How did the Heruka get his wings? And why did the Guru have a feather in his cap? Avian symbolism in rNying ma iconography.

Introduction

Within Tantric or Vajrayana Buddhism, probably the most important and prestigious objects of meditation and worship are a class of wrathful male deities known as Heruka in Sanskrit, or khrag 'thung (literally, 'Blood Drinker') in Tibetan. These are the great meditational deities that are represented in the famous tantric scriptures, which are often known eponymously after the deity whose worship they describe: Cakrasamvara, Hevajra, Kalacakra, Yamari, and so on. The iconography of such deities shares a lot with the cognate ferocious deities of Indian Saivism: often multiarmed, more rarely also multi-legged,2 they dwell within a frightening cemetery, ferocious, surrounded by flames, embracing a consort, kitted out kapalika style with only a tiger-skin apron, bone ornaments, and face-paints of cremation ash and grease. In their hands they usually hold a blood-filled skull cup and often a khatvanga trident, along with sundry other weapons particular to each deity.

Yet in the influential and popular 'Ancient' or rNying ma school of Tibetan Buddhism, we find something additional and different: while resembling the classic Indian heruka deities in most respects, they display a prominent avian symbolism as well.3 Most notably, the rNying ma herukas frequently have huge wings attached to each shoulder. In addition, a divine bird called khyung, partly indigenous but also identifiable with the Indian garuda, can sometimes circle overhead. Despite its widespread occurrence in Tibetan rNying ma, this kind of avian iconography seems to have been vanishingly rare in Indian depictions of the Buddhist herukas. It seems to be so rare in surviving evidence from south of the Himalayas, but so widespread in Tibet, that we are compelled to inquire if its prevalence there might signal a predominantly Himalayan and Tibetan inspiration. In this paper, I will begin a preliminary investigation of the evidence surrounding this little understood avian symbolism, and explore intriguing clues from the 10th century Dunhuang texts and elsewhere, that might help us understand how, when, why, and from whom, the Tibetan herukas acquired their wings.

1 My thanks to Claudine Bautze-Picron, Daniel Berounsky, Gudrun Buhnemann, Cathy Cantwell, Brandon Dotson, Paul Gerstmayr, Sanjukta Gupta, Georgios Halkias, Dan Hirshberg, Toni Huber, Soloin Kij, Janice Leoshko, Dan Martin, Charles Ramble, Anne de Sales, Geoffrey Samuel, Alexis Sanderson, Anna Sehnalova, Peter-Daniel Szanto, Jay Valentine, Jeff Watt, and Michael Willis, for their generosity in answering various queries, and offering comments, advice, and inspiration.

2 Gudrun Buhnemann informs me that while multi-armed deities are widespread in Indian iconography, multiple legs are rarer. The multiple legs more often found in Tibetan images might also therefore originate in Himalayan iconography (personal communication, 8th May, 2016).

3 It has been known for some time that canonical rNying ma tantras from the rNying ma'i rgyud 'bum (NGB) collection contain a sprinkling of non-Indic materials: during my PhD research in the 1990's, I discovered that the ’go ba'i lha Inga, for example, appear within the Phur pa bcu gnyis, as do some other indigenous categories (Mayer 1996:132-147). I have been interested in this phenomenon ever since I first encountered it in the 1990's, but am only now beginning to consider how it might have developed. This paper is a start.

rNying ma herukas often have wings

A unique feature of the rNying ma pantheon is the organisation of its heruka deities into rationalised sets, and this offers us an easy way to approach the wide prevalence of its distinctive avian symbolism.

First are the 'Eight Pronouncement Deities', or Bka' brgyad, which initially come to our attention in the revelations of the 12th century codifier of the rNying ma, Nyang ral nyi ma'i 'od zer (1124-1192). Attributed to the lineages of Padmsambhava, the legendary founder of the rNying ma tradition, this eightfold set of herukas became the structuring device around which much of the rNying ma scriptural canon was organised, and a recurrent theme throughout rNying ma literature. Iconographically speaking, the Eight Pronouncement Deities are a set of eight almost identical herukas whose main differences are in the implements they carry. Each has three heads, six arms, four legs, and of course two wings, and they are represented with the thick limbs and stocky bodies said to be typical of earlier Indian heruka iconography. There are innumerable textual and visual depictions of these eight.

Secondly, there are the 'Hundred Peaceful and Wrathful Deities' (zhi khro), derived from the famous Guhyagarbha tantra, but perhaps best known in the West though its adoption in the bar do literature of death rituals, made famous by W Y Evans Wentz in his mis-named 'Tibetan Book of the Dead'. Here again, in numerous illustrations, one will find that the wrathful heruka deities have wings: yet as I will point out later, possibly not because of what is written in the Guhyagarbha tantra itself.4 5

Indian herukas don't usually have wings

Yet few if any Indian herukas seem to have had wings (of course, there are a number of winged deities in Indian iconography, such as those from 1st century BCE to 1st century CE Chandraketugarh in Bengal (Bautze 1995: 33, 35), or the various forms of Garuda, but these are not herukas of the tantric period). I consulted a dozen leading Indological specialists in tantric art and literature, but none were aware of anything beyond the most meagre evidence for winged herukas in extant Sanskrit sources.’ I also consulted Jeff Watt, director of the Himalayan Art Resources, following which he began to research the topic and opened a Winged Deity page on the Himalayan Art Resources website. Yet according to the data accumulated on his web page, it would

4 The authenticity of the Guhyagarbha tantra was heavily contested in Tibet, with powerful critics such as Bu st on denying its Indic origins. Yet as van der Kuijp points out, there is a good likelihood that there was once a Sanskrit version: "...SakyasrT's [1127-1225] discovery of a Sanskrit manuscript of this work in Bsam yas monastery should have dispelled part of the critique that was leveled against the authenticity of its Tibetan text, namely that there was no original Sanskrit text that corresponded to the Tibetan translation^]. The fact of the existence of its Sanskrit counterpart was again underscored by Bcom ldan ral gri more than half a century later, and then again in the fourteenth century by G.yung ston Rdo rje dpal (1287-1365) and 'Bri gung Lo tsa ba Nor bu dpal ye shes (1313-87), alias ManikasrTjnana" (van der Kuijp 2016: 289). Could the lack of wings in this text now be considered as further possible circumstantial evidence in favour of its Indic origins? 5 Alexis Sanderson was able to point me to a minor Saiva female deity who has wings: Mantradamarika Pratvangira from the Jayadrathayamalatantra, satka 3, patala 10. Teun Goudriaan and Sanjukta Gupta (1981: 115) mention a winged Saiva Bhairava, Akasabhairava, which is particularly worshipped in Nepal, but Sanjukta Gupta now tells me that she long ago concluded this form is of Himalayan rather than Indian origins (personal communication, 16th November 2015). Likewise, Gudrun Btihnemann has pointed me to a scroll from Nepal depicting a heruka who, like the popular rNying ma form, is three-headed, six-armed with wings, embracing a consort and four-legged, stepping on the 4 Maras. She is not clear if this deity is a form of Cakrasamvara or Heruka, but in any case, it is, like Akasabhairava, Nepalese or Himalayan, rather than Indian. For a discussion, see Btihnemann 2008: 110, and Kreijger 1999: 99.

eem that winged deities are predominantly Bon or rNying ma. Out of the scores of known Indic heruka deities preserved in Tibetan sources, the solitary candidate Watt can find for a winged heruka of Indian origins is bDe mchog rdo rje mkha’ lding (Vajragarudasamvara, HAR Item number 40418), a form of Samvara combined with Garuda, surrounded by a retinue of a further fifty deities in garuda form, from Taranatha’s sGrub thabs rin ’byung brgya rtsa (deity 228), and transmitted in some dGe lugs and Jo nang pa traditions. It is said this tradition was transmitted by Rwa lo tsa ba rDo rje grags, which might indicate Indian rather than Tibetan origins.6 However, in this case, the wings only appear because heruka is here combined with garuda: thus the wings derive from garuda, and are not inherent to the heruka, as in so many rNying ma pa examples. Similarly there is a Hayagriva with Garuda Wings claimed to be of the Atisa tradition (HAR Item Number 40354, rTa mgrin khyung gshog can jo bo'i lugs), but it is not yet clear to what degree this might be of Indian provenance.

More speculatively speaking, we know that there were significant cultural transmissions into Tibet from the Indic world's far north-west, more specifically, from a wider Uddiyana cultural region that encompassed Gilgit and adjacent regions of contemporary Northern Pakistan (Klimburg-Salter 2016), so that one cannot rule out that winged Buddhist herukas might have entered the Tibetan empire from those regions. Some rNying ma tantras even claim to have been translated from the languages of Uddiyana (Tib: O rgyan) or Baltistan (Tib: Bru zha), although the origins and transmissional histories of such texts are not yet entirely clear.7 But if winged herukas did exist in a wider Uddiyana, they have left us no archaeological traces, nor any textual traces in their original languages, so that it is difficult to construct a sound diffusion model based upon them.

Avian symbolism in non-Buddhist Tibetan culture

To test our hypothesis that the inspiration for the rNying ma herukas' wings might be as much Himalayan or Tibetan as Indian, we need to establish that pertinent avian symbolism was prevalent at the time, firstly within the wider Tibetan cultural regions, and secondly within those specific areas where we believe the Buddhist heruka traditions were first introduced to Tibet.

In respect to the first question, it is abundantly clear that dense, varied and extremely prominent uses of avian symbolisms occured within elite and popular levels of the Tibetan and Himalayan religions, it seems encompassing most of their many

6 According to the notes attached to TBRC Resource ID T711

7 One such text, the famous root scripture for rNying ma Anuyoga the Mdo dgongs 'dus, even includes a winged heruka. Yet the provenance and dating of this text is not at all clear. In his doctoral dissertation devoted to this huge text, Jacob Dalton concludes that most of it was probably composed in Tibet, with only the narrative section on the taming of Rudra possibly originating in Baltistan (Dalton 2002: 266ff). In his PhD, Dalton placed the composition of the Mdo dgongs 'dus as we have it in 9th century Tibet, but his dating depends on the dates of gNubs chen sangs rgyas ye shes, who wrote a famous commentary on the Mdo dgongs 'dus. More recently, Dalton has slightly revised his dates for gNubs chen to the late 9th to early 10th century (Dalton 2014: 145). Almost two decades earlier, Vitali had likewise estimated the early 10th century (Vitali 1996: 546-7). But other scholars, notably Samten Karmay (2007: 99-103) who did considerable research on the subject, and following him, Carmen Meinart (Meinert 2002: 290), and now also Jose Cabezon (Cabezon 2013: 239 n.32), place gNubs chen around a hundred years later than Dalton, in the late 10th century, which suggests a possibility that the Mdo dgongs 'dus itself might be later than Dalton's 9th century date. In the light of these uncertainties of origins and dating, I have elected not to rely very much on the Mdo dgongs 'dus as early testimony for winged herukas.

distinctive regional manifestations. Some aspects will be discussed in the course of this paper.

Some scholars have sought to understand this in relation to supposed diffusion patterns involving cognate avian symbolisms found in Iranian, Central and Inner Asian, or Chinese cultural contexts, but that is not my concern here. Likewise there is much debate on classifications of non-Buddhist Tibetan religions, differentiating Imperial religious rituals and beliefs, ancient localised forms of religion, religions witnessed in Dunhuang sources, and from other old textual sources such as dGa' thang bum pa, g.Yung drung Bon, Huber's category of 'bon-designated' practices, contemporary regional traditions, cognate trans-regional ritual forms both ancient and modern, and so on,8 but that need not concern us either. For present purposes, all that matters is to gain some idea of ritual practices possibly encountered by the earliest transmitters of Buddhist heruka traditions to the Himalayas and Tibet.

Much evidence is still awaited, as scholars have not yet published relevant research. Berounsky is currently extending Karmay's work (Karmay 2010) on the gNyan 'bum,

8 These are very difficult questions, which are not yet resolved. Some scholars (for example Samten Karmay) have seen a certain degree of cognate or even shared ritual culture across significant stretches of the Tibetan Empire before the mid-8th century, albeit alongside pervasive regional variation; hence old non-Buddhist ritual texts like the Kin 'bum or gNyan 'bum explicitly claim to incorporate materials from several languages (Zhang zhung. Sum pa, sPu rgyal bod, Nam pa ldong, etc), but still manage to weave these sources into a seamless whole.

Others (such as Toni Huber) suspect such a vision is partly an ideological construct seiwing to legitimise traditional g.Yung drung Bon and modem Tibetan nationalism. Thus Huber emphasises an overwhelming degree of regional ritual differences before the mid-8th century. He argues that before that time, across the regions we now know as ethnic Tibet, although cognate ritual patterns might be clearly discernible to anthropologists such as himself, they were not in any way recognised as shared ritual traditions by the various peoples who practiced them (personal communication, 11th May, 2016). However, a problem with this approach is that it too can be criticised by its opponents, as potentially too heavily weighted towards deconstructionism.

It will take some time before the true situation can be firmly established, partly because existing evidence is sparse and unreliable, and partly because the situation also probably varied over time as well as regionally. But perhaps as great a problem as the lack of evidence, might be the problem of scholarly gaze: observing exactly the same data, some scholars might choose to focus on similarity, while others might choose to focus on difference. For example, some might argue that large elite tumulus burials (attended by sacrifice) were indeed constructed right across the plateau; while others might respond that we have not yet done enough archaeology to ascertain if or to what degree those tumuli truly belonged to cognate ritual traditions. Some might remark that deity’ classes such as gnyan and btsan and religious notions such as sa bdag seem to have been recognised beyond individual regions; while others could respond that they were understood differently in different local cultures. Some might argue that ritual cults using avian symbolism in quite similar ways were quite pervasive, yet others could instead draw attention to the regional differences in those avian cults. Some might assert that ideas such as bla (life force) were very widespread indeed, while others could choose to emphasise how such ideas were understood differently in different areas. And so on.

Perhaps an analogy might be drawn with the multifarious religions of India: while having no original concept of being a single religion called 'Hinduism', and moreover being extraordinarily internally differentiated, it took Islamic and Christian colonialisms to create a new notion of 'Hindu', conceived as the sum of most (but not all) of the 'native' religions of India. But this process raises the question, was there truly a significant degrees of cognate or shared or intetextual religious or ritual culture in India, before Islamic and Christian colonialism? The answer to this question depends to a significant degree upon the scholarly gaze: to what degree should one privilege deconstruction and difference, and to what degree should one privilege similarity, intertextualitv, and interconnection? (for what it's worth, my own instinct is that one should privilege both gazes equally). In partly analagous manner, the very idea of g.Yung drung Bon surely only emerged in opposition to imported Buddhism, but the persistent question remains, to what extent did the pre-Buddhist ritual traditions across the regions of the Tibetan Empire already share recognisably cognate ritual and religious categories, and to what extent were they, as Huber argues, in the eyes of their practitioners at least, mutually unrecognisable and highly divergent regional traditions?

a massive and important non-Buddhist ritual compendium extant by the 10th century in both west and east Tibet, which is ritually obsessed by birds (personal communication 3rd May 2016). Heller is currently investigating avian symbols that crowned ancient gold funerary masks from west Tibet, and at T'ang descriptions of priests with feathered head-dresses who attended the Tibetan emperors (Heller 2016). Huber has not yet published his magnum opus on the 'shamanic' cults that still survive across vast stretches of the Himalayas and East Tibet, which he shows to have had ancient roots, and for whom avian symbolism was absolutely fundamental. Many Dunhuang and other Old Tibetan texts with avian references still require analysis.

An occassional complication is that some early avian symbolism had approximate Indian parallels, complicating questions of origins. For example, the countless Tibetan origin myths based on eggs, which appeared in numerous contexts and were the most widespread type of origin myth in Tibet, partially resembles the more limited Indian idea of hiranyagarbha (Kapstein 2006: 33).

Likewise Tibet's indigenous sacred bird, the horned khyung, became conflated with the Indian garuda, although retaining some of its far broader pervasion and ritual function. Khyung signifies an eagle, or other large powerful bird, and was one of the most important symbols of pre-Buddhist Tibet, with numerous ritual functions. It featured prominently on royal and priestly insignia whether real or imagined, for example on the crowns of the legendary early monarchs of Zhang zhung, called generically the bya ru can kings, meaning the kings with khyung feathered crowns (Vitali 2008). The term bya ru, literally 'bird or khyung horns', occurs several times in Dunhuang texts, to indicate bird feathers placed ritually on the head (Huber 2013: 278-9; Bellezza 2013: 69 note 90; 230).9

Huber's research shows that across numerous ethnic groups in widely dispersed areas across the Himalayas and East Tibet, non-Buddhist priests still continue to place bird feathers in the same way, using exactly the same terminology of bya ru (Huber 2013: 279), and always associated with the mobile life-principle, or bla. The implication is that these traditions are relics of a ritual system that is very ancient, and once existed across the entire area (Huber 2013: 279).

The Old Tibetan Chronicles found at Dunhuang (lines 37-51; see Hill 2006) associates the Tibetan royalty with bird-like ancestors, and Brandon Dotson tells me birds and their wings also feature prominently in the Dunhuang divinatory texts (personal communication, November 4th, 2015). Any search through Dunhuang non-Buddhist materials will throw up numerous bird references, many of them still little understood. 9 Huber writes: "For example, PT 1136, 28 has dbul [>dbu la] bya ru khyung ru ni btsugs for a horse as the subject; of. also PT 1134, 118: glad la ru btags sna bya ru 'ong ’ong', PT 1194, 45: cha yang gsas kyi glad la bya ru khyung ru 'ong ’ong, all cited from versions at OTDO. This meaning does crop up in the context of g. Yung-drung Bon, for example bya ru used as a head-piece of a priest in the Rgyal rabs bon gyi 'byung gnas', see Martin 1991: 125, Martin 2001: 195, and Vitali 2008: 388-392. See Zhangzhung'. 164, bya ru = bya khyunggi rwa." (Huber 2013: 279, note 54)

According to Dan Martin's PhD thesis (Tibeto-Logic blog, June 16 2010, 'Birdhoms'), in other contexts, the term bya ru can refers to horn like protuberances on the head or headgear that are not literally feathers; yet Huber's contemporary fieldwork reports the term used literally to mean feathers. Thus he writes: "Dan Martin and Roberto Vitali have both investigated the g.Yung-drung Bon context and concluded that there it refers to a type of finial ornament or symbol atop a mchod-rten shrine, and/or a crown ornament for rulers....Be that as it may, the term bya ru in the context of Srid- pa’i lha Bon always only refers to bird’s feathers “planted” {btsugs, the invariable verb here) directly upon the top of the head.." (Huber 2013: 279). Similarly Bellezza's philological analysis of old texts likewise understands the term literally to refer to feathers placed on the head or headgear (Bellezza 2008: 506-7; 2013: 69 note 90; 230).

In the more recently discovered early non-Buddhist Mokotoff manuscript, (Bellezza 2013: 15-114), we find a distinct focus on birds, notably in relation to ancient funerary rites. This text has many beautiful illustrations of birds, bird wing wands for ritual use, feathered headgears, and other items with avian ritual symbolism.

The avian theme continues into later canonical g.Yung drung Bon scriptures such as the Ka ba nag po, root text for the extensive Bon Phur pa tradition, revealed by a 12th century native of lHo brag, Khu tsha zla 'od. Its entire and important Chapter Six is dedicated solely to hawk deities, and a great deal of further avian symbolism is contained within other chapters too. Its mandala has fifty winged deities of various types. By comparison, the Phur Chen of Sa skya pa bSod nams rgyal mtshan (1312-1375), the principal ritual text of the 'Khon lugs Phur pa of the Sa skya school, which claims to be the oldest Buddhist Phur pa tradition, does not mention wings at all in the main descriptions of the iconography;10 nevertheless, despite the lack of textual support, the several officially commisioned thangkas do show a more modest sixteen pairs of wings (see for example HAR item number 73510). rNying ma gter ma Phur pa traditions can record slightly more than sixteen pairs, but I do not recall one that has as many as the fifty pairs of the Bon Ka ba nagpo.

Toni Huber (personal communications, November 2015) offers a geographical perspective to avian symbolism that is significant for my purposes, because it specifies areas of Southern Tibet and Bumthang which the Dunhuang text PT44 describes as among the first to receive the Buddhist heruka traditions from India. In contemporary East Bhutan, we find the gShen priests of the local Srid pa'i lha Bon tradition, which Huber (forthcoming) argues was very likely the ancestral religion of the Shar Dung populations in southern lHo-brag, prior to their mid-14th century dispersal southwards. Huber reports that birds, bird identities, bird metaphors, bird mimicry and bird body parts are absolutely central to the cult of these 'shamanic' priests (the term 'shamanic' is of course highly contested, but for reasons of convenience, I shall without further analysis follow Huber's well considered usage, which is shared with other noted anthropologists of the Himalayas such as Michael Oppitz). Huber adds that wherever one finds 'shamanic' ritual specialists all along the eastern Himalayas and along the far eastern marches of the Tibetan Plateau, the same avian-themed cultural patterns are found, and that these share a range of cognate features with Siberian 'shamanic' ritual specialists across north Asia. The pattern in all cases is that the shaman/priest is a "bird" who can fly to the upper world, and is thus always marked in one way or another with bird attributes of various kinds, above all feathered headgear and magic wands made of a bird wing or of bird feathers. He emphasises that these parallels constitute a distinctive set of far-reaching continuities which occur all along the eastern Himalayas, the eastern Tibetan Plateau Marches and up into Siberia; and that this set has a wide collection of parallel reference points, not merely individual isolates. In his forthcoming book, Source of Life. Bon Religion in East Bhutan and the Mon-yul Corridor, Huber will demonstrate in detail the continuity of specific content between the ancient non-Buddhist Dunhuang texts, dGa'-thang Bum pa manuscripts, 15th century Tibetan g.Yung drung Bon texts, and the present-day manuscripts used in the Srid-pa'i lha Bon cult of Eastern Bhutan. He writes, "There is some serious demonstrable time depth here, and I am convinced we are looking at very long-standing patterns" (personal communications, November

10 The visualisation of the root mandala in folio 15 mentions no wings; likewise the section on praising and enjoining in folios 35b-36a, which precisely describes the deity form, describes no wings. Nevertheless, the thangkas claiming to be made exactly in conformity with the text of the Sa skya Phur chen, do show sixteen pairs of wings.

2015). In other words, Huber's analysis adds powerful evidence from comparative ethnography, to the conclusions already arrived at by half a century of philology, suggesting that avian symbolism was fundamental to the religious traditions encountered by Buddhism when it first arrived in Tibet and much of the Himalayas. This seems to have been equally the case for elite and popular traditions.

Bird's Wing Wands

To negotiate this vast and little known terrain, I am therefore going to choose just one item from pre-Buddhist belief—not entirely at random, because it is one which has left us some vestiges of textual evidence. I have no way of proving it was in any way a relevant antecedent for the herukas wings, nor do I envisage any monocausal hypothesis that just one item inspired the wings of the rNying ma heruka, the origins of which might ultimately prove unknowable. In introducing it here, my main aspiration is to stimulate wider discussion of the so far inadequately explored but important field of the impact of non-Buddhist Tibetan religion on Buddhist tantrism. To move the debate forward at this juncture calls for an unusual departure from my normal practice of philological wariness and caution, and the taking of some intellectual risk, and in that spirit, I propose this hypothesis.

This item is the bird's wing wand. Huber points out that shamanic ritual specialists right across the Eastern Himalayas, across the Eastern Tibetan marches, and even up into Siberia, often carry ritual wands made either from a complete bird's wing, or from bird's feathers. The reason for this is that most such ritual specialists have bird beings to act as their magical auxiliaries, so that the wands are used to represent these auxiliary beings and their powers. Huber has documented the continuing use of a bird feather wand in local Tibeto-Burman-speaking highland communities throughout the extended eastern Himalayas (Huber 2013: 282-283).

His upcoming magnum opus Source of Life. Religion in East Bhutan and the Mon-yul Corridor is set to argue that this and related ritual symbolism reveals a close continuity with pre-Buddhist Tibetan religion, a conclusion based on the comparison of his many years of ethnographic research with the evidence preserved in a number of Dunhuang texts. He presents plentiful evidence to connect the contemporary practices, with ancient literary sources. For example (Huber 2013: 281, fig. 9), he describes priests from Arunachal, who carry bird feather wands during ritual performances11 which compare interestingly with a deity wielding two bird feather wands, from an illustration in the 11th century Mokotoff funerary mansucript (Bellezza 2013: 34).12 11 In this case, that must hang over the right shoulder and down the back, as if providing rear wings la

the body ] 12 For example (Huber 2013: 281, fig. 9), he describes priests from Arunachal, who must always carry bird feather wands during ritual performances. They must hang over the right shoulder and down tha back, as if providing rear wings to the body. This compares interestingly with a deity wielding whal appears to be two bird feather wands, from an illustration in the 11th century Mokotoff funerary Biansucript (Bellezza 2013: 34). The wands are described as dbal, a widespread Bon description for | lharp or powerful ritual tool. It is interesting that in the main root tantra of the Bon Phur pa tradition. Ble probably 12th century Ka ba nag po, the phur pa, which is likewise described as dbal, can be made Df bird feathers, which is not generally the case in its Buddhist counterparts. Chapter 5 of the Ka ba nag po thus describes the main Bon Phur pa deity 'Brug gsas chem pa as follows: plipag druggpas gyog VgPis hyi dang go na/ 'dab chags titytng dang rgod Ityi khgnang nas/ gdug pa khra dang glag gi gshog phur byung/ '[Thus,] in the left and right hands of the first pair of his six arms, From out of the mouths of garuda and vulture bird species, Come phurpas [made from] dangerous hawks’ and vultures’ quills.'

The bird's wing wand must be quite old, because it is clearly attested in the Dunhuang texts PT. 1194 (Bellezza 2008: 506, 642) and PT 1289 where the great Bon guru 13 gShen rab holds 'wing feathers' in his right hand and a flat bell or gshang in his left. This theme is replicated in various other archaic or marginal sources such as an illustration of a funeral in the old Berlin gZer mig ms, which is a hagiographical source for gShen rab, and in Na xi texts from Yunnan and Sichuan.

Yet knowledge of it is also fully contemporary, because it occurs prominently in a current g.Yung drung Bon funerary texts Mu-cho ’i khrom- ’dur, where it is known as bya-gshogphyag-cha, or phyag-mtshan gshog-pau

In the words of Charles Ramble,

Birds are very important in Bon, as indeed are wings even when they’re detached from the birds. In most cases the wings aren’t just decorative but have a specified function. In the Mu cho ’i khrom ‘dur literature they - and birds - are used to transport the soul of the dead. In composite chimera-type divinities they are some form of weapon or have some other meaning, (personal communication, 19th November, 2015).

The existing textual work comes from Bellezza (2008), who presents the original Tibetan, and his own translations and interpetations. He believes, I think reasonably given the evidence from the Dunhuang text PT 1194, that the bird's wing wand was a ritual implement used in pre-Buddhist funerary rituals (Bellezza 2008: 429). Although difficult to understand, PT1194 does indeed describe at length the use of bird wings (gshog) as ritual implements, and apparently in the context of the dead and of their funerary ritual (gshirr, durlbduf).

Descriptively, we learn that the bird's wing wand was a complex instrument. It was essentially a bird's wing, which seems to have been augmented with a variety of ritually significant embellishments. In the Dunhuang text PT1194, these include incense (spos), lamps (sgrori), agate (mchong), patterned agate (gzi), and needles (khab) (a transcription of PT1194 is conveniently reproduced in Bellezza 2008: 642ff). In the extant g.Yung drung Bon canonical funerary textMu-cho ’i khrom- ’dur, we find a number of further items, but Bellezza claims some of the ritual adornments to the wing are common to both the Dunhuang text and the Mu-cho ’i khrom- ’dur, which, after examining both Tibetan texts, I found does seem to be the case.

While in the Dunhuang sources a vulture's wing (rgod gshog) alone is indicated, in the Mu-cho ’i khrom- ’dur the vulture wing is reserved for the funerals of honourable elders only, while other types of bird wings are prescribed for deceased persons of other (perhaps less prestigious) social categories (Bellezza 2008: 506ff). In the Mu-cho’i khrom-’dur, the wand is wielded by means of a handle, which is itself ritually significant (Bellezza 2008: 433), and its ritual properties are described in detail. Both Khenpo Tenpa Yungdrung commented: "just as in the cases below, where the phurpas are made from leg bones or meteoric iron, here the phurpas are made from quills of the wing feathers of these birds."

13 PT 1289: v3, 12: gshang dril chen na phyag ma g.yon na snams / gshog the ra ther bu ni phyag ma g.yas na snams /

14 Dan Martin discusses the provenance of the Mu-cho ’i khrom- ’dur in a recent blog. Its history is not yet clear, although he and Samten Karmay have so far found references to 19th and 20th century Treasure texts of that name. See: http://tibeto-logic.blogspot.co.uk/2015/12/an-archaic-book-of-dead.html. Bellezza argues at some length that the funeral rites of the Mu cho’i khrom ‘dur still retain significant features found in earlier funerary sources, such as Dunhuang texts (Bellezza 2008: 497-499).

sources alike describe how the different sections of the wings were held to have different ritual properties (for example, PT 1194 differentiates gyas, g.yon, 'og, etc). Conceptually, Bellezza believes that for the Mu-cho ’i khrom- ’dur and the Dunhuang text alike, the bird's wing wand had a function of showing and conveying the deceased person along their path, and clearing their way from obstacles such as harmful spirits; and indeed, the word for path (lani) does occur frequently in PT 1194. However, the accurate interpretation of Dunhuang texts of this sort is notoriously difficult, so that we can expect many years of continuing debate on what PT 1194 precisely intends to say. Narrative recitations, or smrang, have been preserved in both the Dunhuang source and the Mu-cho ’i khrom- ’dur, and in both cases, the figure of 'Dur gshen rma da plays a significant role (for both Tibetan texts, see Bellezza 2008: 611 and 643).

Possible connections between the bird's wing wand and the heruka's wings

Now as I said before, I have no idea at all if the bird's wing wand was in any way directly relevant to the rNying ma heruka's wings. But sometimes the work of a scholar entails taking risks and asking exploratory and speculative questions, if only to see where they might lead us. So what possible points of convergence might there be?

(i) From the Dunhuang text PT. 1194, which is a funerary text, we can see that a mortuary connection is shared quite prominently. The Buddhist heruka is first and foremost a kapalika deity, that is, a deity who inhabits a charnel ground, and who wears the uniform of the Indian charnel ground dweller. Being a charnel ground deity is absolutely central to his ritual and iconographic identity, and no mere detail. Correspondingly, PT. 1194 suggests that the bird's wing wand might be a significant ritual implement of the famous pre-Buddhist funeral ceremonies, while in the Mokotoff manuscript and the canonical Mu-cho ’i khrom- ’dur, it is undoubtedly that.

(ii) As Charles Ramble points out (see above), certainly in the Mu-cho ’i khrom- ’dur, and possibly also in the Dunhuang funerary text, the bird's wing wand is primarily associated with transferring consciousness to a better world. This idea can work as a metaphor for the Mahayoga soteriology of the rNying ma heruka, whose main function is to enable the practitioner to let go of mundane appearences, to experience existence as divine.

(iii) It seems also to be the case that the bird's wing wand of the Mu-cho ’i khrom- ’dur and the Heruka's wings can be associated with summoning magical helpers or auxiliaries. As we have mentioned above, Huber observes that his Himalayan ritualists have birds or bird-like beings to act as their magical auxiliaries, and wing representations or feather wands are used to represent these magical auxiliaries and their powers.

This might also be what is intended in the Dunhuang text on the bird's wing wand, PT1194. For example, in verse (iv) we are told 'these were the wings of the lha and sras [[[divinities]]]' (te lha dang gsas gyi gshog). In verse (vii), the final verse, we again find the deities being referenced: lha gsas gyi nan thu yang che, but like Bellezza, I find this verse too obscure to be easily translateable. (Bellezza 2008: 509-510).

The bird's wing wand of the Mu-cho’i Khrom-’dur at one stage also seems to be associated with auxiliary deities, and as an instrument to summon the gods. To achieve the summoning, a particular ritual gesture is also important. According to Bellezza's understanding (2008: 430-432, 623-4), by raising the divine wing (gshog bzhengs), the armies of allied deities are marshaled (lha dmag mangpo 'ub kyi 'dusj, these will protect from (srung) demonic obstacles (bgegs).

One of our two earliest rNying ma sources for winged herukas quite closely concurs with this gesture of raising the wing, and its function of causing auxiliary deities to appear that will battle all demonic entities. The early Sa skya patriarch Grags-pa rgyal-mtshan (1147-1216) propagated a rNying ma text he said he got from his father Sa-chen Kun-dga’ sNying-po) (1092-1158), the rDo rje phur pa'i mngon par rtogs pa, which they claimed descended to them through a direct lineage from Padmasambhava. Here each of a series of ten winged herukas have magical zoocephalic helper emanations on either side, who are miraculously made to appear by means of the raising of the male heruka's wings. Thus magical auxiliaries with fangs are made to appear when the herukas raise their right wings and winged magical auxiliaries with claws or talons are made to appear when the herukas raise their left wings.15 (Grags pa rgyal mtshan: 1968: 179; 1992-3: Vol. 8 pp 723-724).

(iv) Another interesting comparison can be made between the handles of the bird's wing wand, and the dang ru or 'collar bone', the point at which the rNying ma heruka's wings are joined to his body. Both are ritually very significant. In the Mu-cho’i Khrom-’dur, for example, the handle of the bird's wing wand is described as the meeting point of heaven and earth. In rNying ma texts the 'collar bones', or the points at which the heruka's wings are joined to his body, are often specified as separate symbolic entities, with their own ascribed spiritual and iconographic qualities. It is not impossible that we might find some cognate parallels here (but this must be approached with due caution, since I am not aware of these handles in the Dunhuang text, and the Mu-cho’i Khrom-’dur might have ben influenced by rNying ma iconography).

(v) In the Dunhuang, later g. Yung drung Bon, and rNying ma sources alike, the wings are highly differentiated according to situation and function. The Dunhuang text PT 1194 differentiates between right wings and left wings, and between white, black, and yellow vulture wings, to be used for different versions of wands appropriate for different occassions. Similarly, the Mu-cho ’i Khrom- 'dur specifies twelve different kinds of birds whose wings should be used for persons of different social categories. Likewise, the rNying ma texts differentiate their heruka's wings according to various ritual criteria. Thus Nyang Rai in his bDe gshegs 'dus pa Phur pa root tantra describes the right wing as made of vajras, and the left wing of jewels.16 This definition is now very widespread, some would say, the standard representation. Yet in fact it is by no means universal. For example, bDud 'joms Rinpoche in his influential Gnam Icags spu gri gives the left dangru or collarbone as made of jewels, and the right collar bone of vajras, while the left wing itself has sword feathers, and the right wing has razor feathers. Mag gsar in his Phur pa'i rnam bshad describes the herukas of the five buddha familes as having wings made of of chakras, vajras, ratnas, crossed vajras, and lotuses, while the four or five 'son' kilas have their left collarbones made of jewels, and their right collar bones made of vajra, but with vajra wings on the right, and razor wings on the left (Mag gsar 2003: 224). There are a great many other such variations too. 15 Grags pa rgyal mtshan 1992-3: Vol. 8 pp 723-724): pho nya niche ha can rnams g.yas kyi gshog 'clegs pa'o/ sder mo can thams cad yah kyi g.yon gyi gshog 'degs pa'o. For a further discussion of these zoocephalic deities, and their possible relationships to Bon, see Cantwell and Mayer 2015. 16 See folio 345v, line 5-6, of his bDe bar gshegs pa thams cad kyi 'phrin las 'dus pa phur pa rtsa ha'i rgyud, chapter 1, in Volume Ya of the sGang steng b NGB (photo 230, 3rd page).

(vi) More generally, as Huber points out, for the contemporary ritual specialists in his ethnographic studies, wings represent the Bon 'shaman's' power of flight up into the sky and back, to fetch the sky deities.17 In the ancient Dunhuang sources studied by Bellezza, although difficult to translate, it seems likely that the wings serve as weapons to banish evil forces, while conveying the soul of the dead to good destinations. Thus Bellezza describes the bird wing wand as "that powerful ritual implement used as a weapon to dispel demons that attack the dead and as a viaduct for the regulation of their consciousness principles" (Bellezza 2008: 506); and "as instrumental in forging a physical pathway to heaven" (Bellezza 2008: 507): while I find Bellezza's soteriological language of 'regulation of consciousness principles' and 'heaven' open to question, I think we can at least agree with him in the broader ritual aspects, that the wings were believed to dispel demons that attacked the dead, and help convey the dead to suitable destinations. As Charles Ramble likewise points out, wings function to transport the souls of the dead to desirable destinations, while also (especially in the numerous Bon sources describing composite chimera-type divinities), functioning as weapons (personal communication, 19th November, 2015).

All these functions are entirely acceptable symbols for a tantric Buddhist heruka.

Part Two

Earliest evidence for winged herukas

If one asks when wings first appeared on Buddhist herukas, the answers are intriguing. As far as I am currently aware (of course, more evidence might still come to light), the very earliest evidence for a winged Buddhist heruka is the Dunhuang text IOLTibJ321. This 85 folio manuscript contains the famous rNying ma tantra, the Noble Lasso of Methods, or Thabs zhags, embedded as lemmata within its word-by-word commentary. Paleography dates this manuscript to the mid tenth century (van Schaik 2013: 127), while text critical analysis of its scribal errors has shown that the Dunhuang manuscript itself was already the result of at least two cycles of recopying (Cantwell and Mayer 2012: 26-67). Statements in the main body of its commentary, supported by others in the marginal notes, seem to indicate that the root tantra itself (as opposed to the commentary) represents the revelations of Padmasambhava (Cantwell and Mayer 2012: 91-98). And here, it seems for the first time, we find the winged heruka described in Chapters 12 of the root tantra.18 (Cantwell and Mayer 2012:171-172; 293-300). 17 'RituaI practice consists of the accurate and systematic chanting of the rabs together with the actions of rites often specified in the rabs, all aimed at bringing the deities down from the sky into a highly purified environment, hosting them, gaining powers from them, and dispatching them upwards once more. This process can often involve an elaborate verbal ritual journey undertaken by the priest up to the thirteenth level of the sky and back, in order to invite and escort the deities. The main aspirations of participants/sponsors during ritual is to gain various life powers directly from the sky deities while they are temporarily on earth.' (Huber 2013: 264). 18 Folio 49r. 6, describing the central heruka: Idbus kyi khrag thung chen po ni / tdbti dgu phyag kyang bco brgyad ste! tgshog pa rdo rje gtams pa gdengs /; and in the next verse, describing his attendent herukas: [49v.4] Ide 'i dkyil 'khorkhro bo ni I ldbu gsum phyag [5] drug gshogpa can / Yet it is interesting that another Dunhuang text of similar antiquity, (IOL Tib J 306), which is unattributed to any author, also describes a rNying ma style heruka with three heads, six arms, and four legs, but makes no mention of any wings.

More suprisingly still, nor does the famous Guhyagarbha tantra, generally considered the most influential source for the entire rNying ma Mahayoga tradition. It too describes only the three heads, six arms, and four legs, without mentioning any wings. This is surprising, because the Guhyagarbha tantra is considered the root text for the hundred peaceful and wrathful deities, which as I have already described, constitute one of the fundamental groupings of herukas within the rNying ma pantheon, and which are very often (although not always) represented as winged, throughout the voluminous art and literature. Clearly, the wings that are so often associated with the Guhyagarbha'?, famous set of heruka deities, do not derive directly from the main Guhyagarbha root tantra, but from some other, related source.19 Like IOL Tib J 306, the Guhyagarbha root tantra is not attributed to Padmasambhava.

The other earliest source for a winged rNying ma heruka, which we have already mentioned above, is the rDo rje phur pa'i mngon par rtogs pa, one of the oldest Sa skya texts on Vajrakilaya, which is claimed to originate with 'Khon klu'i dbang po's receipt of the transmission directly from Padmasambhava.20 And this rDo rje phur pa'i mngon par rtogs pa does indeed describe wings.21

The significant point here might be that our two earliest proven sources for the winged heruka, the Dunhuang Thabs zhags manuscript and the rDo rje phur pa'i mngon par rtogs pa, both claim Padmasambhava as their originator. This is in contrast to the two other, equally old sources for the rNying ma heruka, PT306 and the Guhyagarbha tantra, which neither have wings, nor any claimed link with Padmasambhava.

The feather in Padmasambhava's hat

What is more interesting still is that these two earliest known sources for the winged heruka, the Thabs zhags and the earliest Phur pa tradition, are portrayed in the Dunhuang literature not merely as descending from Padmasambhava, not merely as being transmitted by him, but, rather, as being redacted or even composed by him. I should also mention, these are also two out of only three tantric traditions credibly attributable to Padmasambhava; the third, the Man ngag Ita 'phreng, does not describe any iconography, so we don't know if it too envisaged winged herukas.

As already described above, Padmasambhava's involvement in its redaction or authorship is stated explicitly in IOLTibJ321, where he seems to be said to be the producer of the Thabs zhags root tantra (Cantwell and Mayer 2012: 91-98). It is likewise made abundantly clear in PT44, the well known Dunhuang rabs (indigenous

19 I browsed through Klong chen pa's Guhyagarbha commentary Phyogs bcu mun sei, but have not so far found wings there either. 20 For this reason, the 'Khon lugs Phur pa is generally considered the oldest and most faithful bka' ma tradition of Phur pa in Tibet: by contrast, most of the rNying ma traditions are considered gter ma, or otherwise, bka’ ma reconstructions of a later date, based on the NGB texts. 21 This is in agreement with many of the rNying ma canonical tantras upon which the Sa skya ritual tradition claims to be based, which do tend to describe the heruka as winged: see for example the Phur bu myang ’das chapters four and nine, and the rDo rje khros pa chapter three (Cantwell and Mayer 2007: 135, 153, 244). Yet what is fascinating here is that later Sa skya texts, such as the vast and hugely detailed 17th century Phur chen liturgy and commentary by 'Jam mgon A myes zhabs (1597— 1659), do not describe any wings. Phur chen p. 15a.4 to 15b.5 describe the main deity, and no wings are mentioned. Amyes zhab comments on this passage (page 284 lines 2 to 3) and here also does not mention any wings. Nevertheless, officially commisioned and iconographically correct Sa skya thangkas do still typically show sixteen pairs of wings, as we saw above.

Tibetan style ritual narrative) that narrates the origins of the Phur pa tradition:22 23 so abundantly, that the first modem study of this text, Bischoff and Hartman 1971, bore the slightly exagerrated title, Padmasambhava's Invention of the Phur-bu: Ms. Pelliot Tibetain 44. More precisely, the main thrust of PT44 is to describe how Padmasambhava achieved a comprehensive re-redaction of the KTlaya materials he had procured from Nalanda, adding extra protector deities to its mandalas, and rearranging its texts.

Yet, intriguingly, we also get the impression that these redactions have been specially made to suit the needs of Padmasambhava's Himalayan dispensation, because even if this is only suggested implicitly for the Thabs zhags, it is quite explicit in PT44. A canonical NGB Phur pa tantra, the Phur pa bcu gnyis, the title of which is mentioned in Dunhuang sources, offers a kind of corroboration to the narrative of PT44, by including within two of its chapters, the main redactive innovation that PT44 describes Padmsambhava as making to the KTlaya tradition he received from India: the introduction of new specifically Himalayan protector deities. Likewise, in similar vein, PT307 describes Padmasambhava incorporating into the tantric Buddhist pantheon the indigenous deities that, it would appear, later became known as the brtan ma bcu gnyis. Cathy Cantwell and I have already discussed these issues at length in previous publications (Cantwell and Mayer 2013; 2009), so I won't go into detail here. I should reiterate, however, that these two mandalas with winged herukas, the Thabs zhags and Phur pa, are also the only two Tantric Buddhist systems

22 PT44, like the structurally similar Dunhuang Padmasambhava text PT307, seems to represent the adoption by the early Padmsambhava school of the distintively indigenous Tibetan ritual structure of the smrang or rabs. I do not know of such a structure from Indian sources and am uncertain if it exists on a comparable scale within the gSar ma traditions; nor have I so far found it elsewhere in Dunhuang Buddhist texts other than in these two examples relating to Padmsambhava. However, it became a standard strategy within later rNying ma, which continues to use the very same snirangtrabs narratives described in PT44 and PT307, as well as other similar smrangtrabs. See Cantwell and Mayer: 2009. 23 For a detailed analysis of the various points in the text where Padmasambhava is mentioned as producer of the Thabs zhags, see Cantwell and Mayer 2012: 87-98, especially 91-98.

In its redaction, the Thabs zhags root tantra (IOLTibJ321) took an already well known more exoteric Yogatantra mandala of deities as its basis, which it then develops into a more esoteric Mahayoga system by introducing female consorts for the male deities, a separate mandala of wrathful female deities, and of course, the winged male heruka. An appropriately placed marginal note then cites the famous Indian master of the more exoteric Yogatantras, Santigarbha, who himself visited Tibet at the request of the Tibetan state, where he translated a famous Yogatantra text that was extremely important in the late Imperial period, the Sarvadurgatiparisodhana tantra, and, according to some sources, also consecrated bSam yas monastery. Santigarbha is described as commending and approving the new text, which constitutes a further more esoteric development of the Yogatantra tradition with which he is associated. And Santigarbha respectfully attributes this development to Padmasambhava.

What makes these marginal notes historically interesting is that the late 8th century, the time when Santigarbha and Padmasambhava lived and taught, was indeed the epoch of the advent of kapalika-flavoured tantrism within Buddhism, for which the Thabs zhags is a classic early example. The late 8th century was thus precisely the period when Yogatantra materials were increasingly beginning to be redacted and adapted, most specifically to include the distinctive kapalika elements that were henceforth to transform them into the variously called Yogottara, Mahayoga, Yogirn and Yoganiruttara materials of the more developed Vajrayana. And a striking and salient feature of this process of transformation was its bold adoption and utilistation of non-Buddhist materials, both Sanskritic and non-Sanskritic, in some cases to an extent seldom known before in Indian Buddhism: indeed kapalika iconography itself represented a wholesale adoption by Buddhism of an originally Saiva symbolic system. So it is not impossible that in the very same period that his contemporaries further south were busy adapting Saiva and other non-Buddhist symbolisms in India to Tantric Buddhist uses, Padmasambhava (or his school) were engaged in a parallel process in Nepal and Tibet, adopting and adapting Himalayan avian symbolisms to Tantric Buddhist uses.

specifically attributed to Padmasambhava's redaction within the entire Dunhuang literature, and they are simultaneously our two earliest known sources for the winged Heruka.

How are we to understand this Padmasambhava connection? What might make sense of it is the ingenious historical hypothesis proposed by Matthew Kapstein (2000: 155-160), which has the great advantage of fitting all known demonstrably old sources of evidence for Padmsambhava: that is, the Dunhuang sources we have just described, the various versions of the Testament of Ba, and the official Imperial lexicographical text, the sGra sbyor bam po gnyis pa. On the basis of their evidence, Kapstein proposes that Padmasambhava was a late 8th century Indian teacher of esoteric tantrism who taught primarily in the Himalayan regions and the south of Tibet (this because PT44 specfies Yang la shod in Nepal, Bumthang in the modem Bhutan, and Lhobrag on the boundaries of modern Bhutan and Tibet, as areas of activity of Padmasambhava's students, although the cave at Brag dmar near bSam yas is also mentioned). However, Kapstein suggests, Padmasambhava was quite likely not part of the official Tibetan imperial conversion program centred on bSam yas and central Tibet, even if he visited there briefly. Nevertheless, his independent following eventually became powerful enough that his tradition was within a generation or two incorporated into the official versions of Buddhism, and increasingly so after his style of radical tantrism began to wax so fashionable in India and Tibet after the 9th and 10th centuries (Mayer 2016).

An additional feature of Kapstein's theory is that it places some of the early Padmasambhava school's main spheres of religious activity in the regions where Huber's Bon priests still practice a bird-oriented shamanism of apparently great antiquity. Lho brag, an area specified in PT44 as an early location of Padmasambhava's tradition, is also where, some centuries later, Nyang ral and Chos dbang were to so greatly magnify Padmasambhava's cult, along with its avian iconographies, codifying it into formulations that set the pattern for the later rNying ma pa. Lhobrag and neighbouring Bumthang, which is also mentioned in PT44, are also the areas where their great near-contemporary, the polymath Khu tsha zla 'od, produced his own densely feathered Bon Phur pa mandala that I have mentioned above, the Ka ba nagpo.

PT44's citations of Lhobrag and Bumthang, when correlated with Huber's findings, does indeed yield highly specific evidence that the early Padmasambhava tradition coexisted alongside indigenous traditions with prominent avian imagery (but as we have seen, it is also reasonable to infer a geographically much wider prevalence of avian imagery).

The question then emerges, might Padmasambhava's (or the early Padmasambhava school's) attempts to engage with the non-Buddhist rituals of this (and other) areas some centuries earlier, account for the avian elements evident in the traditions linked to Padmasambhava in PT44, IOLTibJ321, and the earliest Sa skya Phur pa literature? This is not implausible. Amongst the most salient features of Indian tantrism of the period was a propensity to absorb and adapt useful elements from other ritual traditions.24 The prevalence of non-Buddhist elements within Indic Buddhist tantras is by now very well known. Likewise I note above (footnote 5) that Teun Goudriaan and Sanjukta Gupta (1981: 115) mention a winged Saiva Bhairava, Akasabhairava, which Sanjukta Gupta tells me she believes is of Himalayan rather than Indian origins (personal communication, 16th November 2015).

In this regard, we seem to find an even more suggestive piece of evidence, in the form of Padmasambhava's enigmatic magical hat. In the classic Tibetan iconography, Padmasambhava sports a vulture's feathers in his lotus hat, described as a sacred headgear imbued with extraordinary miraculous powers. 25 And from Padmasambhava, this trope continues through subsequent rNying ma Buddhism: the Bhutanese Raven Crown is a well known example, and related surely are the vulture feathers that miraculously sprouted from the head of Rig 'dzin rgod kyi Idem 'phru can (1337-1409), which were said to be a sign of Padmasambhava's blessing. But as far as I am aware, a feathered hat is not witnessed anywhere else within Indian siddha iconography, nor am I aware of quite such an emphasis on a hat's miraculous powers within wider Indian Vajrayana symbolism. Yet as Huber and others have pointed out, the wearing of such a magical feathered headgear is ubiquitous amongst Himalayan shamans, arguably constituting their most distinctive insignia. It is also a tradition of great antiquity, traced to T'ang descriptions of Tibetan royal sorcerers (wu) wearing bird hats (Xin Tang Shu. 130f, as translated in Pelliot 1961, cited in Walter 2009: 33), witnessed in surviving early Tibetan sources, such as the Mokotoff ms., and in legend back to the legendary bya ru can monarchs of ancient Zhang zhung.

More remarkably still, PT 44 presents us with a curious narrative that has for many years baffled Tibetologists and until now resisted all analysis, not least because it is no longer part of the rNying ma narrative tradition. This narrrative tells us that Padmasambhava actually stored the powerful protective deities within his magical hat, keeping them there for prolonged periods of time, where they remained firmly under his control and from where he could let them out when required. More specifically, the protective deities that he kept in his hat were the local Himalayan goddeses that he had tamed at Pharping in Nepal, and which he had then introduced into his reformulation of the VajrakTlaya mandala he had received from Nalanda, as its new set of worldly protector deities. Such a ritual use of a hat, as a store for powerful worldly protective deities, is not something I have so far encountered elsewhere within Indian siddha narratives. It is not typical of Indian Buddhist symbolic thinking, where only Buddhas and other enlightened beings might dwell in the hat or head, but dangerous worldly protector deities are far more likely to go on the feet or arms. And while almost all the rest of the narrative of PT44 is continued within the later rNying ma tradition, this particular but prominent detail, of Padmasambhava keeping the tamed deities in his hat, seems to have subsided from view. Yet locating their deities in their headgear does very much seem to be a ubiquitous pattern within Himalayan 'shamanism' (and in the eighth century, it might well have been ubiquitous among more elite priests within the Tibetan empire as

25 Jeff Watt writes on the HAR website: "The vulture feather decoration at the peak of a lotus hat appears to have originated with the iconography of Padmasambhava sometime during or after the lifetime of Nyangral Nyima Ozer. A lotus hat can be adorned by one or three vulture feathers, a half vajra scepter, or left unadorned. The symbolic use of the feather seems to be almost exclusively used in the Nyingma Tradition of Tibetan Buddhism. A single feather is generally the norm but examples of three feathers both in painting and sculpture can be found for Padmasambhava and others. Godemchen (1337-1409 [P5254]) typically has three feathers protruding upwards at the crown of the hat. "(https://www.himalayanart.org/search/set.cfm7setIDM559) But Jeff Watt's statement needs some qualification: since we have no surviving sources for the iconography of Padmasambhava's hat earlier than Nyang ral, we simply cannot know what went before, and we certainly cannot simply assume that Nyang ral invented the feather in the guru's hat, merely because his writings are the earliest surviving sources for it.

well). As Toni Huber wrote to me recently (personal communication, November 25th 2015):

'the hat or better 'headgear' of all Himalayan shamans and many in Siberia is THE ritually potent item of costume. This is because it is the seat of the hereditary auxiliary spirits, as well as other deities the shaman moves around the cosmos during ritual journeys...My book will contain many pages (heavily illustrated) discussing aspects of shaman's headgear for these reasons, and feathers or bird wings feature on many, many Himalayan examples.26

Yes, the feather headgear represents a ritual defense against negative forces, but the material culture of feathers is only really framed in this manner because [feathers] are directly associated with the auxiliary spirits. Bya-ru [i.e. feathers] empowered by the auxiliaries of Bon shamans are still placed upon the crown of the head (and on the hat if a layperson wears one) of ritual participants in my research region to protect vitality for the same reason, as well as elsewhere in the eastern Himalayas where shamans still function (the Mishmi Hills, for instance). From this perspective, which is only ethnographic (although I have a description from 1260 of a lay religious specialist named Bal-po Ka-ru-'dzin who has a bya-spu inserted into his hat), Padmasambhava's, Gling Ge-sar rGyal-po's and other hats with feathers in them are nothing special, just another example of a very wide-spread form of shaman's headgear which, as sure as god made little apples, did not derive from siddhas in South Asia."

In conclusion, my admittedly speculative, but I hope nevertheless interesting, two-limbed hypothesis is as follows:

(i) That Padmasambhava might well have been, as Kapstein suggested, and as the Dunhuang text PT44 describes, an Indian siddha who was active in in Pharping and the southern slopes of the Himalayas, and southern Tibet, in the late eighth century; and whose early Tibetan followers were active in Bumthang, Lhobrag, and bSam yas;

(ii) and that in adapting his ritual teachings to the culture of his Himalayan target audiences—both the local 'shamans' he might have met in his periods of mountain retreat, and, beyond them, the grander priests of the Tibetan aristocracy and royal court—Padmasambhava (or perhaps only his early school, we cannot conclusively tell) might have redacted Indic tantrism to include elements of Himalayan and Tibetan avian symbolism, attaching bird's wings to his herukas, placing a bya ru or vulture feather upon the great guru's own hat, and in addition, displaying the quintessentially Himalayan ritual practice of storing his worldly protective deities within his headgear. As further evidence of adaptation to local cultures, we must recall that two of the three surviving early Padmasambhava texts from Dunhuang, PT44 and PT307, although unambiguously Buddhist in content, both appear to adopt or adapt the quintessentially indigenous ritual structure known as the smrang: and not only that, but the narrative of these two smrangs are dedicated to explaining how Padmasambhava incorporated sets of indigenous deities into the Buddhist pantheon. Might Nyang ral and Chos dbang's deep devotional connection with Padmasambhava, and their unprecedented magnification of him, be connected with the fact that his teachings had already flourished in their homelands of Lho brag for so many years? 26 As well as the examples Huber cites, Anne de Sales has recently sent me photos of Kham-Magar jhankris (shamans/priests) who wear turbans densely feathered with pheasant feathers. Among the Tamangs, peacock feathers are used.

Might this also be why the Herukas of their Treasure cycles so often wear wings, and also have the Tibetan sacred bird, Khyung, circling just above their heads? For in these characteristics, the influential herukas of Nyang ral and Chos dbang are distinct from those of their purely Indic counterparts. Bibliography

Tibetan sources:

Grags pa rgyal mtshan (Sa sky a rje btsun grags pa rgyal mtshan) Works on rDo rje Phur pa found in The Complete Works of Gragspa rgyal mtshan, in the Sa skya bka' 'bum, compiled by bSod nams rgya mtsho, The Complete Works of the Great Masters of the Sa skya Sect of the Tibetan Buddhism, Volume 4, The Toyo Bunko, Tokyo, 1968. In particular, rDo rje phur pa'i mngon par rtogs pa, Volume Nya: 355r-367v (=lr-13v in the separate pagination of this group of texts, found on pages 175-182 in the Western style bound book), Phur pa'i las byang 367v- 384r (=13v-30r in the separate pagination of this group of texts, found on pages 182-190) [both of these texts seem to be edited versions of Sa chen Kun dga' snying po's work], and rDo rjephurpa'i sgrub skor 385r-400r (190-8) [TBRC gives publication details of the Dehra Dun Sakya Centre printing, 1992-1993, reproduced from the sDe dge edition in 15 volumes, TBRC Resource Code: W22271.]

Mag gsar Kun bzang stobs ldan dbang pa, 2003. Phurpa'i mam bshad he ru ka dpal bzhadpa'i zhal lung (bcom ldan 'das dpal chen rdo rje gzhon nu'i 'phrin las kyi mam par bshad pa he ru ka dpal bzhad pa'i zhal lung). sNgags mang zhib 'jug khang (Ngak Mang Institute), Beijing: Mi rigs dpe skrun khang.

Nyang ral, Nyi ma 'od zer, n.d. bde bar gshegs pa thams cad kyi 'phrin las 'dus pa phur pa rtsa ba'i rgyud, sGang steng b NGB, Volume Ya, chapter 1, folio 345v (photo 230, 3rd page), line 5-6.

Nyang ral, Nyi ma 'od zer bKa' brgyad bde gshegs 'duspa'i chos skor, 1979-1980, 13 volumes. Paro, Lama Ngodrup, Kyichu Temple (reproduced from the complete mtshams-brag manuscript). An electronic version is available from the Tibetan Buddhist Resource Center, New York (W22247). Works specifically referred to: 1. bka' brgyad bde gshegs 'dus pa las: 'phrin las 'dus pa rtsa ba'i rgyud, Volume Ga: 321-387. [This is the same root tantra as the bde bar gshegs pa thams cad kyi 'phrin las 'dus paphur pa rtsa ba'i rgyud included in the various rNying ma’i rgyud 'bum editions, see above. It is also found within other compilations of the bKa' brgyad bde gshegs 'dus pa, in particular, 1978, 13 volumes. Gangtok, Sikkim, Sonam Topgay Kazi, Kah thog rdo rje dgan gyi par khang (W1KG12075), Volume Ga: 273-373; and 1977-1978, 4 volumes. Dalhousie, Damchoe Sangpo, reproduced from a collection from the library of Kyirong Lama Kunzang (W1KG9588), Volume 2: 257-309. Note that it is not found in a 2 volume bKa' brgyad bde gshegs 'dus pa

rituaI practice manuscript collection (B. Jamyang Norbu, New Delhi, 1971, W00KG09391), which includes much later material.] 2. bka' brgyad bde gshegs 'dus pa las: byin brlabs phun sum tshogs pa phur pa'i bsgrub pa bi ma las mdzadpa zhi ba'i mchog [yon] tan spo ba'i cho ga. Volume Ta: 351-365. 3. phun sum tshogs pa don gyi man ngag sa ma 'grel, Volume Ta: 367-375; 4. phun sum tshogs pa rgya cher 'brel ba, Volume Ta: 377-417); 5. Further associated materials, Volume Ta: 343-349, 419-503). 6. Medicinal

Accomplishment texts, Volume Nya: i. khro bo'i dgos bsgrub: bdud rtsi sman bsgrub kyi bsgom rim: 347-375; ii. mchodrdzas bdudrtsi'i bsgrub pa: 377-419; iii. bdud rtsi sman sgrub thabs lag khrid du bsdebs pa: 421-507. (See also edited versions and 15 compilations of his various texts in the works of gTer bdag gling pa; and in Kong sprul's Rin chen gter mdzod]

Sa skya Phur chen: dPal rdo rje gzhon nu sgrub pa'i thabs bklagspas don grub, by Sa skya pa bSod nams rgyal mtshan (1312-1375), dPal sa skya'i chos tshogs, Raj pur, India (Tibetan date given: 992). Two different electronic editions of the Sa skya Phur chen are available. One is included in Volume 18 of the dpal chen ki laya'i chos skor phyogs bsgrigs [[[si khron]] zhing chen mi rigs zhib 'jug su'o, bod kyi shes rig zhib 'jug khang, khreng tu'u 2002

(http://tbrc.Org/#library_work_Object-W24051), p,137ff, and one included in the rGyud sde kun btus, Volume 16. This itself has two available versions: Lungtok & Gyaltsan, Delhi 1971-1972 [W21295], p.427- 505; and Sachen International, Kathmandu 2004 [W27883], p.479-568. Both are apparently based on the sDe dge blocks, although the Sachen International version is an entirely remade computer input version edition rather than a photographic reproduction. Works in other languages:

Bautze, Joachim Karl, 1995. Early Indian Terracottas. Brill, Leiden.

Bellezza, John Vincent, 2013. Death and Beyond in Ancient Tibet. Archaic Concepts and Practices in a Thousand-Year-Old Illuminated Funerary manuscript and old Tibetan Funerary Documents of Gathang Bumpa and Dunhuang Verlag der Osterreichischen Akademie der Wissenschaften, Vienna.

Bellezza, John Vincent, 2012. 'The horned eagle: Tibet’s greatest ancestral and religious symbol across the ages'. Flight of the Khyung blog, entry for January 2012. Tibet Archeology website, as retrieved 20th March 2016: http://www.tibetarchaeology.com/january-2012/

Bellezza, John Vincent, 2008. Zhang Zhung Foundations of Civilization in Tibet A Historical and Ethnoarchaeological Study of the Monuments, Rock Art, Texts, and Oral Tradition of the Ancient Tibetan Upland. Verlag der Osterreichischen Akademie der Wissenschaften, Vienna.

Bischoff, F. A., and Charles Hartman. 1971. "Padmasambhava's Invention of the Phur-bu: Ms. Pelliot Tibetain 44." In Etudes tibetaines dediees a la memoire de Marcelle I.alou, pp. 11-27. Paris: Adrien Maisonneuve.

Buhnemann, Gudrun, 2008. Buddhist Iconography and Ritual in Paintings and Line Drawings from Nepal. Lumbini International Research Institute, Lumbini.

Cabezon, Jose Ignacio, 2013. The Buddha's Doctrine and the Nine Vehicles: Rog Bande Sherab's Lamp of the Teachings. Oxford and New York, Oxford University Press.

Cantwell, Cathy, and Robert Mayer, 2015. 'The Winged and the Fanged'. Pages 153-170 of From Bhakti to Bon. Festschrift for Per Kvoerne, ed. Charles Ramble and Hanna Havnevik. The Institute for Comparative Religion in Human Culture/ Novus Forlag, Oslo 2015.

Cantwell, Cathy, and Robert Mayer, 2013. ‘The mythologisation of Padmasambhava in post-imperial Tibet’, in Cuppers, Mayer and Walter, eds, Tibet after Empire. Culture, Society and Religion between 850-1000, Lumbini International Research Institute, Lumbini.

Cantwell, Cathy, and Robert Mayer, 2012. A Noble Noose of Methods, the Lotus Garland Synopsis'. A Mahayoga Tantra and its Commentary, Vienna, The Austrian Academy of Sciences Press.

Cantwell, Cathy, and Robert Mayer, 2009. "Enduring myths: smrang, rabs and ritual in the Dunhuang texts on Padmasambhava", in Pommaret, Frangoise and Jean-Luc Achard (eds) Tibetan Studies in Honor of Samten Karmay, Amnye Machen Institute, Dharamsala.

Cantwell, Cathy, and Robert Mayer, 2007. The Kilaya Nirvana Tantra and the Vajra Wrath Tantra: two texts from the Ancient Tantra Collection, Vienna, The Austrian Academy of Sciences Press, vi, 289 pages, plus editions of texts on cd (529 pages).

Dalton, Jacob, 2002. The Uses of the Dgongs pa 'dus pa'i mdo in the Development of the Rnying-ma School of Tibetan Buddhism. Unpublished PhD dissertation, Michigan 2002. UMI Number 3057935.

Dalton, Jacob, 2014. 'Preliminary Remarks on a Newly Discovered Biography of Nupchen Sangye Yeshe', pp. 145-163 in Andrew Quintman and Benjamin Bogin, eds, Himalayan Passages: Tibetan and Newar Studies in Honor of Hubert Decleer. Wisdom Publications, Boston.

Goudriaan, Teun, and Sanjukta Gupta. 1981. Hindu Tantric and Sakta Literature. Wiesbaden, Harrasowitz.

Heller, Amy, in press, 2016. “The role of Birds and Deer in Early Tibetan rituals: preliminary remarks”, in Tsering Thar, Samten Karmay and Amy Heller (eds), Zhangzhung and Early Tibet, Proceedings of the BeijingMinzu Daxue Seminar Sept 18-21 2015.

Hill, Nathan W., 2006. 'The Old Tibetan Chronicle — Chapter I', pages 89-101 in Revue d'Etudes Tibetaines (Langues et Cultures de l'Aire Tibetaines, CNRS, Paris), Vol.10, April 2006.

Huber, Toni, forthcoming. Source of Life. Bon Religion in East Bhutan and the Mon-yul Corridor.

Huber, Toni, 2013. "The Iconography of gShen Priests in the Ethnographic Context of the Extended Eastern Himalayas, and Reflections on the Development of Bon Religion", pp263-294 in Franz-Karl Ehrhard & Petra Maurer (Hrsg.) Nepalica-Tibetica Festgabe For Christoph Cuppers Band 1. IITBS International Institute for Tibetan and Buddhist Studies GmbH, Andiast.

Kapstein, Matthew, 2006. The Tibetans. Blackwell Publishing, Malden and Oxford.

Kapstein, Matthew, 2000. The Tibetan Assimilation of Buddhism. New York, Oxford University Press.

Karmay, Samten G., 2007. The Great Perfection (rDzogs chen). Second Edition. Brill, Leiden.

Karmay, Samten G., 2010, Tibetan Indigenous Myths and Rituals with Reference to the Ancient Bon Text: The Nyenbum (Gnyan 'bumf In: Cabezon, Jose Ignacio, ed., Tibetan Ritual. Oxford: Oxford University Press.

Klimburg-Salter, D, 2016. 'Along the Pilgrimage Routes between Uddiyana and Tibet: The Gilgit MSS covers and the Tibetan decorated book cover', in Erika Forte, Liang Junyan, Deborah Klimburg-Salter, Zhang Yun, and Helmut Tauscher, eds., Tibet in Dialogue with its Neighbours: History, Culture and Art of Central and Western Tibet, 8th-15th Century. China Tibetology Publishing House, 2016.

Kreijger, Hugo E, 1999. Kathmandu Valley Painting: The Jucker Collection. Serindia Publications, London.

van der Kuijp, Leonard W.J, 2016. “The Lives of Bu ston Rin chen grub and the Date and Sources of His Chos ‘byung, a Chronicle of Buddhism in India and Tibet”, Revue dEtudes Tibetaines, no. 35, April 2016, pp. 203-308.

Martin, Dan, 2015. 'An Archaic Book of the Dead', in Tibeto-logic blog, Monday, December 07, 2015 http://tibeto-logic.blogspot.co.uk

Martin, Dan, 2014. 'Bird Dogs of Tibet', in Tibeto-logic blog, Monday, March 10, 2014. http://tibeto-logic.blogspot.co.uk

Martin, Dan, 2010. 'Birdhoms', in Tibeto-logic blog, Wednesday, June 16, 2010. http://tibeto-logic.blogspot.co.uk

May er,Robert, 1996. A Scripture of the Ancient Tantra Collection. The Phur-pa bcu-gnyis. Kiscadale Publications, Oxford.

Meinart, Carmen, 2002. 'Chinese Chan and Tibetan Rdzogs Chen: Preliminary Remarks on Two Tibetan Dunhuang Manuscripts', pp.289-308 in Henk Blezer, ed. Tibet, Past and Present: Religion and secular culture in Tibet. Tibetan Studies II (PIATS 2000), Brill, Leiden.

Ramble, Charles, 2014. 'Real and imaginary Tibetan chimeras and their special powers.' In Mongolo-Tibetica Pragensia ’14, Volume 7, No. 2, Special Issue "Indigenous Elements in Tibetan Religions", Edited by Daniel Berounsky. Charles University, Prague.

Ramble, Charles, 2013. 'Both Fish and Fowl? Preliminary Reflections on Some Representations of a Tibetan Mirror-World.' Pp 75-89, in Franz-Karl Ehrhard & Petra Maurer (Hrsg.) Nepalica-Tibetica Festgabe For Christoph Cuppers Band 1. IITBS International Institute for Tibetan and Buddhist Studies GmbH, Andiast.

van Schaik, Sam, 2013. “Dating Early Tibetan Manuscripts: A Paleographical Method.” Scribes, Texts and Rituals in Early Tibet and Dunhuang. Eds. Brandon Dotson, Kazushi Iwao and Tsuguhito Takeuchi. Weisbaden: Reichert Verlag, 2013. 119-135.

Vitali, Roberto, 1996. The Kingdoms of Gu.Ge Pu.hrang According to the Mnga'ris rGyal rabs by Gu.ge mkhan.chen Ngag.dbang grags.pa. Tho.ling gtug.lag.khang, Dharamsala.