-869. Although von Schroeder dates the sculptures to circa 1200, I believe they belong to the 11th century. 240 Along the same lines, Heather Stoddard suggested that surviving murals at Grwa thang reflect the aesthetics employed at Samye in the late 8th century. Stoddard 2018: 96-97. In addition to the imagery in the early style, the murals also include some depictions in the Pala style of northeastern India, which was introduced to Tibet in the 11th century. On the history and paintings of Grwa thang, see Heller 2002. 115 Fig. 45 Vairocana               Fig. 46  Amoghasiddhi Part of the sculptural Vajradhātu maṇḍala, clay sculptures, 11th C., Rkyang bu temple, Central Tibet. Not extant anymore (von Schroeder 2001: fig. XIII-17, 19). sculptures, which depict these groupings, show the same distinctive elements of the Southern Indian artistic mode as they are found in the imagery from Mogao and Yulin. For example, Rkyang bu temple was established shortly before 1037241 and contained a sculptural representation of the Vajradhātu maṇḍala with four-faced Vairocana at its centre (Fig. 45).242 Vairocana in this form was likely installed on the third floor of the main temple at Samye. Another four-faced Vairocana is recorded in a temple with a Chinese roof (rgya phibs) at Rtis gnas gsar, which was founded during the Imperial period and renovated in the 11th century.243 The Vajradhātu sculptures in Rkyang bu temple are decorated with spiked armbands and swagged necklaces that are characteristic of the Southern Indian mode in Hexi art (Figs. 45, 46, 47). The sculptures' style and splendor show striking parallels with the 8th and 9th century Hexi imagery (see examples in Figs. 14, 15, 16). Their square and tiered thrones are 241 Vitali 1990: 56-59. 242 This figure was misidentified as Sarvavid-Vairocana in earlier scholarship. 243 Vitali 1990: 51; von Schroeder 2001: 840-841. 116 Fig. 47 Ratnasambhava and attending deities as a section of the sculptural Vajradhātu maṇḍala, 11th C., Rkyang bu temple, Central Tibet. Not extant anymore (von Schroeder 2001: fig. XIII-21). Fig. 48 One of sixteen bodhisattvas surrounding a figure of Vairocanābhisaṁbodhi (worshipped as Amitāyus), 11th C., G.ye dmar temple, Central Tibet (Photo: Lionel Fournier, courtesy of Amy Heller). similar to those in the Hexi depictions as well. It is important to stress that these sculptures are more closely related to the earlier Hexi examples than to contemporaneous 10th and 11th century ones (see Figs. 17, 20, 21). The contemporaneous imagery from the Hexi corridor shows a considerable decline in artistic quality that does not match the refinement and monumentality of these Tibetan sculptures. This would seem to confirm that the sculptures represent the revival of the Imperial aesthetics, which in turn were influenced by the 8th and 9th century art from the Hexi Corridor. Groups of the eight bodhisattvas are recorded in all the temples. In addition, at G.ye dmar there was a group of sixteen bodhisattvas, which probably represented the sixteen bodhisattvas of the Bhadrakalpa, a prominent group in the Yoga tradition and members of the Vajradhātu 117 Fig. 49 One of eight bodhisattvas surrounding a figure of Vairocanābhisaṁbodhi, 11th C., G.ye dmar temple, Central Tibet (Govinda 1979: 45). Fig. 50 Vairocanābhisaṁbodhi ( worshipped as Amitāyus) and one of eight attending bodhisattvas, G.ye dmar temple, Central Tibet (Govinda 1979: 45). maṇḍala.244 In all these groups, the bodhisattvas accompany crowned figures displaying the dhyāna mudrā, who have been identified either as Vairocanābhisaṁbodhi or as Amitābha/ Amitāyus.245 These deities are differentiated only by the presence/absence of a bowl/vase in their hands. It is possible that initially all of them represented Vairocanābhisaṃbodhi. The 244 On the sixteen bodhisattvas of the Bhadrakalpa, see Tanaka 2018: 169-170. 245 von Schroeder 2001: fig. xiii-8, 9, 11, 13, 34B. 118 Fig. 51 Vairocanābhisaṁbodhi (worshiped as Amitāyus) and one of eight attending bodhisattvas, clay sculptures, 11th C., H 310 cm and 330 cm, Snye Thang monastery, Central Tibet (photo: Alain Bordier, 1998; von Schroeder 2001: pics. 199A, 200F). bowls/vases could have been added later when the cult of Vairocana declined in popularity.246 As discussed, examples of bodhisattvas surrounding Vairocana displaying the dhyāna mudrā are relatively rare and, apart from Tibet, can be found in Southern India and in the Hexi corridor. The Tibetan sculptures under review depict Vairocana and bodhisattvas either in the Tibetan mode (clad in Tibetan-style robes) (Figs. 48, 49, 50), or the Indian mode (Fig. 51). These two modes of representation can already be found in the early 9th century rock carvings in Eastern Tibet (Figs. 52, 53, 54) discussed in section 2.1.2.3. Significantly, the sculptures in both Tibetan and Indian modes are adorned with spiked armbands that is a clear indication of the influence from the Hexi corridor. In Figs. 48, 49, and 50, I have highlighted the armbands 246 The same conversion is evident in the early 9th-century rock carving at 'Bis khog in Eastern Tibet, where Vairocana (identified by inscription) holds a bowl in his hands (see Fig. 53) . Another example of such a conversion can be seen in the main temple of Tabo. See Klimburg-Salter and Luczanits 1998: 101, 143-146, fig.146. 119 Fig. 52 Sketch of Renda high relief rock carving, Fig. 53 Sketch of 'Bis khog high relief rock carving (also dated 804/816, Bragyab county, Chamdo known as Bimda), dated 806, Yulshul, Qinghai, prefecture of TAR, China (Shanxi Province China (Huo 2017: 8). Academy of Archaeology and TAR Relics Preservation Institute: 2014: 9). Fig. 54 Sketch of Sgar thog high relief rock carving, early 9th C. near the town of Sgar thog in Smar khams county, Chamdo prefecture of TAR, China (Yang, Lu and Zhang 2017: 236). 120 with blue arrows; in Fig. 51, some remaining spikes are clearly visible in the armbands of the bodhisattva. Despite substantial renovation, the figures of Vairocana and the bodhisattva in Fig. 51 still convey the same splendid aesthetics that we see in the late 8th century images from Yulin Cave 25 (in particular compare Vairocana in Fig. 51 with Vairocana in Fig. 6). As it was already stated, since no direct contacts between Tibetan Empire and Southern Indian polities are known, the source of the Southern Indian features in Tibetan art can only be Hexi/China. Also, it is very important to note that not all of the discussed sculptures display these elements. They are mostly found in sculptures which represent subjects of the Yoga tradition and groups of Vairocanābhisaṁbodhi with bodhisattvas.247 The same correlation is evident in Hexi art. As discussed, imagery played a pivotal role in the transmission of Esoteric teachings, hence it can be deduced that these parallels between the Tibetan and Hexi depictions reflect not only the artistic transmission, but also the transmission of the Yoga teachings as well as of the cult of VAEGB from the Hexi corridor to Tibet. Although the Chinese textual antecedents and visual evidence from Hexi provide solid grounds for this conclusion, much research needs to be done in the area of Tibetan textual scholarship. Such research could potentially uncover additional textual evidence as well as possible traces of Chinese lexical influence in the relevant Tibetan texts. From a larger perspective, these stylistic and iconographic traits from Hexi detected in the early Tibetan imagery provide a strong support to the theory of the apotropaic function of Samye and, by extension, to the theory of the Chinese background of the Tibetan adoption of Buddhism. Finally, in parallel with the Southern Indian mode in Hexi art, the artistic mode represented by these Tibetan sculptures gradually vanished in Tibet after the 11th century. Its disappearance likely reflects a decline of interest for the older teachings and their imagery caused by the arrival of new Tantric teachings and their imagery in the Pāla style from northeastern India.248 Conclusion In tracing the artistic background and doctrinal affiliations of a group of images from Mogao and Yulin this research elucidates the intricate web of politics, religion and art that connected the Pallava Kingdom of Sothern India, Tang China and Imperial Tibet. Addressing a period of particularly intense Buddhist interexchange between South and East Asia, this article 247 For depictions, which do not display the element of the Southern Indian mode, see sculptures at G.ye dmar in von Schroeder 2001: fig.xiii-10; and murals at Grwa thang in Vitali 1990: Plates 29-34. 248 In the forthcoming article, I argue that in this process the Hexi corridor played a crucial role as well, and these northeastern Indian teachings and Pala aesthetics were initially introduced to Tibet from the Hexi corridor. See footnote 174 and Khokhlov: forthcoming. 121 contributes to reshaping our understanding of several key issues. Firstly, it highlights considerable Southern Indian traits in Buddhist art of Tang China and identifies a distinctive artistic tradition, which was most likely established in China by Vajrabodhi, a renowned Buddhist master as well as a painter from the Pallava Kingdom of Southern India. Specific stylistic and iconographic features found in the Dunhuang imagery can be confidently traced to the art of the Pallava kingdom in Tamil Nadu and the Pallava influenced art in the Western Deccan. The subjects of these Dunhuang images belong to the Yoga tradition and the cult of the eight great bodhisattvas, which were propagated in China by Vajrabodhi and Amoghavajra. On the other hand, the article helps in reconstructing artistic features of Buddhist art of Tamil Nadu that was largely lost in India. In addition, the Pallava influenced styles of the discussed Chinese and Tibetan images constitute material evidence for the popularity of the Yoga tradition and the cult of the eight great bodhisattvas in Tamil Nadu in the 8th century. Lastly, but not least, this research defies the idea of "early Tibetan Buddhist art" as represented by the artwork created in the Hexi Corridor during and after the period of Tibetan rule. Moreover, it shows quite the opposite and suggests that the conquest of the Hexi corridor enabled the Tibetan appropriation of Chinese religious and artistic trends, which were popular in the region in the 8th century. Furthermore, the artistic and iconographical considerations presented in the article support the theory that the Tibetan adoption of Buddhism as a state religion was based on the Chinese model of National Protection Buddhism established by Amo- ghavajra. Appendix 1 A list of paintings from Mogao and Yulin that display the Southern Indian artistic mode discussed in this article. This list is not exhaustive and contains only the most important published examples. 1) A wall painting of Kubera with two attendants in Yulin Cave 15. The cave is traditionally dated to the Tibetan period. Discussed in section 2.4, Fig. 22. 2) A wall painting of Vairocanābhisaṁbodhi and the eight great bodhisattvas in Yulin Cave 25. The deities are identified by accompanying Chinese inscriptions, Tibetan period, late 8th century. Discussed in section 2.1, Fig. 1. 3) A portable painting of the Amoghapāśa maṇḍala from Mogao Cave 17, 8th/9th century, Musée Guimet, Paris, MG 26466. Discussed in section 2.3, Fig.18. 4) A portable painting identified as the Pure Land of Bhaiṣajyaguru from Mogao cave 17, 122  dated to 836, British Museum. Ch.xxxvii.004. Discussed in section 2.6, Fig.29. 5) Two wall paintings d epicting maṇḍala-like compositions centred around Vairocanābhisambodhi and Vajrasattva, Mogao Cave 14, second half of the 9th century. See section 2.2, Fig. 14 & 15. 6) A portable painting depicting a heavenly assembly from Mogao Cave 17, 9th century, British Museum, Ch.ivi.0034. 7) A portable painting depicting Mañjuśrī riding a lion from Mogao Cave 17. 9th century, British Museum, Ch.XXVI. a. 007. 8) Four large images of seated Avalokiteśvara with four Offering Goddesses, painted on four slopes of the ceiling in Mogao Cave 161, 9th/10th century. Discussed in Wang 2018: 145- 150, Fig. 63-66. 9) Two wall paintings d epicting maṇḍala-like compositions centred around Vairocanābhisambodhi and a buddha, Yulin Cave 20, 9th century. Discussed in section 2.2, Fig.16. 10) A portable painting depicting a standing figure of Samantabhadra, Mogao Cave 17, 9th/10th century, Musée Guimet, MG. 17770. Discussed in section 2.5, Fig. 25. 11) A portable painting depicting a standing bodhisattva, Mogao Cave 17, 9th/10th century, British Museum, Ch. Iv. 007. Discussed in section 2.5, Fig. 26. 12) A portable painting depicting a standing bodhisattva, Mogao Cave 17, 9th/10th century, British Museum, Ch.Iv.0031. 13) A portable painting depicting a standing bodhisattva, Mogao Cave 17, 9th/10th century, British Museum, Ch.Iv.004. 14) A portable painting depicting a seated figure of Avalokiteśvara, Mogao Cave 17, dated by the museum to the 9th century, but likely belongs to the 10th century, British Museum, Ch.00401. 15) A portable painting depicting seated bodhisattva, Mogao Cave 17, dated by the museum to the 9th century, but likely belongs to the 10th century, British Museum, Ch.00377. 16) A portable depicting a maṇḍala of Avalokiteśvara, Mogao Cave 17, dated by the museum to the 8th century, but likely belongs to the 10th century, British Museum. Ch.xxii.0017. 17) A portable depicting a maṇḍala of Avalokiteśvara, Mogao Cave 17, dated by the museum to the 8th-9th century, but likely belongs to the 10th century. British Museum. Ch.Iv.0024. 18) A portable painting of Vairocanābhisaṁbodhi and the eight great bodhisattvas (the deities are partially identified by Tibetan inscriptions), Mogao Cave 17, late 9th or early 10th century, British Museum, Ch.0074. Discussed in section 2.1, Fig. 13. 19) A portable painting of the Amoghapāśa maṇḍala, Mogao Cave 17, 10th century, Musée Guimet, EO.3579. Discussed in section 2.3, Fig. 20. 20) A portable painting of Amoghapāśa maṇḍala, Mogao Cave 17, dated by the museum to the 8th century, but likely belongs to the 10th century. Musée Guimet. EO.1131. Discussed 123  in section 2.3, Fig. 21. 21) A portable painting of the ten stages of bodhisattvahood, Mogao Cave 17, 10th century, Musée Guimet, MG. 26465. Discussed in section 2.6, Fig.30. 22) Two wall paintings d epicting maṇḍala-like compositions centred around Vairocanābhisaṃbodhi and a buddha, Yulin Cave 38, 10th/11th century. Discussed in section 2.2, Fig. 17. 23) Wall paintings inside a Heavenly King stūpa ( 天王堂 ) at Mogao, 10th century. The content was only partially published. Six multi-armed deities, each surrounded by four Offering Goddesses, are painted on the ceiling; eight or more standing bodhisattvas and other deities are painted on the walls. Published in the Dunhuang Research Academy's 2003 The Complete Collection of Dunhuang Grottoes, Volume 10, pp. 197-212. Bibliography Acri, Andrea. 2016. Esoteric Buddhism in Mediaeval Maritime Asia: Networks of Masters, Texts, Icons. Singapore: I. SEAS-Yusof Ishak Institute. —— 2018. "Maritime Buddhism." In Oxford Research Encyclopedia of Religion. Oxford University Press. http://oxfordre.com/religion/view/10.1093/ acrefore/9780199340378.001.0001/acrefore-9780199340378-e-638, accessed 14.11.19. Aiyappan, Ayinapalli and Srinivasan, P.R. 2000. Story of Buddhism with Special Reference to South India. Chennai: Commissioner of Museums, Govt. Museum. Almogi, Orna. 2008. "How Authentic Are Titles and Colophons of Tantric Works in the Tibetan Canon? The Case of Three Works and Thier Authors and Translators." In Beitrage Zur Zentralasienforschung 14: 87-124. Asha Rani, Mathur. 1988. The Great Traditions: Indian Bronze Masterpieces. New Delhi: Festival of India. Auboyer, Jeannine. 1941. Les Influences et les Réminiscences Étangères au Kondō Du Hōryūji, Dessins de l'Auteur. Paris: P. Geuthner. Barber, A. W. 2008. "Two Mahāyāna Developments Along the Krishna River." In Buddhism in the Krishna River Valley of Andhra, eds. Sree Padma (Holt) and A.W Barber. Albany: State University of New York Press, 151-167. Bautze-Picron, Claudine. 1997. "Le groupe des huits grandes bodhisatva en Inde: genèse et développement." In Living a Life in Accord with the Dhamma: Papers in Honor of Professor Jean Boisselier on his Eightieth Birthday, ed. Natasha Eilenberg. Bangkok: Silpakorn University, 1-55. —— 2000. "Nasik: The Late Mahāyāna Caves 2, 15, 20 & 23-24." In South Asian Archaeology 1997, Proceedings of the Fourteenth International Conference of the European 124  Association of South Asian Archaeologists. Rome: Istituto Italiano per l'Africa e l'Oriente, 1201-27. —— 2008. "Antagonistes et complémentaires, le lion et l'éléphantdans la personnalité du Buddha." In Penser, dire et représenter l'animal dans le monde indien, eds. Nalini Balbir and Georges-Jean Pinault, Paris: Librairie Honoré Champion, Éditeur (Bibliothèque de l'École des Hautes Études, Sciences historiques et philologiques, Tome 345), 523-72. —— 2015. The Forgotten Place: Stone Images from Kurkihar, Bihar. New Delhi: Archaeological Survey of India. Beckwith, Christopher. 1983. "The Revolt of 755 in Tibet." In Contributions on Tibetan Language, History, and Culture. Proceedings of the Csoma de Karas Symposium held at Velm-Vienna, Austria, eds. Ernst Steinkener and Helmut Tauscher, Vol. I. Vienna: Arbeitskreis fur Tibetische und Buddhistische Studien Universitat Wien, 1-16. Boisselier, Jean. 1979. Ceylon Sri Lanka. Geneva: Nagel Publishers. Bopearachchi, Osmund. 2014. "Sri Lanka and the Maritime Trade. Bodhisattva Avalokiteśvara as the Protector of Mariners." In Asian Encounters exploring connected histories, eds. Parul Pandya Dhar and Upinder Singh. New Delhi: Oxford University Press, 161-87. Brauen, Martin. 1997. The Mandala: Sacred Circle in Tibetan Buddhism. 1st ed. Boston: Shambhala. Chandra, Lokesh. 1980. "Oḍḍiyāna: A New Interpretation." In Tibetan Studies in Honour of Hugh Richardson. eds. Aris, M. and Aung San Suu Kyi. Warminster and Oxford: Aris and Phillips, 73-78. Chou, Yi-liang. 1945. "Tantrism in China." Harvard Journal of Asiatic Studies 8.3-4: 241-332. Cleary, Thomas. F. 1993. The Flower Ornament Scripture. Boston: Shambhala. Dalton, Jacob. P. 2011. The Taming of the Demons: Violence and Liberation in Tibetan Buddhism. New Haven and London: Yale University Press. —— 2017. "On the Significance of the Arya-Tattvasamgraha-Sadhanopayika and Its Commentary." In Chinese and Tibetan Esoteric Buddhism. eds. Yael Bentor and Meir Shahar, Leiden: Brill, 321-337. Dalton, Jacob.P. and van Schaik, Sam. 2006. Tibetan Tantric Manuscripts from Dunhuang: A Descriptive Catalogue of the Stein Collection at the British Library. Leiden and Boston: Brill. Dayalan, Duraiswamy. Art, Architecture and Cultural Link Between Maharastra and South India (Andhra Pradesh & Tamil Nadu). https://www.academia.edu/25250507/ART_ ARCHITECTURE_AND_CULTURAL_LINK_BETWEEN_MAHARASHTRA_AND_ SOUTH_INDIA_ANDHRA_PRADESH_and_TAMIL_NADU. Davidson, Ronald M. 2002. Indian Esoteric Buddhism: A Social History of the Tantric Movement. New York: Columbia University Press. —— 2005. Tibetan Renaissance: Tantric Buddhism in the Rebirth of Tibetan Culture. New 125  York: Columbia University Press. de Visser, Marinus Willem. 1915. The Bodhisttva Ti-tsang (Jizō) in China and Japan, Berlin: Oesterheld & Co. Verlag (Erste Sonderveröffentlichung der Ostasiatischen Zeitschrift). Debreczeny, Karl. 2011. "Wutai Shan: Pilgrimage to Five-Peak Mountain." In Journal of the International Association of Tibetan Studies, no. 6: 1-133. —— 2019. "Faith and Empire: An Overview." In Faith and Empire: Art and Politics in Tibetan Buddhism, ed. Karl Debreczeny. New York: Rubin Museum of Art, 17-53. Dessein, Bart. 2003. "The Glow of the of the Vow of the Teacher Samantabhadra 'Puxian Pusa Xingyuan Zan' (T.297) Samantabhadracaryapranidhanaraja." Acta Orientalia Academiae Scientiarum Hungaricae 56 (2/4): 317-38. Dohanian, Diran K. 1977. The Mahāyāna Buddhist Sculpture of Ceylon. New York: Garland House. —— 1983. "Sinhalese Sculptures in the Pallava Style." Archives of Asian art 36: 6-21. Dunhuang Academy ed. 敦煌研究院编 2003. Dunhuang shiku quanji 敦煌石窟全集 / The Complete Collection of Dunhuang Grottoes, vol. 10. Hong Kong: The Commercial Press 香港 ：商务印书馆 . —— 2013. Zhongguo shiku: Dunhuang Mogaoku 4 中 国 石 窟· 敦 煌 莫 高 窟 4/ Chinese Grottoes: Dunhuang Mogao Grottoes. Beijing: Cultural Relics Publisher 北京 ：文物出 版社 . —— 2014. Yulin Ku 榆林窟 / The Yulin Grottoes. Nanjing: Jiangsu Arts Publisher 南京 ：江苏 美术出版社 . Fontein, Jan. 1968. The Pilgrimage of Sudhana. A Study of Gaṇḍavyūha Illustrations in China. Japan and Java. The Hague: Mouton. —— 2012. Entering the Dharmadhātu, A Study of the Gandavyūha Reliefs of Borobudur, Leiden/Boston: Brill (Studies in Asian art and archaeology, vol. 26). Fraser, Sarah E. 2004. Performing the Visual: The Practice of Buddhist Wall Painting in China and Central Asia, 618-960. Stanford, CA: Stanford University Press. Gentry, James. 2019. "An Inquiry into Source-Text Multilingualism in Tibetan Scriptural Translations." abstract of a talk delivered at the 15th Seminar of the International Association for Tibetan Studies in Paris, July 2019. [https://www.academia.edu/39792552/ An_Inquiry_into_Source- Text_Multilingualism_in_Tibetan_Scriptural_Translations.] Giebel, Rolf W. 1995. "The Chin-kang-ting ching yü-ch'ieh shih-pa-hui chih-kuei: An Annotated Translation." In Journal of Naritasan Institute for Buddhist Studies 18: 107- 201. —— 2001. Two Esoteric Sutras: The Adamantine Pinnacle Sutra, the Susiddhikara Sutra. Berkeley, Calif: Numata Center for Buddhist Translation and Research. Giès, Jacques. ed. 1994-96. Les Arts de l'Asie Centrale. 2 vols. London: Serindia. Goble, Geoffrey. 2012. "Chinese Esoteric Buddhism: Amoghavajra and the Ruling Elite." PhD 126  thesis. Indiana University. —— 2013. "The Legendary Siege of Anxi: Myth, History, and Truth in Chinese Buddhism." Pacific World: Journal of the Institute of Buddhist Studies 3 (15): 1-32. Govinda, Li Gotami. 1979. Tibet in Pictures: Text and Photos. Berkeley, CA: Dharma Pub. Granoff, Phyllis. 1968/69. "A Portable Buddhist Shrine from Central Asia." In Archives of Asian Art 22: 80-95. Guy, John. 1994. "The Lost Temples of Nagapattinam and Quanzhou: A Study in Sino-Indian Relations." In Silk Road Art and Archeology 3: 291-310. —— 1995. "The Avalokitesvara of Yunnan and Some South East Asian Connections." In South East Asia and China: Art, Commerce and Interaction, Percival David Foundation Colloquies on Art and Archeology of Asia, eds. Rosemary Scott and John Guy. London: University of London, 64-83. —— 2014. "Introducing Early Southeast Asia." In Lost Kingdoms: Hindu-Buddhist Sculpture of Early Southeast Asia. ed. John Guy. New York: Metropolitan Museum of Art. Heller, Amy. 1994. "Ninth Century Buddhist Images Carved at Ldan Ma Brag to Commemorate Tibeto-Chinese Negotiations." Tibetan Studies: Proceedings of the 6th Seminar of the International Association for Tibetan Studies, Fagernes, 1992. Oslo: Institute for Comparative Research in Human Culture, 335-49. —— 1997. "P.T. 7A , P.T. 108 , P.T. 240 and Beijing Bstan 'gyur 3489: Ancient Tibetan Rituals Dedicated to Vairocana." In The Pandita and the Siddha: Tibetan Studies in Honor of E. Gene Smith, ed. Gene Smith. Dharamsala: Amnye Machen Institute, 85-91. —— 2002. "The Paintings of Gra thang: History and Iconography of an 11th century Tibetan Temple." In Tibet Journal 27(1/2): 39-72. Horlemann, Bianca. 2012. "Buddhist Sites in Amdo and former Longyou from the 8th to the 13th Century." In Old Tibetan Studies Dedicated to the Memory of Ronald E. Emmerick, PIATS Oxford 2003. ed. Cristina Scherrer-Schaub. Leiden: Brill, 119-157. Huntington, John C. 1981. "Cave Six at Aurangabad: A Tantrayana Monument?" In Kalādarśana: American Studies in the Art of India, ed. J.G. Williams. Leiden: E.J. Brill, 47-55. Huo Wei 霍巍 . 2017. "Qinghai Yushu Darirulai fotang de kaogu diaocha yu xinfaxian" 青海 玉树大日如来佛堂的考古调查与新发现 [Archaeological Investigation on Parinirvana in Yushu of Qinghai and New Findings]. Qinghai Minzu Yanjiu 青海民族研究 /Qinghai Ethnic Research 28 (1): 4-9. —— 2018. "Zangdong Tubo moya zaoxiang yu Tangbo jiaoliu shiye xia de Jiannan Yizhou" 藏 东吐蕃摩崖造像与唐蕃交流视野下的剑南益州 [Rock Reliefs in East Tibet and Yizhou in the Context of Cultural Exchanges between Tang and Tubo]. Zangxue xuekan 藏学学刊 /Journal of Tibetology 16: 301-313. Indrapala, Karthigesu. 2005. The Evolution of an Ethnic Identity The Tamils in Sri Lanka c.300 127  BCE to c.1200 CE. Sydney: MV Publications for the South Asian Study Centre. Kapstein, Matthew T. 2000. The Tibetan Assimilation of Buddhism: Conversion, Contestation, and Memory. Oxford University Press. —— 2004. "The Treaty Temple of De-Ga G.Yu-Tshal: Identification and Iconography." In Xizang Kaogu Yu Yishu 西藏考古与艺术 / Tibetan Arachaeology and Art. ed. Huo Wei and Y. Li. 霍巍 , 李永宪编 Chengdu: Sichuan renmin chubanshe 成都：四川人民出版社 , 98-127. —— 2009. "The Treaty Temple of the Turquoise Grove." In Buddhism Between Tibet and China, ed. Matthew T. Kapstein. Boston: Wisdom Publications, 21-72. —— 2014. "The Treaty Temple of De ga g.yu tshal: Reconsiderations." Zangxue xuekan 藏学 学刊 /Journal of Tibetology 10: 32-34. Karmay, Heather. 1975. Early Sino-Tibetan Art. Warminster: Aris and Philips lts. Karmay, Samten Gyaltsen. 2007. The Great Perfection (rDzogs chen) A Philosophical and Meditative Teaching of Tibetan Buddhism. Leiden and Boston: Brill. Khokhlov, Yury. (forthcoming). The abstract: https://www.academia.edu/36748689/From_ India_to_the_Doorstep_of_China_Indian_Pala_Style_in_the_Hexi_Corridor, accessed 14.11.19. Klimburg-Salter, Deborah E. and Christian Luczanits. 1998. Tabo: A Lamp for the Kingdom: Early Indo-Tibetan Buddhist Art in the Western Himalaya. New York: Thames and Hudson. Lee, Sonya S. 2010. Surviving Nirvana: Death of the Buddha in Chinese Visual Culture. Hong Kong University Press. Li, Channa. 2016. "Translationship Lost in Transmission: Elusive Attributions of Two Tibetan Sūtra Translations." Revue d'Etudes Tibétaines 37: 207-30. Lin Li-Kuang. 1935. "Puṇyodaya (Na-t'i), un propagateur du tantrisme en Chine et au Cambodge à l’époque de Hiuan-tsang." Journal Asiatique 227: 83-100. Lin Wei-Cheng. 2014. Building a Sacred Mountain: The Buddhist Architecture of China's Mount Wutai. Seattle: University of Washington Press. Linrothe, Robert N. 2014. "What They Saw: Kaśmīrā in the Eyes of Western Himalayas." In Collecting Paradise: Buddhist Art of Kashmir and its Legacies, exh. cat. New York: Rubin Museum of Art, 29-107. Liu Yongzeng 刘永增 . 2008. "Mogaoku di148ku nanbeikan tianjing tuxiang jieshuo" 莫高窟 第 148 窟南北龛天井图像解说 [Explanation of the Ceiling Images in the Southern and Northern Alcoves of Mogao Cave 148]. In Dunhuang bihua yishu jicheng yu chuangxin guoji xueshu yantaohui lunwenji 敦煌壁画艺术继承与创新国际学术研讨会论文集 / The Inheritance and Innovation of Dunhuang Wall Art: Proceedings of the International Conference. ed. Duhuang Academy 敦煌研究院编 . Shanghai cishu chubanshe 上海辞书 出版社 . 518-536. 128 Lo Bue, Erberto. 2000. Tibet. Templi scomparsi fotografati da Fosco Maraini, Torino: Ananke. Long, Mark. 2009. Caṇḍi Mendut, Womb of the Tathāgata, written, compüiled, translated and illustrated, New Delhi: Aditya Prakashan.. Malandra, Gery Hockfield. 1993. Unfolding a Mandala: The Buddhist Cave Temples at Ellora. Suny Series in Buddhist Studies. Albany, N.Y.: State University of New York Press. —— 1996. "The Mandala at Ellora/Ellora in the Mandala." Journal of the International Association of Buddhist Studies 19: 181-207. McCoy, Michelle. 2016. "Maṇḍala of Avalokiteśvara of the Unfailing Lasso (Amoghapāśa)." In Cave Temples of Dunhuang: Buddhist Art of China's Silk Road. eds. Neville Agnew, Marcia Reed and Tevvy Ball. Los Angeles: J. Paul Getty Trust, 236-39. McNair, Amy. 2007. Donors of Longmen: Faith, Politics, and Patronage in Medieval Chinese Buddhist Sculpture. Honolulu: University of Hawai'i Press. Misra, R.N. 1981. Yaksha Cult and Iconography. New Delhi: Munshiram Manoharlal. Nagasava, Jake. 2017. https://treasuryoflives.org/biographies/view/Buddhaguhya/10546, accessed on 01.10.2019. Okada, Ken. 2004. "Observations on Buddhist Stone Sculptures from the Longxing Monastery." In New Perspectives on China's Past: Chinese Archaeology in the Twentieth Century. ed. Xiaoneng Yang. New Haven, CT: Yale University Press, 381-91. Orzech, Charles D. 1995. "The Legend of the Iron Stūpa." In Buddhism in Practice, ed. Donald S. Lopez, Jr. NJ: Princeton University Press, 314-17. —— 1996. "Maṇḍalas on the Move: Reflections from Chinese Esoteric Buddhism Circa 800." Journal of the International Association of Buddhist Studies 19: 209-45. —— 1998. Politics and Transcendent Wisdom The scripture for Humane Kings in the Creation of Chinese Buddhism. University Park: The Pensilvania State University Press. Orzech, Charles D., Sørensen, Henrik H. and Payne, Richard K. eds. 2011. Esoteric Buddhism and the Tantras in East Asia. Leiden and Boston: Brill. Osto, Douglas. 2004. The Gaṇḍavyūha-sūtra: a study of wealth, gender and power in an Indian Buddhist narrative. Ph.D. dissertation, School of Oriental and African Studies, University of London. —— 2010. "A New Translation of the Sanskrit Bhadracari with Introduction and Notes." New Zealand Journal of Asian Studies 12 (2): 1-21. Owen, Lisa N. 2010. "Demarcating Sacred Space: The Jina Images at Kalugumalai." In International Journal of Jaina Studies 6 (4): 1-28. —— 2012. Carving Devotion in the Jain Caves at Ellora. Leiden: Brill. Padma, Sree and A.W Barber. 2008. Buddhism in the Krishna River Valley of Andhra. Albany: State University of New York Press. Rao, Gopinatha T.A. 1915. "Bauddha Vestiges in Kanchipura." Indian Antiquary 44: 127-29. Reis Habito, Maria. 1999. "The Amoghapasa Kalparaja Sutra: A Historical and Analytical 129  Study." Studies in Central and East Asian Religions 11: 39-67. Revire, Nicolas. 2012. "New Perspectives on the Origin and Spread of Bhadrāsana Buddhas throughout Southeast Asia (7 th-8 th centuries)." In Connecting Empires and States. Selected Papers from the 13th Conference of the European Association of Southeast Asian Archaeologists, Vol. 2, eds. Mai Lin Tjoa-Bonatz, Andreas Reinecke and Dominik Bonatz. Singapore: NUS Press, 127-143. Rhie, Marilyn M. 1988. Interrelationships between the Buddhist Art of China and the Art of India and Central Asia from 618-755 A.D. Napoli: Istituto universitario orientale. Sanderson, Alexis. 2007. "The Śaiva Exegesis of Kashmir." In Mélanges tantriques à la mémoire d'Hélène Brunner, Tantric Studies in Memory of Hélène Brunner, eds. Dominic Goodall and Andre Padoux. Collection Indologie 106, EFEO, Institut français de Pondichéry (IFP), 231-442. Schmid, Neil. 2006. "The Material Culture of Exegesis and Liturgy and a Change in the Artistic Representations in Dunhuang Caves, Ca.700-1000." Asia Major, 3d series, Vol.19, No 1/2, China at the Crossroads: a Festschrift in Honor of Victor H. Mair, 171-210. —— 2011. "Dunhuang and Central Asia." In Esoteric Buddhism and the Tantras in East Asia, ed. Charles Orzech, Henrik Sørensen and Richard Payne, Leiden and Boston: Brill, 365- 374. Sen, Tansen. 2016. Buddhism, Diplomacy, and Trade: The Realignment of India-China Relations, 600-1400. Lanham: Rowman & Littlefield. Skilling, Peter. 2011. "Buddhism and the Circulation of Ritual in Early Peninsular Southeast Asia." In Early Interactions between South and Southeast Asia: Reflections on Cross- cultural Exchange, eds. Pierre-Yves Manguin, A. Mani, Geoff Wade. Singapore: Institute of Southeast Asian Studies. New Delhi: Manohar, 371-384. Sha Wutian 沙武田 2013. Tubo tongzhi shiqi Dunhuang shiku yanjiu 吐蕃统治时期敦煌石窟 研究 / Studies on Dunhuang Grottoes during the Tubo Period. Beijing: Zhongguo shehui kexue chubanshe 北京 ：中国社会科学出版社 . ——2016. Yulinku di 25ku 榆 林 窟 第 25 窟 / The Yulin Cave 25. Beijing: The Commercial Press 北京 ：商务印书馆 . Shanxi Province Academy of Archaeology and TAR Relics Preservation Institute 陕 西 省 考 古 研 究 院 , 西 藏 自 治 区 文 物 保 护 研 究 所 2014. "Xizang Chayaxian Danmazha moya zaoxiang kaogu diaocha jianbao" 西 藏 察 雅 县 丹 玛 札 摩 崖 造 像 考 古 调 查 简 报 [Archaeological Report on Brag gYab County lDan Ma Brag Cliff Carvings]. Kaogu yu wenwu 考古与文物 /Archaeology and Relics 6: 7-14, 52-3. Sharf, Robert. 2017. "Buddhist Veda and the Rise of Chan." In Chinese and Tibetan Esoteric Buddhism. eds. Yael Bentor and Meir Shahar. Leiden and Boston: Brill, 85-120. Sharrock, Peter D. and Emma C. Bunker. 2016. "Seeds of Vajrabodhi: Buddhist Ritual Bronzes from Java and Khorat." In Esoteric Buddhism in Mediaeval Maritime Asia: Networks of 130  Masters, Texts, Icons. Singapore: I. SEAS-Yusof Ishak Institute, 237-252. Shen, Hsueh-man. 2019. Authentic Replicas: Buddhist Art in Medieval China. Honolulu: University of Hawai'i Press. Shim, Yeoung Shin. 2013. "Four Heavenly Kings: Iconography and Symbolism seen Through Literary Evidence and Imagery." Unpublished PhD thesis. The University of Pennsilvania. Snellgrove, David L. 1978. The Image of the Buddha. London: Serindia Publications: UNESCO. —— 1987. Indo-Tibetan Buddhism Indian Buddhists and their Tibetan Successors. London: Serindia Publications. Soper, Alexander C. 1960. "South Chinese Influence on the Buddhist Art of the Six Dynasties Period." Bulletin of the Museum of Far Eastern Antiquities 32: 47-112. Sørensen, Henrik H. 2000. "Perspectives on Buddhism in Dunhuang during the Tang and Five Dynasties Period." In The Silk Roads: Highways of Culture and Commerce, ed. Vadime Elisseeff. New York and Oxford: Berghahn Books, 27-48. —— 2011. "Central Divinities in Esoteric Buddhist Pantheon in China." In Esoteric Buddhism and the Tantras in East Asia, ed. Charles Orzech, Henrik Sørensen and Richard Payne. Leiden and Boston: Brill, 90-132. Sørensen, Per. 1994. The Mirror Illuminating the Royal Genealogies: Tibetan Buddhist Historiography: an Annotated Translation of the XIVth century Tibetan Chronicle: rGyal- rabs gsal-ba'i me-long. Weisbaden: Harrassowitz. Srinivasan, Sharada. 1999. "Lead Isotope and Trace Element Analysis in the Study of over One Hundred South Indian Metal Icons." Archaeometry 41: 91-116. Stoddard, Heather. 2018. "Drathang Monastery." In Murals of Tibet, eds. Thomas Liard, Robert Thurman, Heather Stoddard, Cameron Bailey, Jakob Winkler. Milan, New York, Berlin, London: Taschen, 96-105. Sundberg, Jeffrey R. and Rolf Giebel. 2011. "The Life of the Tang Court Monk Vajrabodhi as Chronicled by Lü Xiang ( 吕向 ) : South Indian and Śrī Laṅkān Antecedents to the Arrival of the Buddhist Vajrayāna in Eighth-Century Java and China." Pacific World: Journal of the Institute of Buddhist Studies 13 (3): 129-222. Takata, Tokio. 2019. "Tibetan Dominion over Dunhuang and the Formation of a Tibeto-Chinese Community." In Buddhist Road Paper 6.1. Special Issue: Central Asian Networks. Rethinking the Interplay of Religions, Art and Politics Across the Tarim Basin (5th-10th c.), ed. E. Forte, 85-106. Tanaka, Kimiaki. 2000. Tonkō: Mikkyō to bijutsu 敦煌 : 密教と美术 . Kyoto: Hōzōkan. —— 2018. An Illustrated History of the Maṇḍala : From Its Genesis to the Kālacakratantra. Boston: Wisdom Publications. Google books. https://play.google.com/store/books/details/ An_Illustrated_History_of_the_Mandala_From_Its_Gen?id=l8VWDwAAQBAJ&hl=en_ US 131 Thapar, Romila. 2003. Early India: From the Origins to AD 1300. Berkeley, CA: University of California Press. Toyka-Fuong, Ursula. 1998. "The Influence of Pala Art on 11th-Century Wall-Painting of Grotto 76 in Dunhuang." In The Inner Asian International Style 12th-14th Centuries, eds. Deborah Klimburg-Salter and Eva Allinger. Vienna: Verlag der Osterreichischen Akademie der Wissenschaften, 67-96. van der Kuijp, Leonard W.J. 1992. "Notes Apropos of the transmission of the Sarvadurgatipariśodhanatantra in Tibet." Studien zur Indologie und Iraninistik (StII) 16: 109-125. van Schaik, Sam. 2015. Tibetan Zen: Discovering a Lost Tradition. Boston and London: Snow Lion. van Schaik, Sam. and Lewis Doney. 2007. "The Prayer, the Priest, and the Tsenpo: An Early Buddhist Narrative from Dunhuang." Journal of International Association of Buddhist Studies 30 (1-2): 175-217. van Schaik, S. and Imre Galambos. 2012. Manuscripts and Travellers: The Sino-Tibetan Documents of a Tenth-Century Buddhist Pilgrim. Berlin: De Gruyter. Verardi, Giovanni. 2018. The Gods and the Heretics, Crisis and Ruin of Indian Buddhism, New Delhi/Buenos Aires: Aditya Prakashan & Fundación Bodhiyāna. Vitali, Roberto. 1990. Early Temples of Central Tibet. London: Serindia Publications. von der Schulenburg, Stephan., Elke Hessel, Karsten Schmidt, Matthias Wagner eds. 2015. Buddha: 108 Begegnungen/Encounters. Koln: Wienand Verlag. von Schroeder, Ulrich. 1990. Buddhist Sculptures of Sri Lanka. Hong Kong: Visual Dharma Publications —— 2001. Buddhist Sculptures in Tibet. Hong Kong: Visual Dharma Publications. Wade, Geoff. 2014. "Beyond the Southern Borders: Southeast Asia in Chinese Texts to the Ninth Century." In Lost Kingdoms: Hindu-Buddhist Sculpture of Early Southeast Asia, ed. John Guy. New York: Metropolitan Museum of Art. Wang, Eugene Y. 2005. Shaping the Lotus Sutra Buddhist Visual Culture in Medieval China. Seattle and London: University of Washington Press. Wang, Michelle C. 2018. Maṇḍalas in the Making: The Visual Culture of Esoteric Buddhism at Dunhuang. Leiden: Brill. Wangdu, Pasang and Hildegard Diemberger. 2000. dBa' bzhed: The Royal Narrative Concerning the Bringing of the Buddha's Doctrine to Tibet. Wien: Verlag der Österreichischen Akademie der Wissenschaften. Wayman, Alex and Ryujun Tajima 1998. The Enlightenment of Vairocana. Delhi: Motilal Banarsidass Publishers. Weinberger, Steven N. 2003. The Significance of Yoga Tantra and the Compendium of Principles (Tattvasaṃgraha Tantra ) within Tantric Buddhism in India and Tibet. Ph.D. 132  dissertation, University of Virginia. —— 2010. "The Yoga Tantras and the Social Context of Their Transmission to Tibet." Chung- Hwa Buddhist Journal 23: 131-66. Wessels-Mevissen, Corinna. 2001. The Gods of the Directions in Ancient India: Origin and Early Development in Art and Literature (Until c. 1000 A.D.). Monographien zur indi- schen Archäologie, Kunst und Philologie, Band 14. Berlin: Dietrich Reimer Verlag. Whitfield, Roderik. 1982-86. Art of Central Asia: The Stein Collection in the British Museum. 3 vols. Tokyo: Kodansha International, in cooperation with the Trustees of the British Museum. Wong, Dorothy C. 1993. "A Reassessment of the Representation of Mt Wutai from Dunhuang Cave 61." Archives of Asian Art, 27-52. —— 2007a. "The Case of Amoghapasa." Journal of Inner Asian Art and Archeology 2: 151-58. —— 2007b. "The Huayan/Kegon/Hwaŏm Paintings in East Asia." In Reflecting Mirrors: Perspectives on Huayan Buddhism, ed. by Imre Hamar. Wiesbaden: Harrassowitz. —— 2012. "The Art of AvataṁsakaBuddhism at the Courts of Empress Wu and Emperor Shomu/Empress Komyo." in Avataṁsaka Buddhism in East Asia: Huayan, Kegon, Flower Ornament Buddhism: Origins and Adaptation of a Visual Culture, eds. Imre Hamar and Robert M. Gimello. Wiesbaden: Harrassowitz Verlag, 223-260. —— 2018. Buddhist Pilgrim-Monks as Agents of Cultural and Artistic Transmission: The International Buddhist Art Style in East Asia, ca. 645-770. Singapore: NUS Press. Wong, Dorothy C. and Eric M. Field. 2008. Hōryūji Reconsidered. Newcastle: Cambridge Scholars. Yang Qingfan, Lu Suwen and Zhang Yanqing 杨 清 凡 , 卢 素 文 , 张 延 清 . 2017. "Xizang Mangkang Gatuozhen xinfaxian Tubo moya shike diaocha jianbao" 西 藏 芒 康 嘎 托 镇 新 发 现 吐 蕃 摩 崖 石 刻 调 查 简 报 [Archaeological Report on the Newly Discovered Tubo Dynasty Rock Sculptures in the Town of Sgar thog in Smar khams County, Tibet]. Zangxue xuekan 藏学学刊 /Journal of Tibetology 16: 233-251. Yang Zeng. 2018. "A Biographical Study on Bukong 不 空 (aka. Amoghavajra, 705-774): Networks, Institutions, and Identities." Unpublished Ph.D thesis. The University of British Columbia. Zhang Wenbin ed.张文彬编 2000. Dunhuang: Jinian Dunhuang cangjingdong faxian yibai zhounian 敦煌 ：纪念敦煌藏经洞发现一百周年 / Dunhuang: A Centennial Commemoration of the Discovery of the Cave Library. Beijing: Chaohua Chubanshe 北京 ：朝华出版社 . ◆ Author: Yury Khokhlov, Independent researcher, the United Kingdom. 133  Abstracts 尼泊尔上穆斯塘地区桑宗墓地所见本教施垛仪轨的考古学证据 马克·奥登德弗 （美国加利福尼亚大学人类学与遗产研究系） 在考古发掘中，发现确凿清晰的有关古代仪式方面的证据一向是很困难的。但也有例外，比 如我们在尼泊尔穆斯塘北部地区发掘的古代墓葬，其中所出一系列遗物很可能属于所谓的本教施 垛（Mdos rgyab）仪式。公元 5 世纪晚期的桑宗 5 号墓中，我们发现了铁三脚架、铜容器及木杯 和竹杯子，这一套器物与民族志描述的有关本教施垛仪式中所用的道具惊人地相似。本文解释了 何为本教的施垛仪式，并且依据桑宗 5 号墓的墓葬背景和出土的人工制品，可认为至少在公元 5 世纪本教已见于穆斯塘北部。 305  羊距骨 ：西藏最古老的博具和占卜工具的考古学和民族史初探 约翰·文森特·贝勒扎 （美国弗吉尼亚大学） 本文旨在通过考古学和民族史的证据，对藏族文化中流行的一种物件——“羊距骨”做综合 性考察。首先提出了鉴别墓葬中那些具有文化意义羊距骨的标准，并比较了欧亚大陆其他地区同 类实例。本文特别关注到一件西藏的红铜合金的羊距骨，这为讨论该类物品在高原古代文化中的 意义提供了佐证。文本还讨论了羊距骨在藏族博具、占卜、计数及其他仪式、观念中的作用。最 后，本文聚焦羊距骨的跨文化意义，尝试说明羊距骨能够帮助我们进一步理解史前晚期（约公元 前 1200 年至公元 600 年）及历史时期欧亚大陆跨文化交流的一些面貌。 Re-examination of Materials about the Communications between Nepal and China in the Early Tang Dynasty Wang Bangwei (Peking University) The Bod-Balpo Ancient Road has a long history, which can be traced back to the Paleolithic Age on the basis of archaeological discoveries. As for the history of the Bod-Balpo Ancient Road, we can distinguish three periods, namely, pre-, during, and post-Tang dynasty to the Qing dynasty. In the specific case of the Bod-Balpo Ancient Road, we need to consider the situation of East Asia, South Asia and even Central Asia at that time, the chronological order of the literary records and the recent archaeological discoveries. In the meantime, although many researchers have noticed and quoted the literary records on the Bod- Balpo Ancient Road, there are still some problems that require further studies. In addition, we also should pay attention to the relationship between the Buddhist monks and diplomatic envoys and merchants, they always traveled together at that time. In fact, the routes taken by Buddhist monks coincide with those taken by commerce. 306  在河西走廊和西藏发现的不空之影响 尤利·霍赫洛夫 （英国独立学者） 本文分析了莫高窟和榆林窟被认为是受到吐蕃影响的早期藏传佛教艺术遗存。与此观点相反， 本文认为这些图像实际上是受到了南印度艺术的影响，尤其是泰米尔纳德帕拉瓦王国艺术的影响， 这与 8 世纪上半叶来自帕拉瓦王国的著名佛教大师金刚智在中国的传法密切相关，更与不空 8 世 纪中叶在河西的活动有关。这些图像属于金刚智所创造的一种独特艺术传统，他也是一位伟大的 画家。吐蕃统治时期的艺术只是吐蕃占领河西以前已经存在的艺术和宗教传统的继续延续而已。 不仅如此，吐蕃控制河西走廊使得吐蕃能够移植当时当地流行的宗教和艺术潮流。因此，吐蕃将 佛教作为国教正是基于不空在唐代所提出的以佛法护国的模式，本文支持这一理论。 A Preliminary Survey on Mthing shog Manuscripts Zhu Lishuang (Center for Dunhuang Studies, Lanzhou University) Mthing shog, translated as blue-black paper 蓝 黑 纸 in Tibetan-Chinese dictionaries, is a kind of colored paper that we find in Tibetan culture. Its production covers three steps: 1) several layers of papers are laminated with a wheat paste; 2) a specially made coloring material is applied to the surface of the paper; and 3) the paper is burnished with a gzi bead or a smooth piece of conch shell or other smooth and hard precious stones. The recipes and substances of making coloring materials for mthing shog vary. For instance, one formula uses black ink, which is made from soot (sre nag) or burnt resinous pine wood (sgron shing), which is mixed with yak and/or sheep brains; one formula uses a mixture consisting of black vitriol (nag mtshur), euphorbia (thar nu), borax (tsa la), wheat beer (gro chang), and myrobalan (a ru ra); one formula uses a solution made with ground white cowry shell ('gron bu), ground cardamom (dzā ti) and myrobalan juice; one formula uses indigo (rams). Judging from literary sources and extant objects, the expression mthing shog derives from shug bu mthing ga of the Tibetan dynastic period. It is mthing shog that was used for writing royal edicts. Beginning from the 10th century, mthing shog is most often used to write Buddhist sutras. The most prized manuscripts frequently use gold ink or a combination of gold and silver ink. The main purpose of preparing mthing shog manuscripts is for devotional purposes. These are deposited in shrines and monasteries and prepared for the accumulation of merit; they are not intended for circulation. 307  Chinese scholars often term the black-blue paper of Tibetan cultural tradition as 瓷 / 磁 青 纸 , "porcelain green paper," an indigo-dyed paper in Chinese culture. However, as mentioned above, there are many different materials that are used to make mthing shog. Mthing shog includes indigo-dyed paper, but they are not identical. The gold or/and silver manuscripts produced in China proper, especially those in the Ming and Qing courts, may be written on indigo-dyed paper or mthing shog. As for gold or/and silver manuscripts from the Tibetan cultural area, the situation is more complicated, and it seems that not all gold or/and silver texts are written on mthing shog. An Interpretation of the Mural of Lce btsun Shes rab 'byung gnas in Zhwa lu Monastery Jia Yuping (Collage of Art and Design, Chengdu University) This paper discusses the mural depicting Lce bstun Shes rab 'byung gnas (the 11th century founder of Zhwa lu) in the Sgo gsum ma lha khang of Zhwa lu monastery. The history of the monastery and the Lce clan from the 11th to the 14th century are fully presented through the images and inscriptions in a series of murals. Combining the literature on the genealogies of Zhwa lu and related Tibetan sources, this paper discusses some stories about Zhwa lu that were hitherto not well known. These include aspects of its religious background and its connections with the Indian subcontinent, as well as its contact with western Tibet during the period of the second diffusion of Buddhism. Its governance through a combination of politics and religion makes Zhwa lu one of the earliest monasteries that was under the control of a family- owned theocratic system. On the Mural Painting of Buddha's life in the Corridor of the Assembly Hall on the First Floor of Zhwa lu Monastery Meng Yu (School of Chinese Classics, Renmin University of China) The 101st mural painting in the corridor of the Assembly Hall on the first floor of Zhwa lu Monastery is concerned with the biography of the Buddha, from his birth to nirvana. Twenty significant scenes of the Buddha's life were chosen and depicted in this mural. This essay provides a detailed discussion of each scene and concludes that the 101st mural painting is based on the Lalitavistara and other sources. The life 308 of the Buddha is a popular subject in Buddhist art and literature. It can be found in many texts, besides the one in Skyes rabs brgya ba. In addition to texts, the illustrations of the Buddha's life can also be found in other parts of Tibet. This essay also briefly compares the stories of the Buddha's life in the Skyes rabs brgya ba and other literary sources in Tibetan, and its representation in Zhwa lu and other places. A conclusion is drawn that, prior to the sixteenth century, the textual and pictorial representations of Buddha's life in Tibet share obvious similarities. 桑噶地区赤洛纳特观音崇拜的艺术史证据 林瑞宾 （美国西北大学） 自 19 世纪以来，赤洛纳特寺（Triloknāth Mandir）就一直吸引着宗教史、人类学、艺术史和 建筑各领域学者的兴趣。该寺位于喜马偕尔邦钱德拉巴嘎河谷的屯德村（藏语称“热帕”村），属 于西喜马拉雅佛教文化圈。竹巴噶举派喇嘛达仓热巴（1574-1651 年）于 17 世纪时到过这里，在 其所撰朝圣行纪中称该地为“乌苌那” ，引起了图齐的关注。赤洛纳特寺诱人之处部分缘于以下现 象，即现在该地居民将寺中的白色大理石造像视为湿婆神并加以供奉，而来自拉胡尔、桑噶及其 他地区的佛教信徒则认为这是一尊佛教造像。近年来，一位佛教僧人和一位印度教祭师被委派至 该寺以协助不同宗教的朝圣者。寺中另一独特之处是该寺建筑为通常见于北印度婆罗门教建筑的 西柯罗高塔，寺内却供奉佛教雕塑，造像材质为白色大理石，在喜马拉雅地区佛教造像中极为罕 见。关于该寺的很多问题都观点不一，如寺院的创建时间，这尊造像何时被安置在寺中，寺院建 筑最初是否为印度教寺院而在后来容纳了一尊佛像，抑或建寺伊始就是佛教寺院？暂且遑顾上述 众多未决之疑问，本文拟集中探讨这尊造像的图像问题，善趣观音（Sugatiśaṃdārśana Lokeśvara） 及其与克什米尔的联系，相邻的桑噶地区的图像推测应系模仿今赤洛纳特寺白色大理石造像。大 部分桑噶地区的造像所属时期均相对较晚，因此本研究对赤洛纳特寺白色大理石像的断代、以及 此像何时取代了另一尊较早的灰色小石像诸问题也将提供一些依据。 309  Research on the Seals Granted by the Ming Dynasty to Tibet Li Shuai (Post-Doctor, Department of Archaeology, Sichuan University) The types of seals granted by the Ming dynasty to Tibetan individuals include at least four types: religious — title seals, knighthood seals, official seals and stamp seals. The difference of these seals depends on the social status of the grantees. The type of seal and the motivation of granting them reflected the flexibility and diversity of Ming policies towards Tibet. The seals did not only function as political symbols, they also had practical functions in Tibet. They were used to announce local official statements, to handle local affairs, etc. Some of the seals continued to be in use until well into the Qing dynasty and thus form important evidence that their recipients maintained their political status and were able to influence the governance of Tibet. Research on the Official Seals of the Chieftain (Tusi) in the Tibetan Region of Sichuan Province in the Qing Dynasty Liu Sha (Sichuan Museum) Based on the literary sources, this article investigates seventeen seals of the chieftains (tusi, 土司 ) of the Tibetan region of Sichuan Province in the Qing Dynasty in terms of their nature, characteristics, origin, how they were awarded and how they functioned. This paper considers that the official seals used by the chieftains were mainly granted by the central government of the Qing Dynasty. According to the official seal system of the Qing court, the texture, shape, size, impression and inscription of these seals have special provisions. As political tokens, these official seals played a multifaceted role in local society aside from having practical functions. They were highly valued by the chieftains over generations. Through comparison, it was found that the official seals used by the chieftains in the Sichuan Tibetan areas were quite different from those that were in use in Central Tibet. This was no doubt owed to the differences in their administration. The study of the official seals of the chieftains serves to further explore the policies and characteristics of the Qing Dynasty's governance of the border areas. 310  The Imperial Kapala Drums of the Qianlong Emperor Lin Huan (The Palace Museum, Beijing) This paper studies the kapala drums (ḍamaru) that are found in collection of the Palace Museum. They belong to the Qianlong period of the Qing dynasty. The ḍamaru drums were mainly made of human skull caps, ivory, jade or wood and are also called "hand drum" or "tambourine," and they belong to the inventory of musical instruments or ritual instruments. The drums in the Palace collection mostly came from the Tibetan areas in the form as tribute to the Qing court. Based on archival sources and extant objects, we know that the Qianlong Emperor had asked the court workshops partly to change these drums and also ordered the court workshops in Suzhou to imitate these with different materials. The extant drums show different influences from Tibetan, Han Chinese and Mongolian culture. They reflected the Qianlong Emperor's aesthetic taste. These drums were decorated with turquoise, coral, beeswax and tridacna, etc. and thus became luxury objects. However, the shape of the drums and their packaging were not overly modified, and maintained the simple nature of the artifact itself. These drums also demonstrate that under the "great unification" pattern, the mature handicraft skills of the border areas could be transplanted to the court (and even to Suzhou). These skills were quickly mastered and reached a very high standard. Fa Wang Xin Sheng Zhuan —— Stories of Qianlong Emperor's Incarnations Li Ruoyu (The Palace Museum, Beijing) The Palace Museum collection, no. Zong 23584, contains a book that is titled Fawang xinsheng zhuan 法 王新胜传 . It was composed by the 6th Panchen Lama Blo bzang dpal ldan ye shes (1738-1780) in 1780 when he visited Rehe for the celebration of the Qianlong Emperor's birthday. It contains eleven stories of the Qianlong Emperor's previous lives thereby creating a system of the emperor's reincarnations. In this system, the Qianlong Emperor's earlier incarnations included kings who protected Buddhism, siddhas and great gurus of the Bka' gdams pa and Dge lugs pa schools. This paper studies the date and background when this book was produced. It also includes translations of parts of the stories for the first time in the hope of achieving a better understanding of the history of this period and this kind of Tibetan text. 311  西藏的胎藏界曼荼罗传统及其彩砂曼荼罗制作 田中公明 （日本中村元东方研究所） “两界曼荼罗”于 9 世纪初自唐传入日本，不仅构建了日本佛教造像的核心体系，并对整个 日本文化产生了深远影响。在这两类曼荼罗中，与“金刚界曼荼罗”相关的遗存较为丰富，其原 型出自印度，尼泊尔及西藏继承了的印度的曼荼罗传统，并有大量文献记录和绘画保存至今。然 而，留存下来的“胎藏界曼荼罗”相关文献却非常少，其原因在于基于《大日经》的密教体系在 古印度很早就已衰亡。此外， 《大日经》在吐蕃时期自印度传入西藏，迄今仍有一些胎藏界曼荼罗 的实例得以留存，尽管这类遗存极为稀少。此前，本人曾于 2001 年在日本京都日本文化国际研究 中心举行的一次国际学术讨论会，及 2003 年于牛津举行的第 10 届国际藏学会上，就西藏的胎藏 界曼荼罗传统及相关遗存发表过演讲。但当时限于交通条件，我未能亲自考察位于安多的拉加寺、 夏琼寺，这两座寺院至今仍几乎每年都要制作胎藏界曼荼罗的彩砂坛城 ；尽管 1994-1996 年间当 我担任富山市瑜伽禅修博物馆的首席讲师时，藉馆内复原藏传佛教两界曼荼罗之机，就已经获得 了一些源自拉卜楞寺、拉加寺的胎藏界曼荼罗资料。近年来，由于安多地区交通大为改善，使我 这样的国外学者得以至上述寺院实地考察。本文主要在笔者于 2014、2015 年田野考察的基础上， 拟通过对安多寺院中胎藏界曼荼罗彩砂坛城的制作进行初步介绍，并与唐密—东密传统的胎藏界 曼荼罗相比较，进而概括总结其特点。 312  Journal of Tibetology (Vol.21) Edited by Center for Tibetan Studies of Sichuan University Chengdu, China ISBN 978-7-5211-0206-2 First Published in December 2019 China Tibetology Publishing House Beijing, China  图书在版编目(CIP)数据 藏学学刊 . 第 21 辑 / 四川大学中国藏学研究所编 . —北京：中国藏学出版社， 2019.12 ISBN 978-7-5211-0206-2 Ⅰ . ①藏… Ⅱ . ①四… Ⅲ . ①藏学－文集 Ⅳ . ① K281.4-53 中国版本图书馆 CIP 数据核字（2019）第 289412 号 藏学学刊〔第 21 辑〕 四川大学中国藏学研究所 主编 责任编辑  张荣德 藏文责编  顿珠次仁 装帧设计  翟跃飞 出版发行  中国藏学出版社 印    刷  中国电影出版社印刷厂 版    次  2019 年 12 月第 1 版第 1 次印刷 开    本  787 毫米 ×1092 毫米 1/16 字    数  300 千字 印    张  20.25 定    价  58.00 元 书    号  ISBN 978-7-5211-0206-2 图书如有质量问题，请与本社联系 E-mail: dfhw64892902@126.com 电话: 010-64892902 版权所有 侵权必究  《藏学学刊》稿约 《藏学学刊》（བོད་རིག་པའི་དུས་དེབ། Journal of Tibetology）系中国教育部普通高等学校人文社会科 学重点研究基地四川大学中国藏学研究所主办的藏学类专业性学术刊物，CSSCI（2014-2018）来 源集刊，创刊于 2004 年，旨在搭建藏学研究的学术交流平台。从 2014 年开始，本刊由年刊改为 半年刊，接受中、英文稿件，设有论文、书评、译文等栏目，热忱欢迎国内外藏学研究者惠赐稿件。 来稿时请注意以下事项 ： 1. 来稿请注明字数并提供作者简介，包括姓名、出生年月、性别、民族、籍贯、职称、学位、 工作单位、联系方式和主要研究方向等，其中姓名和工作单位名称请提供正确的英文译名。 2. 来稿必须为原创性作品，此前未经公开发表。严禁抄袭、剽窃和一稿多投，如有发现，将 永不刊用该作者来稿。 3. 为保证文稿的准确性，电子版来稿须同时提供 word 和 pdf 两种文档格式并遵循本刊用稿规 范（详见本所网页）。来稿字数不限，提倡言简意赅，以一万字左右为宜。 4. 本刊每期收稿截止日期为每年六月三十日和十二月三十日。该日期以后收到的稿件将纳入 下一期的编辑工作。 5. 本刊取舍稿件惟以学术为标尺，并实行 2-3 名专家匿名审稿制度，根据专家审稿意见决 定稿件取舍。本刊在尊重原作的前提下，有权对拟刊用稿件作必要的删改并告知作者，作者如果 不同意对文稿作文字性修改或适当删节，请在来稿时予以说明。 6. 本刊所刊文章均为作者个人观点，不代表编辑部意见，文责由作者自负。 7. 本刊已加入期刊数字化网络系统，作者若无此意愿，请在来稿时注明，否则视为默许。 8. 稿件请直接投寄本刊编辑部。自寄稿之日起，若二个月内未接到用稿通知，可自行处理。 来稿除图版原件外，一律不退稿，敬请作者谅解。 9. 本刊对刊用文稿将支付作者稿酬，并赠送样刊五册。 10. 本刊热诚欢迎国内外学者或编辑部与本刊建立资料互赠交流关系。 地 址 ：中国四川省成都市望江路 29 号 四川大学中国藏学研究所《藏学学刊》编辑部 邮 编 ：610064 电子邮件 ：zangxuexuekan@163.com 电话 / 传真 ：+86-28-8541 2567 网 址 ：http://www.zangx.com  Call for Contributions Founded in 2004, the Journal of Tibetology (བོད་རིག་པའི་དུས་དེབ། 藏 学 学 刊 ) is a peer-reviewed bilingual scholarly journal dedicated to publishing papers in the field of Tibetan Studies. Featuring articles and reviews in either Chinese or English, the journal is published biannually by the Center for Tibetan Studies of Sichuan University, Chengdu, PRC. The Journal of Tibetology welcomes the submission of academic and unpublished (and original) work, including the Chinese translation of foreign research and serious, critical reviews of books or review articles, in any area of research that deals with the Tibetan cultural area. Essential Guidelines: 1. The manuscript should contain information on the total number of words/characters and author's details, including the following information: (1) full name and institutional affiliation; (2) academic title; (3) contact information, (4) the primary field of research. 2. The Journal of Tibetology has a zero-tolerance plagiarism policy. 3. Essays and reviews should be submitted electronically in Microsoft Word file and PDF format, using Unicode, to zangxuexuekan@163.com and conform to the style sheet of the Journal of Tibetology that is found at the website of the institute. If this is not done, they will automatically be not considered for publication. There is no limit on the length of the manuscript, but we strongly encourage the manuscript to be concise with approximately 10,000 words in English or Chinese. 4. The deadline for the submission of articles and reviews is June 30 of each year. Articles and reviews received and accepted for publication after this date will be considered for the next issue. 5. The Journal of Tibetology uses a double-blind review process. Each manuscript is sent to two or three referees for double-blind peer review. Based on their recommendations, the editor then decides whether the manuscript will be accepted as is, whether it needs to be revised, or whether it will be rejected. 6. The views and opinions expressed in the articles and reviews are those of the author alone and do not reflect the views or opinions of the editor(s) or the editorial board. The author is responsible for his/ her own views. 7. The journal is already part of the digital network of Chinese journals. If any author does not want to have his or her article published online, please note this upon submitting the manuscript. Otherwise, the editor will take it as tacitly granting permission to do so. 8. The manuscript should be directly submitted to the editor. If one has not received a notice of acceptance for publication within two months since the day the manuscript was submitted, it is within one’s discretion to submit the manuscript elsewhere for publication. Excepting the originals of maps, charts or photographs the editor will not return the manuscript to the author. 9. The journal will pay remuneration to the author and send five copies of the issue of the journal in which his or her article appears. 10. Scholars or editorial boards in China and abroad are warmly welcomed to establish a growing network of exchanging copies of journals with the editor of the Journal of Tibetology. Correspondence should be addressed to: Dr. Zhang Changhong Center for Tibetan Studies of Sichuan University Chengdu, Sichuan Province P.R. China 610064 Email : zangxuexuekan@163.com Tel/Fax: +86-28-8541 2567 Website: http://www.zangx.com