Vedic, Medieval, and Contemporary Concepts in the Nepalese Agnihotra Ritual

Michael Witzel (From: Festschrift J. Heesterman, 1992) <Diagrams and maps not included> MEANINGFUL RITUAL. Vedic, Medieval, and Contemporary Concepts in the Nepalese Agnihotra Ritual. A casual visitor to the southern outskirts of the town of Patan, which is situated south Kathmandu on the other side of the Bāgmati River, may come across one of the many two-storeyed and house-like temples, which are home to such deities as the Navadurgā or Aṣṭamātṛkā. They are known as dyaḥcheṃ (vulgo: dyocheṃ) or dyaḥga (dyoga) in the Newārī language, a word which ultimately derives from Sanskrit devagṛha, "god's house". This temple, however, is known as Agiṃmaṭh, Agnisālā or Agnisthal, or by its Sanskrit equivalents as agnimaṭha, agniśālā, agniṣṭhala. The farmer families (jyāpu) who live around the temple in this quarter of the town called Thambu can point out that it is Agni, the God of Fire, who resides here, and that, indeed, women come here every morning to pay homage to this god. One can always meet a few of them, clad in the traditional Newār dress, a black sari with red hem, who come carrying brass plates with the usual pūjā offerings. At this temple, however, these offerings are often supplemented by small pieces of wood. Indeed, if one looks through the door of the temple one can see three fire altars which are about a three or four feet high. A fire is burning on them or smoldering on a huge heap of ash, and a priest, often assisted by his wife, burns wood, ghee or various sorts of grains on them. An Indologist or Vedic specialist, however, who might visit this temple will be struck by the pictures of Śiva, Viṣṇu, Brahmā that are painted on the fire altars, something one is not used to in ancient, Vedic or medieval, Purāṇic sacrifice. The vedisant will ask himself why all these women are coming and offering. Is this a Vedic ritual or is it one of the everyday pūjās, for which women usually come and bring offerings early in the morning? The priest takes these offerings, separates the grains which he throws, handful by handful, into the fire while mumbling prayers. The visitors then receive the prasād of the god, some flowers, they bow and leave. Sometimes, an old woman might just come and bow to the fire and go away again. Others have a brief chat or, less frequently, an extended conversation with the priest's wife. In such cases, they retire to an antichamber of the temple to talk in private. All of this does not seem to present anything unusual: a priest offering, though not in front of a statue of some god but directly into a fire, and women (though hardly ever men) who come to make their pūjā. Yet, as we will discover, this temple, its ritual and the priestly traditions underlying them are something very special: actually, they are unique in the whole of the subcontinent, and even in those parts of Asia which exhibit cultural influences from India. An anthropologist who strays by the temple may observe that women do not pray to a figure of a god, as is usual in South Asia: they do not, for example, touch the door posts of the temple where there are some four figures of gods on the toraṇa crowning the entrance, though the lowermost one even is one of Agni, the main deity of this temple. Inside, however, one finds only the aniconic, physically present God Fire. The anthropologist may also note the frequent and lively interaction, between the visitors and the priest, and especially those involving wife of the priest. This is something that is not common at the pūjās at temples or small shrines. The very participation of the priest's wife strikes the observer: usually, it is men who offer, but here, one can also see the priests' wife offer grains into the fire. On the other hand, one will notice that it is the usual gods of Nepal who surround the temple: A Bhairava who gets a bali1 offering, a Gaṇeśa, a Nāga, and Lakṣmī. The temple, therefore, just seems to specialize in the worship of the otherwise hardly prominent Purāṇic deity Agni, the Fire God. Asking why local people come to the Agnimaṭh, the anthropologist gets something like a standard answer: to ward off evil, or to offer something to counter the bad effects of the planets. He may be told that a local astrologer, a jośi, has sent the visitors to the Agniśālā to worship and to offer. This happens especially, he is told, on birthdays, and in this way, people insure themselves against the bad influences of planets in the year to come, or they do so for another member of the family. This yearly visit seems to be one of the many rituals and pūjās performed to insure one's future, one may conclude. It seems to be a Ne0alese development, as people otherwise do not flock to Agni temples, - there are not any of 1 On the role of Bhairava who receives the remnants of the offerings, see Ucchiṣṭa Bhairava, see Sanjukta Gupta, Dirk Jan Hoens, Teun Goudriaan, Hindu Tantrism, Leiden 1979, p. 156. them one would know about in India. Yet Nepal differs from India in so many respects - why not in this one as well? If our imagined anthropologist knows something about Tantra, he may also speculate that this ritual may have something to do with the commonly found male/female opposition. The god Fire, just as the 'element' Fire is quite generally imagined as male, and the women visiting the temple may represent the female element. This seems to be echoed inside the temple by the priest and his wife. Perhaps there are more correspondences, but, unfortunately, the shape of the altars (square, round, half circle) does not offer a clue: while, in the Tantric systems, the half-circle may be represent an inverted triangle and therefore the female element, the male counterpart is missing here: the round fire represents heaven or the underworld, stored energy, the use of which permits to supercede the natural world and to reach salvation or emancipation; and the square fire is a symbol of extension, of a surface, and of the earth.2 Unless earth is regarded as a representation of the female element, this aspect would be missing in the Agniśālā design. But then, the round fire would repeat the male aspect, having the form of heaven.3 - However, an identification with the three famous gods of the Hindu Trinity, Brahmā (round fire), Viṣṇu (square), and Śiva (Rudra, semi- circular) is habitually made by the priest and the pictures of these gods are found on the respective altars. This identification certainly does not fit the characters of the gods and the interpretation of their forms in Tantra: while Brahmā is the creator, his fire has a shape which would indicate stored energy; Viṣṇu is the preserver, but his fire indicates extended energy; and finally, the half-circle of Śiva would rather represent the female aspect, and this certainly does not agree with the nature of this god. The anthropologist therefore is left with many unanswered questions and whatever he may ask, the priest and the visitors will not lead him much further in an understanding of this ritual. For example, another common answer offered by the priest is that it is performed for his own benefit and for that of the King and the country. Another type of visitor, an Indologist, - or as one prefers to say in Nepal, a Sanskritist, - may recognize the basic features of a homa sacrifice. There are several fires into which an offering of ghee, grasses, or grains is made. There are three more altars in the western half of the hall, which one can barely discern in front of the dark eastern wall of the temple, if one leans into the temple and the priest does not reprimand one, as he does the children playing in the area, for doing so. All of this 2 In Vedic thought, however, earth is round, and heaven is square, (see below). 3 See the preceding note. might fit one's general knowledge about various types of fire rituals in the Epics and Purāṇas. The indologist may know, for example, of a Bhāgavata Purāṇa recitation lasting seven days which is accompanied by offerings into five fires, or he may have read about the elaborate lakṣahoma ritual which requires an offering of grain made 100,000 times. He must be puzzled, however, by the pictures of the Hindu trinity on the three fires. He may also wonder what the priest's wife is doing in such a homa ritual. When the priest is asked for details, he turns out to be a Brahmin (and not, for example, a karmācārya priest of the farmer caste, an ācājyu Jyāpu) who knows some Sanskrit, but our Indologist will certainly be surprised to hear that this ritual is called an agnihotra. He begins to wonder: an Agnihotra with fires representing Brahmā, Viṣṇu and Śiva? After all, these gods are not prominent in the Veda, and they are not found at all in Vedic texts which describe and interpret the Agnihotra offering. They gained prominence only well after the end of the Vedic period, and they never were part and parcel of a Vedic ritual such as the Agnihotra. On the other hand, one is then immediately projected back into a Vedic framework when one learns that a small figure covered by red sindura dust, which is placed at the foot of the large Viṣṇukuṇḍa fire altar, is called Mitrāvaruṇa (sic!). The gods Mitra and Varuṇa, Vedic dual deities, are not prominent after the Vedic period: Varuṇa feebly survives as the god of the ocean and Mitra is virtually unknown. So the Indologist he may wonder what kind of Vedic survival he is witnessing or what kind of a mixture with post-Vedic ritual elements this may be. If our visitor happens to be a Vedic specialist, he will at first be pleasantly surprised: at one glance, he can recognize the three holy fires prescribed for a Vedic Agnihotra ritual: the left one,4 situated in the east, is the square āhavanīya fire, the right (western) one is round, having the typical shape of the gārhapatya fire, and the middle one has the form of a half-circle, the shape prescribed for a dakṣināgni. A Vedic specialist will also have recognized the peculiar ancient shape of the offering ladle used by the priest. Even the presence of the priest's wife does not astonish him: she must take part in the Agnihotra, so the Veda prescribes. If he happens to catch a few Mantras spoken by the priest in a low voice, he may recognize some of them as Vedic, and perhaps even as belonging to the set of Agnihotra yajuṣ formulae. But then, the attention of our Vedisant will be drawn to the many aspects of this ritual which do not fit his mental picture of an Agnihotra: there are the pictures of 4 The door of the temple opens to the north, to a courtyard which is walled in by a cemented brick wall behind some trees. Brahmā, Viṣṇu and Śiva which also gave their names to the fires on which are painted. The three extra altars in the eastern half of the hall, he hears, are the Sūrya, Kumāra and Maṇi(nāga) kuṇḍas. This definitely includes one altar more than allowed in the Vedas, - and their names are untypical and new again. And what does he have to make of the figure of Mitrāvaruṇa? Is this a survival from Vedic times, - a development perhaps not quite impossible in this secluded Valley which has seen only one Muslim invasion of three weeks in the fourteenth century and was hardly exposed to other foreign influences since then. Also, the priest belongs to the group of Brahmins who are among the oldest layer of immigrants to the Valley, the Newārī speaking Rājopādhyāyas. But then, one has never heard of or read about any figure of Mitrāvaruṇa (singular!). Even then, one may speculate, as a dual deity, both Mitra and Varuṇa might perhaps be represented by just one figure... This, however, will not be the end of questions for our Vedisant. He notices that Mitrāvaruṇa is taken out of the temple hall at the end of the pūjā and carried by the priest towards a tree in the courtyard, which he calls a varuṇa tree. The Mitrāvaruṇa figure then makes a round of other gods in the courtyard: Viśvakarmā, Siddhilakṣmī, Kumāra, Gaṇeśa, Bhairava, and Nāga. Finally, a Vedic scholar will be baffled by the ritual itself, of which he recognizes only some basic features such as the pouring ghee or the throwing of grains into the fire, a substitute for the usual Agnihotra milk. This substitution, however, he knows, is allowed already in the Veda. But the the rest of the many actions escapes him: Why are multiple offerings of grain made by the priest or his wife from plates of leaves, why are metal spoons of strange shapes used in place of to the well known wooden agnihotrahavanī ladle? Why is Mitrāvaruṇa worshipped in an Agnihotra ritual at all? What does the priest do at the sixth altar, where he sits for an extended period of time, at the end of the ritual, reciting long passages of Mantras of which one can occasionally grasp a few lines dedicated to ancestors and the King... The sixth altar does not contain a fire, one notices on closer inspection, but is a structure covered with dry leaves, flowers, and strings. He learns that this is a kalaśa, the (Maṇi)Nāga-Kalaśa: this certainly is something post-Vedic as well.5 And of the offerings made outside of the temple, one portion is thrown on the Bhairava stone in front of the door, some other portion is put on the wall of the courtyard: this is the śvabali and the kākabali, he learns. Even to the Vedic specialist, therefore, this ritual does not make sense. It represents to him, at best, a curious mixture of old and new, a Vedic pattern interspersed with later traits, and a curious survival of a Mitrāvaruṇa worship. Why, 5 On Maṇināga, see n. 86 on Varuṇanāga. finally, is this Agni venerated at all, while the three other Agnihotras in the Kathmandu Valley are not visited by the populace, and why is their Agni not worshipped by the local people, as it indeed should be the case? A Vedic ritual, one knows, basically is a private ritual not intended for public participation or for daily worship of outsiders. The three specialists, whom we imagined in the preceding paragraphs, cannot provide, even taken together, a cogent and tenable explanation of this ritual. As we shall see, important aspects for an understanding have been noted. But it will be necessary not only to put these observations and conclusions together but to add many other, more hidden facets. Only then, we can arrive at an understanding of this particular Agnihotra ritual,6 can gauge its importance for the people of present day Patan, can evaluate the legends told about the temple and its priests, can understand the functions of the many, too many, small ritual details involved, and finally can correctly describe and comprehend the historical trends which are underlying the development from a 3000 year old Vedic ritual to the contemporary Agnihotra as performed at Patan.7 *** 1. The Vedic and the Tantric Agnihotra. 6 Indeed, many aspects became clear only when I got access in 1985 to the manuscript used by the priest for the daily Agnihotra; an older MS. of 1433, mainly treating the fortnightly New and Full Moon Sacrifice, had been available in a film made by the Nepal German Manuscript Preservation Project in 1976 (see below). 7 The three other Vedic Agnihotras regularly performed in the Valley are not treated here in detail; for these, cf. my earlier article (Agnihotra-Rituale in Nepal, in Formen kulturellen Wandels und andere Baiträge zur Erforschung des Himalaya, Nepalica 2, ed. B. Kölver u. S. Lienhard, St. Augustin: VGH Wissenschaftsverlag 1986). Other fire rituals are not treated here as well, e.g. Vedic rituals at weddings or at nāmakaraṇas, or non-Vedic rituals such as the common fire rituals at the reading of Purāṇas, the lakṣahoma (last performed at Bhaktapur in the Summer of 1976 to avert an earthquake) and other Śāntihomas, the various other fire rituals connected with Bhairava, etc. (cf. in this volume, the article by B. van den Hoek), and finally, the various Buddhist homas. I will describe and interpret most of them, in a separate work in much greater detail. For the Buddhist homa, see author, The Buddhist forms of fire ritual (homa) in Nepal and Japan. (Summary). 31st CISHAAN (Tokyo-Kyoto), Proceedings, ed. by T.Yamamoto, Tokyo 1984, p.135.  First, I will try to put some order into these impressions and give some answers to the questions put; I begin to do so with the the oldest form of this ritual, the Vedic Agnihotra. <1> In its barest form the Vedic Agnihotra consists of an offering of milk or, optionally, of ghee into the three sacred fires. This takes place in the evening and is accompanied by a few mumbled prose sentences, the central ones of which are the mantras: "the light is Agni, Agni is the light" (agnir jyotir, jyotir agniṃ, VS 3.9a, KŚS 4.14.6). They establish an identification of agni and jyotiṣ, i.e. the heavenly light (the sun) and the light here on earth (Agni, the fire). This magical identification, which can also be seen in the old Iranian fire ritual,8 assures the survival of the sun through the dangers of the night and secures its rise next morning, when the fire made to blaze again from its embers. After offering into the fires, they are venerated and then the coals are shoved together and covered with ash so that they can be kept glowing until the next morning. Then, they are rekindled before sunrise, with the same ritual and the same mantras, except that the central yajuṣ formula is changed to: "The light )s the sun, the sun is the light" (sūryo jyotir, jyotiḥ sūryaḥ VS 4.15.9-10) -- a magically effective inversion of the Mantra used in the evening. While this ritual dates from Vedic times and thus is at least 3000 years old, it has survived at Patan in a peculiar form apparently not found anywhere else in the Indian subcontinent. Otherwise, as well, the secluded nature of the Kathmandu Valley has allowed the preservation of an unusual amount of historical and literary documents on the Agnihotra. This lucky coincidence allows us to follow its development, its interpretation, and its diversification in this outlying region of the subcontinent through several millenia. Actually, the Agnihotra is still one of the simpler Vedic rituals, yet it contains about 100 separate actions. As space does not permit to detail all these actions - they have been described well by Dumont9 - I refer to the diagram (below) in order to provide an idea of the old and the "new" Agnihotra. Obviously the main parts, such as 8 Found in Yasna Haptaŋhāiti 36.3,6 where the offering fire and the Fire of Ahura Mazdā is identified with the Sun; see J.Narten, Yasna Haptanhāiti, Wiesbaden 1986; see also, below n. 13 . - For the Vedic identification of the fire and the sun, note the following: Sun entering the fire: TB 2.1.2.9, KB 2.8, ŚB 2.3.1.3, 2.3.1.36, 11.6.2.2, JB 1.8, 1.9, 1.11, 1.23; Fire entering the rising sun: KB 2.8, ŚB 2.3.1.36, 11.6.2.2, JB 1.23; Sun born (from fire): KS 6.3:51.9, ŚB 2.3.1.5, etc. 9 P.E. Dumont, L'Agnihotra, Description de l'agnihotra dans le rituél védique d'après les Śrautasūtras..., Baltimore 1939. -- I will describe and interpret this ritual, along with the three other (Vedic) Agnihotras of the Kathmandu Valley, in a separate work in much greater detail. the offerings (of ghee) into the fire, the three (plus two) sacred fires into which offerings are made, have been retained. As in Vedic times, both the priest and his wife, the Agnihotriṇī, take part. <2-8> The basic structure of the Vedic and contemporary Agnihotra is similar, both in text and in actual performance today. To use P. Thieme's description of the available oldest Indian ritual, as alluded to in the Ṛgveda, these rituals are forms of guest worship. <10> Guests / gods Ritual - preparation of the offering ground | | preparation of the fires, vedi | -- the guests/gods are invited | - āvāhana of gods... (to the offering ground), | -- they are washed, refreshed, | - offerings of milk, etc. feasted, and praised | - with several subframes with hymns, | -- they are sent off (to be | - visarjana of gods invited again another | time). - final rites (covering the fires, their worship: agnyupasthāna) Actually, there is a great similarity in Ṛgvedic, 'classical' Vedic and in modern pūjā ritual which has not been noticed sufficiently so far.10 Again, for want of space I can only summarize the comparison of all the details which could be compared when a detailed study of the Vedic and the modern Agnihotra at Patan is made: 1. The order of some of the actions has been changed. 2. There are a few intrusions in the old ritual: Varuṇa, the Nāga, the long śrāddha at the end... (and, incidentally, to take refuge with the by now standard terminology of Great/Little Tradition, does not explain everything here).; 10 See author, Die Kaṭha-Śikṣā-Upaniṣad und ihr Verhaltnis zur Śīkṣāvallī der Taittirīya-Upaniṣad. WZKS XXIV (1980), p. 37 sq.  3. There is an elaborate Tantristic framework into which the ritual is set. I begin with the last item, the Tantric influences and its framework. The Vedic ritual starts with the cleaning of the offering ground, strewing of grass etc., that is, with 'household' preparations for receiving guests in a still rather simple society. The Tantric Agnihotra of Patan, however, stresses the mental preparation as well: not only is the offering ground cleaned and cleansed but so is the body and the mind of the offerer. This is easily observable in the meditation of both the priest and his wife, and their breathing exercises. Producing the god mentally (bhavanā) and identifying him with parts of one's own body (nyāsa) follows.11 This is new and Tantric: all limbs or parts of the body are identified with those of the god. - Only then do the actual practical preparations take place, such as rekindling the fire. It is within these Tantric and "practical" frames12 that the actual ritual begins with the vrata: the taking up of the observance governing the sacrifice. The structure of these frames both in their Vedic and Tantric forms is similar.13 11 See for example the MS of 1433 A.D., fol.38b (uncorrected): śrāyaṃ-prātahunad agnau sarvvaṃ pāpapranāsanaṃ | sakalaṃm agnihotrasya tena nityaṃ prahūyate || nyāso-sthānāgniviharaṇa-vratagrahaṇaṃ | pūrvvavat toyenāgnipradakṣaṇī kṛtya || arjavātana- lokapālārccaṇaṃ | sarvvatra pūrvvavat samidhahomaṃ || 12 I stress the Vedic idea of frames, differently from Staal's procedure, as demonstrated again in his recent book Jouer avec le Feu, Paris 1990. To conceive the ritual as built up of frames or boxes is in concordance with Vedic practice. For example, Ved. avāntaradīkṣā means "the lower, inner consecration", i.e. the one which has been inserted into the normal consecration of the Soma ritual; on the framework structure of Vedic ritual, see already author, Agnihotra-Rituale in Nepal, p. 172, n. 34; 31st CISHAAN (Tokyo-Kyoto), Proceedings, ed. by T.Yamamoto, Tokyo 1984, p. 534, and: On the origin of the literary device of the 'Frame Story' in Old Indian literature. Hinduismus und Buddhismus, Festschrift für U. Schneider, ed. by H.Falk, Freiburg 1987, pp. 380-414. 13 Again, the structure of the Iranian ritual is similar: Yasna general, post-Zaraϑuštrian worship | | Gāϑās Y. 28-23 hymns composed by Zaraϑuštra | | Yasna H. 35-36 35.2 welcoming | | | 35.9 voicing of manϑras | | | 35.10 worship (yasna) of Ahura Mazdā | | | central identification: 36.3 fire = of Ahura Mazdā <11> ________________________________________________________________________ VEDIC RITUAL TANTRIC RITUAL - preparation of the offering- preparation of the offering ground | ground and of the mind, body of | | priest (ātmaśuddhi) | | - calling into his body of the || | god (nyāsa) | | | preparation of the fires, vedi | preparation of the fires, vedi | | | -vratagrahaṇa (taking up | | | | of ritual observance) | - āvāhana of gods... | | - āvāhana of gods... | | | | - unifying god with one | | | | mentally produced earlier | - offerings of milk etc. | | - offerings of ghee etc. | - with several subframes | | - with several subframes | - visarjana of gods | | - visarjana of gods | | | | | - vratavimokṣaṇa (giving up the vrata.) | | - final rites (covering the fires,| - final rites (covering the | their worship: agnyupasthāna) fires, pūjā of the fires) - finalizing the nyāsa state14 ________________________________________________________________________ | | | 36.6 fire = Sun | | Yasna H. 37-41 37 worship of A.M., earth, water, etc. | | 40- prayer for all Zoroastrians | Gāϑās Y 43-53 composed by Zaraϑuštra, 53 for the | marriage of his daughter (appended) Yasna general, post-Zaraϑuštrian worship For other, often quite specific similarities between Vedic and Iranian ritual, see Hillebrandt, Rituallitteratur, p. 11. 14 There is a similar overlap in the Buddhist homa, see author, The Buddhist forms of fire ritual (homa) in Nepal and Japan.  Here, we can perceive system of frames within frames within frames so typical /f Indian (and some other) rituals. Ritual does not have, as Staal will have it, a structure similar to an inverted tree, so well known from modern grammarians, but rather a complicated frame structure.15 This model (whether one prefers to call it -with P. Hacker- 'inclusivism' or not) helps significantly to understand not only Vedic ritual but even some of the main features and structures of Indian thought. In the same way as the various ancient and more recent subsections and parts of the Vedic rituals are assembled, old and new are often found next to each other in other contexts as well, the older version having been "enveloped" and superceded, and yet often left unchanged by the new forms. This is the case, on a larger scale, also with the Vedic rituals of the Kathmandu Valley, where the Agnihotra is found in a separate, ancient form next to the 'reformed' Patan Agnihotra.16 To return to the ritual itself: <11a> Space does not allow to compare the Gṛhya and Śrauta fire rituals in detail, and I also cannot take a longer look at the worship of Mitrāvaruṇa which is, as we shall see, a medieval addition. An extra altar for a deity Maṇināga17 is also unknown to the Vedic texts. It is Varuṇa, who in post-Vedic texts is not much more than a god of the Ocean and the waters, who is represented here as the snake god (or Nāga) Vāsuki by/in a water pitcher, a kalaśa, put on the sixth altar, the maṇikuṇḍa. In post-Vedic Hindu (and Buddhist) worship, gods frequently are called into such kalaśas, where they are represented by nuts. The order of worship at the five fires, however, is important. Notably, the fifth, the kumārakuṇḍa (Ved. avasathya), is not used at all. The Vedic evening Agnihotra starts with the taking out of embers from the gārhapatya (brahmakuṇḍa) and continues with taking them to the āhavanīya (viṣṇukuṇḍa), while the dakṣiṇāgni (rudrakuṇḍa) either may burn permanently or may be kindled every time again, also with coals from the gārhapatya. However, this description applies to the evening Agnihotra, while the Patan ritual is more elaborate in the morning, both in its actual performance and the 15 Which is also found in Indian literature (Mahābhārata, Pāṇini, Purāṇas, etc.) and in religion: Buddha as the ninth avatāra of Viṣṇu, Brahmā as 'enveloped' by or emanated from Brahman, etc.; cf. author, On the origin of the literary device of the 'Frame Story' in Old Indian literature. Hinduismus und Buddhismus, Festschrift für U. Schneider, ed. by H.Falk, Freiburg 1987, pp. 380-414. 16 See author, Agnihotra-Rituale in Nepal. 17 Cf. below, section 4. - A special homa ceremony for Varuṇa is described at Agni Purāṇa 64, which makes use of eight fires. description in the MS. In the morning, the fires are rekindled with libations of ghee from the embers of the āhavanīya fire. This means, as has been indicated, that the sun receives power from the fire altar: the ritual mostly was performed before daybreak, and thus the sun was strengthened enough to start its course again: embers are taken out from the āhavanīya (representing the Sky) and moved towards the gārhapatya (Earth) and the dakṣiṇāgni (Moon). One set of offerings deserves special mention and a more detailed investigation, the desapāta18 offerings which are a medieval addition. They are interesting as they can give us an idea of which deities or places of worship were regarded as essential by the author of the new Agnihotra ritual. The following list is given according to two MSS; the unnumbered deities coming from only one of the two MSS. <12> @@see INSET add list <9> INSET ------------------------------------------------------------------------- Location /interpretation Agniśālā gods ------------------------------------------------------------------------- 0 Sūrya Sūryakuṇḍa 1 Gaṇeśa Nārāyaṇa Śiva Rudrakuṇḍa 2 Maṇikumāra Manikakuṇḍa 3 Allima Brahmakuṇḍa? 4 Keśava 5 Viṣṇu Viṣṇukuṇḍa --------------------------------------------------------------------------------------------- Gods of Patan and its surroundings --------------------------------------------------------------------------------------------- Patan circle Valley circles 6 Dhanaṃjeśvara NE: Paśupatināth? Dhanes Bazar? 7 Bhīmasena C: Patan Tārāya 8 Mānabhāirava Manibhairava, <S: Tangal Hospital> 9 Māneśvarī ditto (Śakti of Bhairava) 18 From Newārī pāta "offering plate"; the rite is otherwise called deśabali in Skt. 10 Kumāra ? Agniśālā 11 Ganeśa ? Agniśālā 12 Śikhanārāyaṇa C: W. of Agniśālā <or SW: Parphing> 13 Āryāvalokiteśvara S: Patan 14 Mīnanātha S: Patan --------------------------------------------------------------------------------------------- Agniśālā gods --------------------------------------------------------------------------------------------- Nīlagrīva 15 Triliṅgeśvara N. of Agniśālā, or: W: near Balkhu River 16 Nāga ? Agniśālā 17 Lakṣmī ? Agniśālā, or: S: Mahālakṣmī ------------------------------------------------------------------------- 18 Dhvaka-Bhīmasena <C. Patan?> NE: Dolakha! 19 Vārāhī N: (Dhaṃtila°) <S: Bajravārāhī> 20 Śikho-Bhairava S: Bungamati (°vārāhī) 21 Mahālakṣmī S: Patan Industrial District ------------------------------------------------------------------------- 22 Dathulacche Ināya Vinayaka-Gaṇeśa: Agniśālā N: Dhathukachem? Kūpa (= near Teku) 23 Gaṇeśa ? Agniśālā 24 Kūpa Agniśālā; well 26 Gaṇeśa ? Agniśālā; near the well ------------------------------------------------------------------------- 27 Vaiṣṇavī SW: Nakhu River, near the jail 28 Corī S: Amlekhganj 29 Taḍāga "pond" SW: Nakhu ? 30 Nāṭeśvara 31 Cāmuṇḍā? 32 Vaṭuka-Bhairava S: Baṭuki-Bh. near the Stupa 33 Kumbheśvara N: Patan 34 Bhāga-Bairava W: Kirtipur (= Vyāghra-Bh.) 35 Ratneśvara 36 Sveṣṭadevatā = Devatā of the priest? 37 Paśupati C: Paśupatināth Temple 38 Guhyeśvarī C: ditto, near Paśupatināth 39 Vacchalādevī C: ditto, near Paśupatināth 40 Valasuki-Nāga C: ditto, near Paśupatināth? (Vāsuki-Nāga?) 41 Jayarājeśvarī C: ditto, near Paśupatināth 42 Rājarājeśvarī C: ditto, near Paśupatināth 43 ..]ilasvarī C: ditto, Maṅgaleśvarī? 44 ..]vanī 45 Dolaśikha Nārāyaṇa E: Changu Nar. E: Changu Nar. 46 ..]ti/Bhūnīśa 47 ..]dyanam 48 ..]ajimā ----------------------------------------------------------------------------------------------- Agniśālā ----------------------------------------------------------------------------------------------- 49 Pāñcāgni 5 fires 50 Balithāvali bali places, northern courtyard 51 Thaṃtu "well", south of Agniśālā 52 Gaṇeśa NW courtyard 53 Nasuvalī /Laṣu° bali place? 54 Balkhusayavi Balkhu cremation ground?? 55 Dvāra door god(s): Brahma, Viṣṇu, Rudra, Agni 56 Nāga Nāga at E. doorpost of Agniśālā ---------------------------------------------------------------------------------------------- 57 Vajrayoginī NE: Sankhu/SW Parphing 59 Koṭeśvara NE: Patan: Manohara River 60 Vāgmatī N: Patan, Bagmati River 61 Tīrthas N: Patan: ghāṭs 62 Vajravārāhī S: near Capagaon 63 Nārāyaṇa 64 Kvathu N: Patan, near Bagmati R. or pond on the way from 65 Chāso Jalavinayaka 66 Hanuman NE: Patan: river, or at the Palace gate? 67 Varuṇa 68 Gṛhalakṣmī Siddhilakṣmī of the Agniśālā courtyard? 69 sarve devāḥ All the gods of the town and the Valley ---------------------------------------------------------------------------------------------- iti deśāpāti We will have to return to this list: It even includes the two Buddhist gods Āryāvalokiteśvara19 and Mīnanātha,20 and also many gods surrounding Patan inside the Kathmandu Valley; they are, at first sight, arranged in two or three concentric circles. 19 Who interestingly is not yet named Macchendranāth, as has been the usage of the last few centuries; for the names of Macchendranāth and the history of this ancient Newār rain god (of the village of Buṃga), see J. Locke, Karunamaya, Kathmandu 1980.  Summing up this brief comparison of the old and the medieval ritual it can be stated: - The outward form of the Vedic Agnihotra has been retained (3/5 fires, preserving and rekindling the fire, etc.). - A new (Tantristic) frame has been created to include the Vedic actions (albeit in a changed order and simplified). - This framework also includes many post-Vedic, and notably, Tantric additions: worship of new deities as found in medieval Hinduism as well as some very special deities and entities encountered only here, at the Patan Agnihotra, such as Mitrāvaruṇa, Maṇināga or the Varuṇa tree. In short, the elements of the great Brahmaṇical (and Vedic) tradition of Northern India are clearly in evidence, but there are quite a number of later and notably local developments as well. The influence of the Little Tradition, the local one of the Newār of this Himalayan Valley, makes itself felt. We thus do not witness a simple survival of old Vedic forms (as for instance in the repeated re-introduction at Paśupatināth21 and two other places22 in the Valley of a Vedic form of the Agnihotra during the last two centuries.) The question arising here, therefore, is not the simple one of change and continuity23 in Indian religion and thought, but -- as we can deduce even from the quick comparison of ancient Vedic, medieval, and modern ritual presented so far, -- what must be explained here is a special development, an amalgation which involved various strands of tradition, both ancient and medieval, and both local and supra-regional.24 20 The 'brother' of Macchendranāth, see J. Locke, Karunamaya. 21 See author, Agnihotra-Rituale in Nepal. 22 There even is a Commentary, written by one Nityānanda Parbatīya (thus, apparently, a Nepālī speaking Brahmin), Kātīyeṣṭtidīpakaḥ, samvat 1981, bhādra kṛṣṇa 8, śukre, i.e. at the end of August 1924 A.D. 23 Cf. Gonda, Change and Continuity in Indian Religion, The Hague 1965. 24 For a general discussion of these terms in the framework of Brahmanical culture, see author, Regionale und überregionale Faktoren in der Entwicklung vedischer Brahmanengruppen im 2. Conditions of Change. This leads to the second topic which I should like to address here: the actual conditions of change. These are, as always in the Indian subcontinent, somewhat difficult to fathom due to the perceived, almost complete lack of genuine historical writing in Indian civilization, and often,the actual lack of documents pertaining to the question.25 In Nepal, however, we are somewhat more fortunate in this respect, and a brief look into history is useful to establish a background against which the changes can be projected which took place in the development of this ritual. Ever since 4he earliest inscription of Nepal found so far, i.e. the Caṅgu Nārāyaṇ stele of Mānadeva of 464 A.D., Nepal has been a Hindu kingdom, with Hindu kings and courtiers, - whatever the religion of the mass of the people had been throughout the ages. Thus, the coronation inscription of Aṃśuvarman, put up in 6o5 A.D. at Harigaon, already refers to an Agni temple or ritual. Here, a little to the south-west of the probable site of his palace,26 one of the few scientific excavations carried out in the Valley until recently27 has produced a platform of large bricks with a fire altar, an agnikuṇḍa, most probably the Agni mentioned in Aṃśuvarman's inscription. Then, luckily, among the c. three thousand, mostly unpublished medieval inscriptions, there is one of King Ānandadeva (1127-1167 A.D.), which even mentions the agnimaṭha. The princes Yaśomalla and Someśvara serve as dīkṣitas of the ritual in c. 1140 A.D. The next inscription, one by prince Dharmamalla of 1401 A.D., which was found next to the present Agni temple, already mentions a restoration of the Agnimaṭha at its location at Thambu in south-west Patan. And a slightly damaged inscription in the Mittelalter, in Regionale Tradition in Südasien, ed. H.Kulke and D. Rothermund (Baiträge zur Südasienforschung 104), Heidelberg 1986, pp. 37-76. 25 On this topic see now author, On Indian historical writing, Journal of the Japanese Association for outh Asian Studies (JJASAS) 2, 1990, 1-57. 26 See N. Gutschow, Journal of the Nepal Research Center 1 (1977), p. 89-92, and author, On the location of the Licchavi Capital of Nepal, in Festschrift für P.Thieme (= StII 5/6) 1980, pp. 311-337. 27 See the reports by S.B. Deo; now, there have been excavations at Satyanārāyaṇ (G. Verardi, Harigaon Satya Narayan, Kathmandu: a report on the excavations carried out in 1984-1988, Rome 1988), and in 1984/5, 1988/9 at Dhumakhal, a village west of Caṅgu Nārāyaṇ (Th. Riccardi, forthc.). courtyard of the Agniśālā, dated 1572 A.D.,28 reports certain guṭhi regulations in connection with the Agniśālā. Now it could be asked, in the context of medieval Nepal: Why is the Agniśālā located at Patan, which was situated some five miles away from the medieval capitals of Gvala (Deopatan at Paśupatināth) and some ten miles from Tripura (Bhaktapur)? And why is it not found, for example, at the center of Patan? <14> The reason for this position of the Agniśālā is a surprising one: the division into two moieties of the Kathmandu Valley, perhaps in force since times <15> immemorial: - a male north-eastern one and a female south-western one: Two pilgrimage sites, <16> the bathing places (tīrtha) represent these two halves of the Valley: The ancestor worship for one's mother (śrāddha) is to be performed at the south-western location of Mātātīrtha (near Thānkoṭ), and that of one's father at a north-eastern location, at Gokarṇa. Even the kitchens on the top floor of houses are divided in this way: the south-western corner is the one reserved for women and it is here that they cook. The north-eastern corner is reserved for the gods and their pūjā. The same division into male and female moieties is also found in the town of Kīrtipur.29 As far as the Agniśālā and the Agnihotra ritual are concerned, the division of the Valley into two halves is reconfirmed by the legendary opposition of an origin of Fire at the Maṇicūṛ Mountain in the north-east and its establishment in the south-western corner of the Valley, near the water of the Triveṇi at Cobhar, and near the Campādevī Mountain, - a fact to which we will have to return. The selection, in the early Middle Ages, of Patan as the location for the Agnihotra temple therefore may have been influenced, apart from political reasons, by this mythical division of the Valley. As for possible political reasons, the installation of the Agnihotra ritual and temple can be understood as a deliberate action by a monarch, 28 According to C. Pruscha (ed.), Kāṭhmāṇḍu Valley, The preservation of physical environment and cultural heritage - a protective inventory, Wien 1975. 29 See R. Herdick, Kirtipur, Stadtgestalt, Prinzipien der Raumordnung und gesellschaftliche Funktion einer Newar-Stadt, München 1988, p. 97, 100 sq., and: Neue Kulte in Kīrtipur, in Formen kulturellen Wandels, p. 255. Note that this type of division may be very old: there are two moieties (Deva [N]E : Asura/Varuṇa [S]W) in the Vedic texts, see Kuiper, Varuṇa and Vidūṣaka, Amsterdam 1979, p. 55 sqq., cf. p. 154 sq.; the Nepalese situation obviously is patterned by the later, Sāṅkhya and Tantric stress of male/female characteristics. probably30 Śivadeva (1098-1126 A.D.), who wanted to create a new center. He chose a typical Brahmanical ritual when re-establishing his new Hindu Kingdom. In fact, he even had a very good motive: he needed ideological support against the strong (probably predominantly Buddhist) gentry of the pradhānas of Patan. Consequently, he used the well-known 'trick' of a new king or dynasty, that is to settle Brahmins with rich endowments around the central core area,31 and he founded the temple of an important god which these Brahmins had to worship, in short: we see the establishment of a cordon sanitaire.32 Even the actual performance of the ritual can be detailed on the basis of some historical documents. From the time of Jayasthiti's and Dharmamalla's restoration of a strong Hindu kingdom at the end of the 14th century comes a manuscript33 (dated 1433 A.D.) which deals in detail with the fortnightly Dārśapaurnamāsa ritual and their variants,-- as well as with the daily Agnihotra, be it in a quite a brief fashion. After this early written testimony, there are detailed MSS. from 1601 A.D. onwards to this day.34 Therefore, four periods can be distinguished in the development of the Patan Agnihotra: -- Vedic texts (Kātyāyana Śrautasūtra), 30 According to legend, it was his father Śaṅkaradeva (1060-82 A.D., see below, n. 36 ); but this statement is found only in Wright's Vamśāvalī, that is in a compilation based on older medieval chronicles and on legends of Patan current around 1820 A.D.; see Wright, History, p.159. 31 I suspect, following the settlement pattern in the area, that NW Patan (Sulimā, Gābahā, Puñcalī), closer to the Pulcok Stūpa, was Śivadeva's center of government; this lies some 3000 ft. (2 km) from the older centers around Konti (near the N. Stūpa) and Guita. Note the also role of the Vyāghra Bhairava of Kīrtipur who opposes Patan; he does so just like Śivadeva himself, who is said, by the Vaṃśāvalī to be an incarnation of the Bhairava of Kāmarūpa. (Guita is the medieval Gusaṭalo° and the Licchavi time Gullaṃtaṅga). 32 H. Kulke, Die frühmittelalterlichen Regionalreiche, in Regionale Tradition in Südasien, Baiträge zur Südasienforschung 104, Heidelberg 1986, pp. 77-114; cf. author, Regionale und überregionale Faktoren in der Entwicklung vedischer Brahmanengruppen in Mittelalter, in Regionale Tradition, pp. 37-76. 33 Kesar Library, Kesar Mahal, Dept. of Archaeology, Kathmandu, no. 28; cf. author, Agnihotra- Rituale in Nepal, p. 186; the MS was filmed by the Nepal German Manuscript Preservation Project, as film no. C 3/8. 34 See the description in: Agnihotra-Rituale in Nepal, p. 187 sq. -- Medieval Purāṇic evidence (especially in the Agnipurāṇa), -- the MSS of 1433 A.D. onwards, -- modern performance and the present day MSS. ****** The MS of 1433 A.D. already deals with the new Tantristic ritual in a form that is only marginally different from the ritual of today. The change from a Vedic Agnihotra, and perhaps from an early Paurāṇic ritual (of homa type, as described in Atharvedapariśiṣṭa 30-31), to a 'miśra', Tantristic one must have taken place in Nepal between 605 A.D. (the single fire altar of Aṃśuvarman) and 1433 A.D. (three / five sacred fires of the Agniśālā). Taking into account the inscription35 of 1403 A.D., which already mentions a 'restoration' of the temple, and evaluating the date of the foundation, according to the later chronicles of the turn of the 19th century (vaṃśāvalīs), of the Agnimaṭha by Śaṅkaradeva (1060-82),36 we may justifiably assume that the Tantric ritual, as preserved in the MS of 1433 A.D., is more or less the same as that used by Ānandadeva's son Yaśomalla and prince Someśvara, mentioned at c. 1140 A.D. in the earliest inscription found so far that clearly refers to the Agnimaṭha.37 The Tantric form of Fire Ritual. If this represents the establishment of a Tantristic Agnihotra in the 11th century, the question to be asked now is: What is typical for its form? What is the Tantric form of a fire ritual? The answer to these questions would, of course, fill volumes. Let me summarize them here in a few, very general sentences. In Tantra, the idea that the universe is constituted by two entities, male and female, the God and his female partner, his Śakti, plays a great role. All forms are constituted by the interactions and 35 See D. R. Regmi, Medieval Nepal, Calcutta 1966, Pt. III, p. 40 no. XXXIX and cf. Aiśvaryadhar Śarmā, Agiṃmaṭhaṃ.yā aitihāsik sāmagrī ["The historical materials relating to Agnimath"; in Newārī], Kathmandu 1984, p. 26. 36 This is unlikely because of the detailed description given in the Gopālavaṃśāvalī, fol. 30b sq. (māgha, kṛṣṇā āṣṭāmī mūla nakṣatra, NS 219) about the dominant role of Pradhāna Varapāla Bhāro at Patan; however, Śaṅkaradeva's successor Śivadeva (who is called "the Bhairava of Assām" by the Vaṃś.) is much more likely, cf. on Bhairava, n.31 , n. 57 . 37 D. R. Regmi, Medieval Nepal, vol. III, p. 13, no. XXIV. emanations of these two entities. The male element is the active one here;38 it is Fire, flaming upwards, and represented in diagrams (important in the present context, as we shall see) and in general symbolism by a triangle pointed upwards. The female element is passive, watery, trying to reach an equal level, and is represented by a triangle pointing downwards. These two triangles combined may form a star of David: In that case they are in productive union. Or they may form a rhombus:39 in that case, both are in a state of celibacy. One can easily note that the identification as fire and water of the two basic entities lends itself to a new interpretation of the fire ritual of Vedic times. Whatever is done with a fire can represent an activity of the god. Tantric religion originally was private (and, in that sense, similar to the private Vedic ritual) and usually challenged traditional religion and caste structure, but it has undergone change. For example, originally, the private meditation on fire was a form of sādhana, a very personal meditation, by which the adept tried to create (bhāvanā) a divinity from his mind, e.g. in the present case the god Agni, and inserted him into the fire. This samaya or "conventional" god is then unified with another one, the lokottaradeva Agni, called from "the other world". Both are united in the fire.40 Yet fire is, of course, not united with water, neither in ritual nor in the mind of the adept: Water is sprinkled around Agni, a Vedic practice, but it is not poured into him. Water is also present at the Agniśālā in various other forms: as samudra,41 Maṇināga, kalaśa, perhaps also as the varuṇa tree, and the Mitrāvaruṇa figure made from it. Thus, it surrounds Agni everywhere but does not enter him. Instead, the male God Fire is 38 Normally, "Śakti is the active partner in the cosmic act of procreation while Śiva remains purely passive... In this process the chief place is allotted to Śiva by the Kashmir Śaiva texts. But in Tantrism very often no precedence is given to one of the two." (Sanjukta Gupta et al., Hindu Tantrism, Leiden 1979, p. 55 sq.). -- On fire and water see also S. Gupta, p. 55. She also compares the combination Agni- Ṣoma of the Ṛgveda as well (but note that both these gods are male!). --- Cf. also E. Garzilli, The Bhāvopahāra of Cakrapāṇinātha, (thesis Rome 1991, publ. forthc. 1992), comm. on śloka 9, on the Śakti-oriented Krama school and on the Kula, Pratyabhijñā, Spanda schools who stress Śiva as possessor of power. 39 Which is, among other things, also a symbol of Śiva. 40 Cf. Sanjukta Gupta et al., Hindu Tantrism, Leiden 1979, p. 152 (union of Vāgīśvara and Vāgīśvarī), J.Locke, Karunamaya. 41 An oval-shaped layer of grain, strewn on the ground between the Āhavanīya and Gārhapatya fires, in front of the Dakṣiṇāgni, only during the Dārśapaurṇamāsa rituals. It apparently substitutes for the Vedi, which is sprinkled with water in the Vedic version of the rite, in a line extended between these fires (in my interpretation, representing the Milky Way). envisioned as burning in the womb of the female Goddess Earth, represented either by a pit dug into the offering ground, or, as in the Patan Agniśālā, by the fire altars which, in turn, are constructed with clay bricks (iṣṭikā, fem.). What is poured into the fire altar, is ghee, a sy-bol of semen, and this is poured from a yoni-shaped offering ladle. This male/female dichotomy, which of course reminds of innumerable Vedic examples, is repeated many times in the ritual. If the sādhana of Agni originally was a strictly private affair, it became a public one at the moment it became a royal Agnihotra in the middle ages. Already the oldest inscription mentions princes as yajamānas or dīkṣitas. Some other priests, in addition to the Agnihotrī, always took part in the Dārśapaurṇamāsa and other fortnightly and monthly rites, and they still do so today, albeit rather erratically. This transformation is another case of Brahmanization or, perhaps, to employ the nowadays overused term, of Sanskritization (Srinivasan). While the early Tantric rituals were performed by any initiated adept for himself or his group, the Patan Agnihotra is a typical Brahmanical ritual, with Mantras in (Vedic) Sanskrit, and with an elaborate ritualistic system which only learned priests can understand. The wheel has turned again: as in so many cases, whatever had originated as a 'reformed' or popular ritual, was sooner or later taken over by the Brahmins and inserted into their framework of a highly specialized, difficult and Sanskritized ritual which is open only to members of the higher castes and which can only be performed by the priests themselves. Developments after 1600. By the time the Malla kingdom of Patan became independent42 from Kathmandu (1619 A.D.), the famous tāntrika Viśvanātha, who is believed to have been able to fly from Patan to Benares in a few moments, turned the Patan Agnihotra into a kind of state ritual. This was related to and performed for the King, who sometimes, for example at the time of the introduction of a new Agnihotrī, had to perform certain tasks. Actually, he is referred to in the ritual, and indeed is mentioned to this very day, though the present dynasty has not supported the Patan Agnihotra. Our earliest source for the ritual after the creation of the Patan kingdom in 1619 is an autograph of the Dārśapaurṇamāsa ritual written by the Tāntrika Viśvanātha himself some eighteen years earlier. This MS contains many notes on the ritual in Newārī language. However, the MS of 1433 was composed completely in Sanskrit. The ritual was, at that time, still a strictly priestly affair. Things were, as we have seen, to change: 42 See P. Burleigh, A chronology of the later kings of Pāṭan, Kailash IV (1976), p. 21-71. The MS of 1601 mentions the King, the country and its well-being, the wish for rain, etc. While the introduction of a Tantristic Agnihotra in the middle ages seems to constitute a clear case of Sanskritization that persisted until at least 1433, this is then counterbalanced by regionalization which is clearly visible already in Viśvanātha's MS of 1601 A.D. Still later, the population seems to have been allowed to take part. This is not attested in any document so far, but I assume that this happened as a result of the the conquest of the Kathmandu Valley by the Gorkhas in 1768/69. Indeed, the Agnihotrī mentioned in interviews, though rarely, that the Patan Brahmins had to conform to the new situation: they no longer received land and other grants from the new dynasty and had to appeal to the Patan people to take part in the rites. The effects of regionalization then grew. Gradually, the Rājopadhyāyas involved in the ritual lost much of their learning. They were now restricted to Patan proper: The new rulers of the Śāha dynasty had turned to other rituals, such as a Vedic revival, the Agnihotra of Paśupatināth, which was apparently instituted at the time of Pṛthvī Nārāyaṇ by the king of the neighboring state of Makwanpur. These developments naturally increased the trend to regionalization, one may even say, to communalism or tribalism... Both the last Agnihotrī as well as the present one had learnt the rite from a written text, a manuscript in Newāri viz. Devanāgarī script. "The book is my guru," the last Agnihotrī used to tell me - and therefore, in spite of his general patience and friendliness did not allow its filming.43 **** Two main discernable causes which shaped the Patan Agnihotra therefore are: first, a general development of religion from the Vedic and early Pauraṇic forms to the Tantric stage, accompanied by strong royal interest i. the middle ages, and secondly, a socio- economic situation in which the priests wanted to continue their ritual, --- why, that I will attempt to answer below. 43 Which could be carried out only in 1985 and formed the ultimate basis for this study. In the meantime, also in 1985, I found an older and perfect film of the MS that is used by the present Agnihotrī, Gurujvālānda, dating from the time when it still was kept with the his relatives at Bhaktapur, while it by now is sindura-smeared (and was partly burned, in 1985, by accident); in the colophon it is called Paṃcāgninityahomavidhi; containing 21 fol., written already in Devanāgarī, and dated in VS 1977 = A.D. 1820/1; filmed by the Nepal German Manuscript Preservation Project on Nov. 27, 1977 (unfortunately unknown to me, then) in the project's station at Bhaktapur under the title Agnihotrahomavidhi, film no. G 152/5. That this evaluation is basically correct, is noticeable from a comparison with the "official", traditional Vedic Agnihotra at Paśupatināth.44 This has been performed, at least since the Gorkhā conquest of the Valley in 1769 A.D., twice daily, in half-secrecy. No one comes to watch it, and only the priest offers into the fires. He is quite nervous when I ask him anything about the ritual in the presence of others. When he is alone, he is more open-minded, and he once even told me, - a rara avis indeed - how much money he received from the palace for the daily execution of the ritual. 3. Reasons for the preservation of the Agnihotra. But why has the Patan Agnihotra ritual survived until now, in spite of the lack of royal assistance since more than two hundred years? One may argue that, given the tenacity with which customs and rites are transmitted in South Asia to this very day,45 this is not too surprising. But why exactly this Patan Agnihotra, in its particular mixed (miśra) Vedic-Tantric form? Other older forms of rituals, whether Vedic or not, have vanished without trace. If other answers were not forthcoming, one could at least state that the Agniśālā and the God Agni, installed in aniconic form, - something very rare these days! - play a lively, and even a very important role in the contemporary religious thought and in the beliefs of the inhabitants of Patan. Local people tell stories, for example, of how one of the later Rāṇa prime ministers, early in this century, wanted to cut down the Varuṇa tree in the courtyard of the temple, and how the gods prevented this. They punished the carpenter who had attempted to do so with immediate death, caused by vomiting blood. Even recentl9, they say, snakes lived there, and they vanished only, - a bad omen, - just before the last Agnihotrī died.46 Or, whenever there is a big fire, this means that God Agni has left his house and burns outside.47 At such times, as in 1973 when the Singha Durbar Government Secretariat was burnt down in a big fire that lasted for a week, the fire in 44 For details see author, Agnihotra-Rituale in Nepal. 45 Cf. Gonda, Continuity and Change in Indian Religion, The Hague 1965. 46 Cf. Author, On Nāgas, in: Y.Ikari (ed.), Proceedings of the Kyoto Colloquium on the Nīlamata Purāṇa, Kyoto (forthc. 1992). 47 Note that this idea is already expressed by Ṛgveda 1.1.8, cf. Agnihotra-Rituale in Nepal, p. 182, n. 61. the Agni temple is believed to burn with very small flames.48 There is, apparently, much more to the Agnihotra than meets the eye of the casual observer. Apart from the Patan Brahmins, in Nepal only the "royal" Agnihotrins of Paśupatināth have performed the ritual continuously until today. This is understandable: the king wanted to protect himself and his country, and the priests had easy access to a steady income for the last 200 years, just by performing the Agnihotra twice daily.49 The case of Patan, however, this is different. After the conquest, and with the shift of the Agnihotra to the `National Shrine' of Paśupatināth, the Patan Agnihotrins were left to their own devices, successfully, as we already have seen: they were able to have the people of Patan visit their Fire Temple. Even then, it is not readily understandable that they continued the ritual. They could have argued that this function was now taken over by the new Agnihotra at Paśupatināth. Why bother to undergo the difficult, tedious and arduous observances of their own Agnihotra?50 Would its disappearance leave a big gap? There is much evidence which indicates that it indeed would. The Maṇḍala of Gods <12b> 48 One of my early friends and informants at Patan, Madan Rājopādhyāya, who had studied electronics at Bangkok, told me in 1975 that he had indeed checked the fires in the Agnimaṭh during the week the Secretariat kept burning (note the scientific frame of mind!), and indeed, the fire was very small then, just as expected. -- There are, of course, many more of such cases where present day experience re-enforces traditional beliefs. During the Summer of 1976, a big earthquake was predicted by astrologers to occur in the Valley; consequently, various groups tried to avert it by rituals: virtually all the Hindus of Bhaktapur, led by the Rājopādhyāyas, tried to do so with a lakṣahoma which lasted a whole fortnight; the Bajracharyas of Kathmandu did so near the Kumārī abode at Vasantapur palace with a one day śāntihoma; the Purohitas of the king, secretively, at the old Vasantapur palace itself; and so did the Rājopādhyāyas at Patan (who had been asked by the Royal Palace to do something as well). ass prophecied, the earthquake did occur during the astrologically determined night -- but it was deflected to ritual-less Northern China. Indeed, are all these rituals meaninglessness to the South Asian mind? 49 For details, see Agnihotra-Rituale in Nepal, p. 8 sq. 50 Nowadays some people indeed argue in this way. The present Agnihotrī and his wife took a lot of convincing before they to took up their office. The wife once told me, secretively, that she had "auspicious" dreams that made them take up this hard office and that subsequently "many good things happened to us." Local people phrase it somewhat differently: They claim that money was the convincing factor.  We have already seen that this ritual invokes all the gods of Patan and and its hinterland, - even the Buddhist god Āryāvalokiteśvara, who is celebrated each year in the big Macchendranāth chariot festival.51 This general invocation seems to effect an all-round protection of Patan. The gods surrounding a certain place are, however, invoked in normal homas as well.52 In the case of the Agnihotra, they are named in what turns out to be a particularly interesting arrangement, the reason of which will be in4eresting to discover. A look at the legends concerning the Agnihotra may help to discern the cause. The First Agnihotrī is believed to have come from the north-eastern mountain of Maṇicūṛ and to have taken his bath at the confluence of the Rivers Bāgmatī and the Nakhu (Prabhāvatī) at Cobhar, near the mountain Campādevī in the south-west of the Valley. Another legend, too, associates two corners of the Valley, the (north-)west and the south-east, i.e. the village of Jhūl53 near Thānkoṭ/Mātātīrtha in the west of the Valley where, according to the early 19th century chronicle (vaṃśāvalī),54 the Agnihotrins are said to have begun the original ritual, and, in the extreme south-east, at Gaur (Bengal) where they temporarily were exiled before returning to Nepal. <14> The intersection of both lines is either the Patan Agniśālā itself or the confluence of the Nakhu and Bāgmatī rivers near Cobhar, west of Patan. Still, the very confluence of the two rivers of Bāgmatī and Nakhu constitutes a center as well. Actually, any confluence is not only just one of two rivers but, mythologically, one of three, a Triveṇī of the streams of heaven and earth. The heavenly Gaṅgā (the Milky Way) is supposed to flow underground as well and join any important confluence of rivers: this confluence is, as it were, the center of the world.55 51 For details see J. Locke, Karunamaya, Kathmandu 1980. 52 Interestingly, including the local Buddhist gods as well, for example at one I witnessed and filmed at Kindol, Svayambhunath, in 1975, which included the local gods Ajimā (= Śītalā), Svayambhunātha, Sarasvatī (= Mañjuśrī), etc. These were included only after the Pūrbe Brahmin, who lived at Harigaon east of Kathmandu, had inquired about "local gods" from our gardener... 53 There is another(?) Jhor [jhūl] in the northern corner of the Valley, north of Tokhā. 54 Wright's Vaṃśāvalī, p. 158-9. 55 For the origins of this concept see author, La voie au ciel, Bulletin des Etudes indiennes, 2 (1984), pp. 213-279. Legend of the Kīrtipur Bhairava. Then, there is a legend about the Bāgh Bhairav (Vyāgra Bhairava) of the town of Kīrtipur, situated only three miles west of Patan. This Bhairava, represented in the form of a tiger made of clay,56 always tried to extinguish the sacred fires of the Agnimaṭha: Vyāghra Bhairava looks eastwards and faces the Agniśālā. He put out the fires so many times that, finally, the people of Kīrtipur built a wall in front of him so that he could not see the Agniśālā any more.57 Bhairava is, as is well known, a demonic form of Śiva while the main deity of the Agniśālā is, apart from Agni himself, Viṣṇu who is present in the Viṣṇukuṇḍa (Āhavanīya fire). However, in his more pacific aspect, Bhairava lies, as slab of stone, in front of the entrance of the Agniśālā, and he receives his share of the offerings each time the priest comes out from the temple at the end of the daily Agnihotra and places offerings in front of the tree and of other gods in the courtyard. He also receives part of the śrāddha food and the bali for the dogs (śvabali). This legend establishes, even at surface level, a relationship between one of the most prominent gods of the neighboring town of Kīrtipur and the Agniśālā.58 The fire ritual apparently is so important that even the inhabitants of Kīrtipur took care to preserve it. 56 Cf. R. Herdick, Kirtipur, München 1988, and in Formen kulturellen Wandels, 1986. 57 It would be interesting to begin an interpretation of this tale. It is well known that the king can be an incarnation of Bhairava, such as is the case, according to the Gopālarājavaṃśāvalī, is the case with Śivadeva. (More about this interesting king, by Dhana Bajra Bajracharya, in Contrib. to Nep. Stud.). Thus, the violent Tantric god Bhairava who often has his own fire, is opposed (as Bhairava of of Kīrtipur) to Patan in general and to the peaceful Tantric-Vedic fire at the Agniśālā which is identified with iṣṇu and which he tries to extinguish; again it is the peaceful Bhairava who lies in front of the Agnimaṭh, as a big slab of stone (see above, n. 31, 36, 57 ). Note that these Bhairavas can be regarded as an incarnation of Bhairava of Kāmarūpa (just like Śivadeva), and that the fire of the Agnimaṭh is supposed to have been bought from the same south-eastern direction, from Gaur (Bengal), at least according to the late Vaṃśāvalīs. We thus have a demonic Bhairava, identified with the king and outside of and opposing the (Buddhist) town of Patan, and a more peaceful Bahirava next to the peace- bringing fire of the Agnimaṭh, which is identical with Viṣṇu, the preserver, and situated at the rims of traditional Patan. -- On the other hand, the fire used at Kīrtipur's Indrāyaṇī temple has to be stolen from the fire ritual of Kathmandu (see B. van den Hoek, elsewhere in this volume). 58 It should be noted that a similar relationship also has existed, at least since the early 17th century, between the Cilanco Stūpa of Kīrtipur and the one in western Patan, as the work of R. Herdick (in progress) will show. This echoes the relationship between the Kathesimbu Stūpa at Kathmandu and the Svayambhunāth Stūpa. Pattern of exchange of priestly offices The same is also indicated by the complicated pattern of exchange of offices which the clans of the Patan Brahmins undertake upon the death of an Agnihotrī. There are five, originally six, clans (kawal) with various ritual duties. Only one clan, of course, can supply the Agnihotrī, but the others send their oldest member (thakāli) to the Agniśālā on the days of the New and Full Moon so that he may take part in the Dārśapaurṇamāsa ritual as hotar, udgātar, adhvaryu, brahman and as the (nominal) yajamāna. References to the exchange of offices can be found in some priests' diaries which are about 300 years old, among personal notes entered into ritual handbooks (thyāsaphu),59 and in interviews with the present day Agnihotrī and his clan.<17> According to this information, the various quarters of Patan where Brahmins live are involved in the exchange, and to this scheme, the role of the Agniśālā is, of course, central though the temple is not located at the geographical core of Patan. Yet, the exchange of the offices of the Brahmins associated with the Agnimaṭha crosses the two main intersecting streets which roughly run north-south and west-east,60 with the four well-known "Aśoka" Stūpas at their ends. Thus, should the Agnihotra be discontinued, this ritual interchange between the five clans of the Patan Brahmins could no longer function. Also, one must not forget tha4 the Newārī speaking Brahmins of Patan, the Rājopādhyāya, can only intermarry with the c. 100 Rājopādhyāya families at Bhaktapur, Kathmandu and in a few minor towns. Therefore, the exchange of offices concerns, in one way or another, most Newār Brahmin communities of the Valley.61 Restrictions of Movement <19> Another important aspect which helps to evaluate the importance of the Patan Agnihotra is the restriction of movement of the priest. Once a Rājopādhyāya Brahmin 59 Some of them are published by the local Deo Bhaju (Rājopādhyāya) Brahmins, notably Śrīṣanāṭh Śarmā of Puñcalī, Patan, in the small journal Rolamba, Journal of the Joshi Research Institute, Vol. 2, Oct.-Dec. 1982, no.4 p. 24 and Vol. 3, Jan.-March 1983, no.1, p. 12-13. 60 Actually, the alignment is NW/SE, etc., see N. Gutschow, Stadtraum u. Ritual, p. 154, and U. Wiesner, Zur Frage der vier sogenannten Aśoka-Stūpas in Pāṭan, Nepal, in Zur Kunstgeschichte Asiens, ed. R. Goepper, Köln 1977, p. 189-198. 61 For example, Viṣṇujvālānanda, the present Agnihotrī, is from Kathmandu and his wife from Bhaktapur. has become an Agnihotrī he is no longer allowed to cross rivers. The idea, again, can be traced back to Vedic times.62 The rule means, in fact, that he is restricted to Patan itself.63 In other words, the Agnihotrī priest seems to have been so important to the well- being of the medieval realm of Patan that he was not allowed out of it. Importance of the Ritual. Even nowadays, this feeling still persists: The Agnihotrins usually claim that the Agnihotra sustains the country. That is why the last priest, in talks with me, always stressed that his work is unselfish. He said: "What is being done here, that is done for the whole country; what one person does here, that is done for all; when someone sits here [and offers], he sits everywhere; when we pour water here, then all parts [of the country] get wet..."64 The Agnihotra ritual of Patan therefore was and is of vital importance for the well-being of the town of Patan, its surrounding hinterland, and also for the whole kingdom of Nepal. It was, and is, one of the few spiritual holdfasts of the Newārī speaking Brahmins and the population of a Newār capital that was left without the traditional ritual and ceremonial center after its conquest in 1768/9, with its royal court deserted and its royal rituals discontinued.65 This role of the Agnihotra is observable to this very day. The last Agnihotrī, for example, said: "A Brahmin's work is prescribed; we cannot only do what we wish... This ritual has to be continued... My father sat here and his father..." When he died in 1980 and a 3uccessor could not be found immediately, the people, including even the Buddhist goldsmiths (and occasional priests, the Śākya) and the farmers (Jyāpu), 62 See the Agnyādhāna fragments of the Kaṭha Br., ed. Caland, Brāhmaṇa en Sūtra aanwinsten, Verslagen en Meded. Akad. van Wetensch., Amsterdam 1920 p. 478 sqq.; Surya Kanta, Kāṭhaka- saṃkalanam, p. 3 sqq.; cf. also n. 103. 63 As the town is surrounded by rivers on all sides, except in the south, where the uplands of the old Patan kingdom are limited by the Mahābhārat mountain range. - In recent years the rule has been relaxed somewhat. Already the last Agnihotrī had made visits to his ill son at the tuberculosis hospital at Tokhā, north of Kathmandu. 64 Interviews with the last Agnihotrī, Gurujvālānand, in the Summer of 1979. 65 This is felt strongly even today. The Brahmins, at least, tell that the Newār king, as a kind of irredentist Barbarossa, still can be seen, occasionally appearing at the window of his palace, waiting for his return. interfered and proposed to start their own Agnihotra if the Brahmins did not agree to continue the ritual. This much was even related to me by one of the leading Patan Brahmins66 who also otherwise had not much good to say about the Patan Buddhists, and the same story was also told by some of the Buddhist priests (Bajrācharya) who are occasional visitors of the temple.67 Eventually, the Brahmins found a replacement from among their midst, though obviously not the thakāli of the clan who should have taken over the office. With the small revenue acquired from the visitors, and apparently also with some unspecified income from land that was not confiscated by the national religious trust, the Guṭhī Saṃsthān, the Agniśālā temple and the Agnihotra could be sustained until this very day, and a small income could be given to the Agnihotrī. This is really not very much, and the reward must be rather a spiritual than a mundane one: which spiritual one, an integral interpretation of the ritual may now show. 4. Interpretation The Vedic Agnihotra. Before coming to the interpretation of the medieval and the present day Agnihotra, a few facts must be repeated about the the Vedic form of this ritual: The original intention of the priests who conceived the Agnihotra, as it is alluded to in the Ṛgveda and as it must have existed even much earlier,68 apparently was that of a sun 66 They formed a committee which was to search for a new Agnihotrī, led by the Patan Rājopadhyāya Brahmins (Śrīśanāth, and others). 67 The reason for this again is historical. One of the sons of the famous Tāntrika Viśvanātha (attested between 1601-1666 A.D.) had become a Buddhist (cf. Wright, History p. 159), and even nowadays both parts of the clan still are required, people say, to perform their yearly ancestor worship together as members of a single guṭhi. 68 Fire ritual as such is, of course is much older: cf. the Iranian (Yasna Haptanhāiti), and even the Indo-European evidence of fire worship, cf.: Skt. hutám = Greek xutón = Germanic gudán > Old High German [daz waltand] got, Engl. god, a notion that requires offering of ghee or another ritual libation; (this equation has been noticed independently by K. Hoffmann and C. Watkins in the Sixties). Note also that Ved. agni-s and Latin ignis have masculine gender, while Hittite pehur, Greek pūr, German Feuer are neuter; the same situation pertains to "water" which is fem. in Ved. āpas, Latin aquā, Germanic *akwā, German Aa, Ache (in river names) :: Greek hūdōr, Hittite wadar/wed(a)naš, German charm. In other words, it was a rite that helped the sun through one of its critical points of "turning".69 As is well known, the sun is thought to disappear in the west below the Wasser, Ved. udan-/udr°. -- Even beyond the IE. area, fire worship is found connected with a decidedly male character of fire: In Japanese Shintō fire rituals, which must be investigated and evaluated separately from the Buddhist ho-as introduced only in mid-first millenium A.D., the fire god (Ho- musubi, etc.) is male. He is born from a woman, the Goddess Izanami, just as in the Vedic texts, where he is called mātariśvan, "the one who grows in his mother". Interestingly, there are many more connections between old Jpn., Vedic and other Eurasian mythologies (for which see author, forthc.). The Jpn. version, indeed, "explains" a Vedic riddle: The fire god Hi.no Yagi-haya-wo.no kami (or: Hi.no Kaga-biko.no kami, Hi.no Kagu-tsuchi.no kami), the last child of the primordial goddess Izanami, was carried like other gods by his mother as fetus, but while being born, he burnt her so severely that she died of her wounds. He consequently was cursed by his father Izanagi and then beheaded (Kojiki 1.8). -- Fire worship takes many forms in Japan; for example, in a yearly New Year ritual "God Fire" (thus the local people) "is carried" from a Shintō shrine (Kankura Jinja) at Shingū, Wakayama Prefecture, "down to earth" only by specially assembled, white clad males (the present writer included, in 1990) who are spiritually fortified with sake, an originally sacred drink; at the bottom of the hill Fire is received by a multitude of women and other spectators. --- It is known that there also existed a fire cult among the Mongols and other North Asian peoples (see Walter Heissig, The religions of Mongolia, Berkeley 1980, p.69 sqq., on connections with the Indo-Iranians, p. 76, and N. Poppe, Zum Feuerkultus bei den Mongolen, Asia Major 2, 1925, p.130-145; but note that the fire deity is female in Mongolian and Turkic texts). The relationship of these cults with the Indo-European ones, however, has not been described and determined so far. (For details, see author, in a forthcoming larger work on Eurasian mythology). -- Even more interesting for the early history of fire worship is the observation that proto-Indo-European pre-supposes the masculine viz. feminine gender only of the gods Fire and Water (this is attested, extra-linguistically, e.g. also in Jpn. mythology). Thus, the IE. word for "water" shows a clear suffix -h2 which is a feminine gender marker when compared with its putative Nostratic form (I quote the forms of Illich-Svitych, as arranged by Mark Kaiser, in Vitaly Shevoroshkin, Reconstructing Languages and Cultures, Bochum 1989, p. 125 sqq.): akwā < IE. *akwe- h2, added to Nostr. **•akʌ; Alt. āk(ʌ), Drav. āk, Afr.-As. `ku; and the masculine gender marker -s is added to Nostr. **Henkʌ in IE ṇgni-s, cf. Ural. *eṇkʌ "burn". -- However, the "elements" fire and water are regarded as grammatically neuter and are marked as such in IE by the very archaic neuter gender marker -r/-n, added to Nostr. **piHnʌ in IE. pehu-r "fire", cf. Alt. *pebʌ "hot, to dry" ( > Jpn. hi/ho "fire, light, sun"), Ural. *pīwe "hot, warm", Kart. *px(w)- "warm", Af.-As. *p`w "fire"; similarly, in the isolated word Nostr. **Hoṭʌ: IE. Het-r /*etH-r (> Avestan atar-, Ved. athar-), Alt. *otʌ. -- Beyond Eurasia, one can, of course, further compare the fire rituals of other linguistic and cultural areas, such as that of the Central African Pygmies, see E. M. Zuesse, Ritual Cosmos, Atlanta, Ohio 1979. 69 Other such 'rites of passage' of the sun through the year are: New/Full Moon (for the moon), Cāturmāsyas (for the seasons), Soma ritual (New Year), Gavām Ayana, with its viṣūvat (summer solstice) and mahāvrata (winter solstice) days. The nucleus of these rites may well have been to follow the sun (= Vivasvant, Mārtāṇḍa, asau āditya) and moon (Somo Rājā, King Soma) through the year w)th all its vicissitudes. Notably, the 'turning around' of the fire which is represented by the sun at earth in the evening, to travel underground eastwards, and to re-appear next morning from its underground passage at the eastern horizon; this is similar to Old Egyptian myth, where the sun has to fight `the dragon of the deep' each night.70 Or, the sun is believed to have two sides, a dark and a bright one: at night, the su. travels eastwards with its dark side turned downwards, towards the earth; in the morning, it reverses its course, now with its bright, shining side turned downwards.71 In both concepts of the nocturnal path of the sun there is the danger: that of no return or that of not turning back towards the earth, and there thus looms the danger of continuous darkness, which has to be warded off, for example by performing Agnihotra-like rites. But there are other elements, such as that of making the milk flow which is in the heavenly river, the Vedic Sarasvatī and the later 'heavenly Gaṅgā.'72 This river goddess brings fertility to men and easy childbirth to women. As is well known, in the Veda and up to this very day, it is a man's duty to beget a son to insure the survival of the line of progeny (prajātantu),73 symbols of which are the moj string, the mekhalā cloth,74 and the yajñopavīta cord of the initiated.75 This connection of human beings with the gods is effected by means of the tantu of Agni, the "string" extended from earth to heaven with the help of the fire and its smoke, as has been expressed since Ṛgvedic times in winter solstice is a myth that is found in many versions and rituals in various Eurasian civilizations, see author, The myth of the Hidden Sun, (forthc.) 70 Cf. the Ṛgvedic Ahi budhnya, RV 1.186.5, 4.55.6, 5.41.16; 6.49.14 etc. 71 See E. Sieg, Der Nachtweg der Sonne, Nachr. Ges. Wiss. Göttingen, 1913. 72 This is a nightly but, ultimately, also a seasonal feature: note the seasonal progression of the Milky Way, for which see author, Sur le chemin du ciel. Bulletin des Etudes indiennes 2 (1984), pp. 213-279 and author, forthc. (in Jpn.). In my opinion, the Milky Way is symbolized in the Agnihotra by the Vedi and especially by the pouring of a water line from the Āhavanīya (= heaven) to the Gārhapatya (= earth), e.g., at Kātyāyana Śrautasūtra 4.13.16. Note that this takes place in the evening when the heavenly river 'descends' down to earth (for details, see BEI 2) and that the ground between the fires is called samudra "ocean"! (see below, n. 87). 73 Connecting a son with his father (see TU 1.11, Kaṭha Śikṣā Up. 11). 74 The moj (< mauñja girdle) is worn by the Agnihotriṇī, see below (on Agastya and Arundhatī, and B. van den Hoek, elsewhere in this volume), while wearing the mekhalā is a rather old Vedic custom (Kāṭhaka GS, etc.), substituting for the umbilical cord. 75 Establishing a sort of umbilical cord with the Ṛṣis, see author, in a forthcoming study. The yajñopavīta, of course, refers to the spiritual line of progeny connecting man to the Ṛṣis. stereotype fashion.76 In addition, the relationship of men with their ancestors is stressed by the Agnihotra offerings to the pitṛs; their 'spiritual' ancestors, the Ṛṣis, are similarly 'fed' by the recitation of poetry and by the offering of food. All of this indicates that the 'classical' Vedic Agnihotra, like most other Vedic Śrauta rituals, is a combination of several older, (pre-)Ṛgvedic rituals.77 This can also be noticed through a more formalistic approach: The classical Agnihotra is a combination of a Ṛgvedic, homa type fire ritual (of keeping the fires alive overnight) and an old milk offering to the Aśvins, made in a metal kettle. This rite has been preserved independently, in an artificial but archaic form, as the Pravargya ritual, in which the metal kettle of Ṛgvedic times has been exchanged by one of clay, made by the priest himself in an extremely primitive 76 See for example, RV 10.57.2 tantur deveṣv ātataḥ; cf. also the primitive Mantra tantur asi VSK 6.2.9, TS 3.5.2.3, KS 17.7 etc. 77 This aspect has not been stressed sufficiently enough. For a possible methodology that has not been tested by students of Indian ritual, and which would include the detailed investigation, at each subsequent historical level, of Mantras, performance, procedural variants and their 'theological' discussion, in rituals that are attested early on (RV, AV, YV-Mantra collections) in order to arrive at the earlier stages of a particular ritual, see author, review of: J.Gonda, The Mantras of the Agnyupasthāna and the Sautrāmaṇī, Amsterdam 1980 in Kratylos XXVI (1981/2), pp. 80-85. -- Strangely, the earlier, i.e. Ṛgvedic stage usually is neglected in contemporary ritual studies, while it indeed represents the `pre-classical' period and while many of the early verses (mantra, Avestan maOra 'thought') used in the rituals of the RV have been taken over into the later ritual. (The Mantras, incidentally, have, like most poetry, at least one and mostly more than just one meanings, on the various levels of everyday and poetical language; these are, again, just as in poetry, often difficult to trace by the modern philologist). -- Finally, we must not overlook the possibility and even the probability of several other developments or 'reforms' in the shaping of the pre-Vedic ritual. The prominence given to the Asura/Āditya cult by the Vedic Indians and the Iranians is one of these early features. Dualistic sets of gods are also known from other early Asian and European religions, but there is no groups such as that of (mostly, personified) concepts like Varuṇa, Mitra, Aryaman (note the irregular, artificial formation with a primary suffix -man attached to an adjective or noun, see Wackernagel, Ai. Gr. II 2 p.764 §607!), Bhaga, Aṃśa, Dakṣa, etc. This group includes, incidentally, the gods who are crucial for the emergence of the world and of humankind, Vivasvant/Vīvahuuant or Mārtāṇḍa/Gayō Mar5tan (and Indra, in some versions). Especially their social function must be taken into account when defining their role in early Aryan religion and society. there is the strong probability of a change taking place when the Indo-Iranians encountered the sedentary, agricultural people of the Bactro-Margiana Archaeological complex, in the late second millenium B.C. -- All of this is indicative of a long evolutionary process rather than of a single brief 'reform' of Aryan ritual, a snapshot of which we witness in the RV and in the (older) Avestan texts. However, a major step in the shaping of the 'classical' Vedic ritual took place in the little studied 'dark period' following the RV which can be investigated (which has not been done) on the basis of texts such as AV, RV Khilas, and the YV Mantras. This change was motivated, in my considered view, by political and social change, a process that will be described in more detail in a forthcoming paper (Proceedings /f the Toronto conference of Archaeologists and Linguists, October 1991, ed. by George Erdosy, forthc.) and intentionally archaic fashio., without a potter's wheel. Both the ancient fire and milk offering rites are joined in the Agnihotra, just as the Pravargya is later on (optionally) inserted into the Soma ritual. *** The Tantric Agnihotra. Against this historical background, the medieval Tantristic Agnihotra of Patan emerges as being of similarly multi-valent meaning. Quite obviously, the protective role of Agni within a complicated maṇḍala of Tantric gods surrounding the Agniśālā, the town of Patan, and the Kathmandu Valley is of great importance. Agni provides protection against the intrusion of all kinds of demons,78 unregulated powers that threaten persons and society, and the state, - as well as a similar state of mind of the individual. One of these forces is Bāgh Bhairav from Kīrtipur. Bhairava is a member of the host (gaṇa) of Śiva or even one of his aspects. As was seen earlier, he is very much present right in front to the gate to the Agniśālā and he is regularly appeased there by a bali offering. In his more terrible aspect of a tiger he threatens the sacred fires, even at a distance, from his shrine at Kīrtipur. Such forces are kept at bay. The Agnihotra and the Valley of Kathmandu. <21> * Deśapāta. However, God Agni at his Agniśālā of Patan is located, as we have already seen, at the center of a larger set of deities that surround him. These gods all of whom occur, with their appropriate and often Vedic mantras, in the deśapāta (deśabali) offerings can be located on the map in three concentric circles, each of them divided into two semi- circles, the dividing line of which runs from north-west to south-east, through the Agniśālā. They represent the two old opposing halves of the world, for example those of the gods and that of the fathers in Vedic times79 - or here, rather the male and the female halves of the Valley, which we also find represented in the town of Kīrtipur,80 in 78 See Robert Levy, How the Navadurgā protect Bhaktapur. The effective meaning of a symbolic enactment, in Heritage of the Kathmandu Valley, p. 105-134, and now, in his new book, Mesocosm, Berkeley: Univ. Calif. Press 1991. 79 Cf. Kuiper, Varuṇa and Vidūṣaka, p. 55 sqq.; cf. above, n. 29. 80 See R. Herdick, Kirtipur, München 1988, p. 97, 100 sq. and in: Formen kulturellen Wandels, p. 255. the Agniśālā temple itself, or in the two halves of a Newār kitchen <21> cf. <24>. Such an interpretation agrees with the late medieval legends81 about the origin of the Patan Agnihotra, which indicate a connection of the north-east and the south-west of the Valley. If stress is put on the concept of two moieties, then (N)E82 opposes (S)W.83 The origin of the fire in the north-east is n/t surprising. Not only is this the direction of the gods, and the (south)east generally is the direction of Agni, but the concept is also found in the comparatively late, medieval Agni Purāṇa, which often can be seen to be close to the ideas found in the Patan Agnihotra, including the identification of Agni with Viṣṇu in ch. 271, while ch. 34 deals with the homa ritual at some length.84 * Uttaravāhinī and the human body 81 Datable at c. 1630 A.D. and later, partly in connection with the famous Tāntrika Viśvanātha. Some of them are collected in Wright's Vaṃśāvalī, p. 158-9; see author, Agnihotra-Rituale in Nepal, p. 160. 82 Cf. Agnipurāṇa, ed. Pt.s Śrīyukta-Pañcānana and Śrīyukta-Vīṛasiṃha and Srīyukta-Dhīrānanda, Calcutta (Kalikāta), Śaka 1812, ch. 40.1-2 on the Puruṣa: tad devair nihitam bhūmau sa vāstupuruṣaḥ smṛtaḥ | catuḥṣaṣṭipade kṣetre īśaṃ koṇārddhasaṃsthitam...; 34.36 "his head is in the east... the feet in the north-west and south-west" (prācyāṃ śiraḥ samākhyātaḥ ... jaṅghe vāyavyanairṛte); Agni Pur. 58.1, about digging a fire pit in the north-east at the time of a homa accompanying the establishment of a Viṣṇu figure in the maṇḍapa of a temple (aiśanyāṃ janayet kuṇḍaṃ gurur vahniḥ). 83 Another line connects the (north-)western village of Jhul and Bengal in the south-east which can be interpreted, as has been done above, as a device aiming at making Patan and the Agniśālā the center of two crossing lines. 84 Agni Purāṇa 270.15-16 (Calc. ed. 271.15 sqq., transl. by Manmatha Nath Dutt, Calcutta 1899-1904, p. 998): "The mighty, victorious Viṣṇu is manifest in the fire and in the sun. As fire, he is the mouth of the gods..." (Viṣṇur jiṣṇur bhaviṣṇuś ca agnisūryādi rūpavān | agnirūpeṇa devāder mukhaṃ Viṣṇuṃ parā gatiṃ ||15|| vedeṣu sapurāṇeṣu yajñamūrtiś ca gīyate | āgneyākhyaṃ purāṇan tu rūpaṃ Viṣṇor mahattaram | āgneyākhya-purāṇasya karttā śrotā janārddanaṃ), Agni Purāṇa 34.35 sq.: "Fire is the instrument of emancipation... His head is described to be in the east..." (vimukteḥ kāraṇam vahniḥ paramātmā ca muktidāḥ ||35|| prācyāṃ śiraḥ samākhyātaḥ bāhū koṇe vyavasthitau | īśānāgneyakoṇe tu jaṅghe vāyavyanairṛte ||36|| udaraṃ kuṇḍam ity uktaṃ yonir-yonir vidhīyate.) -- Note that Agni Pur., other than its name implies, mainly deals with Viṣṇu as its main deity (ch. 1.11, 271. 17 sqq., etc.) and with his worship, always stressing the use of (Tantric) homa ceremonies, (almost always with one fire only; such a Tantric homa, addressed to Śiva in the Śaivaite part of the text, is found in the rather detailed ch. 75). Even in the homas addressed to various gods, Viṣṇu comes first (66.7). The text, as it stands now, proclaims to be one of the Madhyadeśa Brahmins following the Pañcarātra doctrine but excludes Brahmins from Kaccha, Kāverī, Koṅkoṇa, Kāmarūpa, Kaliṅga, Kāñcī, Kāśmīra (ch. 39.6). On the problem of composition, see B.B. Mishra, Polity in the Agni Purāṇa, Calcutta 1965, p. 9, 24-26.  Within this scheme, the position in myth of the Agnihotra temple near an uttaravāhinī, a river flowing north, is necessary. This was consistently stressed by the late Agnihotrī. It is provided here by the Nakhu (Skt. Prabhāvatī), one of the few rivers of the Valley flowing northwards.85 <14> In the foundation myth of the Agnihotra ritual, the Agniśālā is situated at the confluence of this river with the southernbound Bāgmatī. The staff of the Agnihotrī, who took his bath at this confluence, persists to this very day as the Varuṇa tree in the temple courtyard (called Varuṇa-Nāga,86 correctly: *varaṇa tree). And indeed, as the priest always insisted, the ground is so damp87 that one has but to dig a few inches to draw water; in fact, it colors the lower part of the altars by constant seeping. The priest does not mention, of course, that the ground is sloping this way anyhow, and that there is a big water tank at the back wall of the temple. This confluence of the Bāgmatī and the Nakhu, a triveṇī, symbolizes that of the three 'mystical' arteries in Tantra. This is the Yogic feature of a symbolical bath of the adept at the confluence of two rivers, i.e. the 'arteries' piṅgalā and iḍā88 inside his body. They are joined there by a third stream, the suṣumnā.89 On Earth, this is the third river, 85 Note that the other one is Karmanāśā, the boundary between Buddhist Patan and Hindu Bhaktapur, just as the North Indian Karmanāśā is the boundary between Āryāvarta and polluted Magadha. The position of the river is, of course, the same as that in Northern India: a northwards flowing river that joins one flowing on a W/E axis; but the question to be answered is which territory in the Kathmandu Valley was regarded as homologous with Magadha? It must have been that of Bhaktapur, a much more Hindu oriented kingdom than that of Patan or even Kathmandu. 86 For Varuṇanāga see also Pott, Yoga en Tantra, p. 90 n. 24 (on the Kalaṅka-śmaśāna), and p. 102 sq., as head of a group of eight Nāgas and king of the Yakṣas (engl. transl.: Yoga and Tantra: their interpretation and their significance for Indian archaeology, The Hague 1966); cf. author, Agnihotra- Rituale in Nepal, in Formen kulturellen Wandels und andere Baiträge zur Erforschung des Himalaya, ed. B. Kölver u. S. Lienhard, St. Augustin: VGH Wissenschaftsverlag 1986, pp. 157-187. 87 Note that the ground between the fires is called samudra "ocean"! 88 She also is regarded as the daughter of Mitrāvaruṇāu, cf. below on Mitrāvaruṇa; on iḍā and piṅgala, see Sanjukta Gupta et al., Hindu Tantrism, Leiden 1979, p. 59. 89 Cf. Sanjukta Gupta et al., Hindu Tantrism, Leiden 1979, p. 145; p. 57: "The doctrine that the human body corresponds to, is even identical with, the universe is seldomly systematically expounded but cearly always self-understood." --- The author of the Newārī booklet on the Agnimaṭha, Aiśvayadhar Śarmā, stressed the same in a small, two page summary (in Nepālī), which he had prepared for me in 1985 when I asked him about his opinion on the Patan Agnihotra: The ritual is both an antaryātra and a bahiryātra. the underground viz. heavenly Gaṅgā or Sarasvatī which -unseen - joins t(e Bāgmatī and Nakhu at Cobhar. If this confluence, which in Tantra is situated at the navel, is the place of the Agnihotra, then the head of the Vāstupuruṣa, who underlies the sacred geography of the Valley,90 is in the (north-)east, his feet in the (south/north)-west, cf. Agni Purāṇa 34.36 prācyāṃ śiraḥ samākhyātaḥ ... jaṅghe vāyavyanairṛte (and cf. also ch. 40.2.) A localization of inner, somatic and psychic features and of processes of religious experience is thus transmitted into the geographical surroundings of Patan, into the well-known localities of the Valley. Here, ritual serves to identify body and nature, it makes this identity visible in one's own surroundings, and it allows one to experience it, for example in pilgrimages (yātra) to the various tīrthas and pīṭhas of the Valley. <15> * Deśapāta (II) It was pointed out earlier that the gods mentioned in the deśapāta list are arranged in (semi-)circles and as such, form two opposing units. Yet another interpretation is possible as well, if one does not look at the list with the bias, so natural in the Indian context, of expecting pradakṣiṇa type movements, that is (concentric) circles, established by an sun-wise, clockwise movement. <21> If one, instead, locates and connects with each other all the deśapāta gods on a map of the Valley, this results in a rather strange and completely unexpected pattern,91 which can be interpreted as the 90 Just as any offering ground; for a picture see M. Slusser, Nepal Mandala. A cultural study of the Kathmandu Valley, Princeton 1982, Vol. II, pl. 108; -- cf. Kurukṣetra as a devayajana "sacred offering ground of the Gods" in Vedic texts, or in medieval Hinduism, the Doāb of the Yamunā and Gaṅgā as the antarvedi, and Allahabad as Prayāga "fore-offering" or perhaps rather "foreward / foremost offering (place)", where one could gain access to heaven by ritual suicide through drowning oneself at the confluence of Yamunā and Gaṅgā. Note also that Agni Purāṇa (111.4) regards the confluence at Prayāga as the vulva of the earth goddess, and as a place where the three sacred fires are found. -- On inner (Tantric) and outer geography, see Sanjukta Gupta et al., Hindu Tantrism, Leiden 1979, p. 57, 60. 91 The same exercise results, for the first time, in an explanation of the 600 odd Nāga locations enumerated in the Nīlamata Purāṇa of Kashmir, vss. 881 sqq. (ed. de Vreese, Leiden 1936). They seem to be mentioned at random, -- quite famous Nāgas next to clearly insignificant local ones. The enumeration, geographically arranged, results in a multiple clockwise listing of all Nāgas of the Kashmir Valley, see the Proceedings of the Kyoto Seminar on the Nīlamata Purāṇa, ed. Y.Ikari, forthc. 1992. shape of the letter expressing <23> the sacred syllable oṃ. Significantly, this is traced not in the North Indian (Nāgarī) style, with the candrabindu92 attached to the u-shape letter but appropriately, in the usual Nepalese fashion, and placed on the map far away towards the right upper (i.e. north-eastern) corner, at Dolakhā, in a province east of the Valley. <24> - From here onwards, as required by the shape of the letter in Nepal, the enumeration of the gods describes a counter-clockwise movement (a-pradakṣiṇa, prasavya), resulting in several circles around the letter oṃ, with an obvious back and forth movement. The enumeration returns several times to the center of the ritual, the Agniśālā. If <22> one draws a map of these movements it will be discovered that they form an 8-petal or perhaps even a 64-petal <26> lotos, a feature typical of a maṇḍala.93 2. Town If the Agnihotra of Patan thus becomes the center of the whole Valley, the focal point of a complicated maṇḍala, then the same principle can be expected to exist in the town of Patan as well: <18> The relations established at the time of the death of an Agnihotrī through the exchange of offices by the Brahmin clans involved (see above) create two triangles which intersect in ṣaṭkoṇa form (a star of David). This figure symbolizes the male god and his female śakti in productive union, in creation, as is befitting in the representation of Viṣṇu / Lakṣmī that is found in the Agnimaṭh (see below). Agni in his temple constitutes the center of the diagram. This conforms with the position of the bīja-mantra in Tantra, where the Mantra from which all develops is found at the center and is surrounded by the various geometrical designs and concentric circles of gods forming a maṇḍala. The pattern involving the Agnimaṭh is datable at ca. 1630 A.D. It is obvious that it does no longer fit the modern location of the six Brahmanical clans but their movements and relocations within the city since then 92 The dot and halfcircle (candra-bindu) denoting the nasalization. -- For a representation of this symbol, in connection with the candrabindu (sun, moon) and the flame (which represent the three secret cakras), see Pott, Yoga en Tantra, p. 40, and cf. p. 36. 93 One may, of course, also interpret this simply as a continuous centrifugal/centripetal movement. But then, one would have to explain the repeated <24> return to the Agniśālā. A centripetal/centrifugal movement as such is rare, but not unknown. A centrifugal one occurs, for example, in enumeration of countries in the Milindapañha, see author, Early Eastern Iran and the Atharvaveda, Persica IX (1980), 86-128; in Nepal, there exists a centripetal enumeration of the major Śiva sanctuaries beginning and ending ending with Paśupathināth, see Nepālamahātmya, ch. 29 and Himavatkhaṇḍa, ed. Yogī Naraharināth, pariśiṣṭa 1, with a list of some 166 tīrthas and 64 liṅgas, arranged in this fashion, and cf. N. Gutschow, Stadtraum, p. 19 and n. 22 (p.195), referring to the Bhāṣa Vaṃśāvalī and another Vaṃśāvalī from Sankhu as sources. are well documented in the thyāsaphus. The older pattern used here is based on these sources and the descriptions of the present day Rājopādhyāyas.94 <26> 3. Town and Valley. We have seen that the pattern of exchange of the offices symbolically unites the town of Patan, and especially their Brahmanical community. The general population naturally also takes note of the exchange and in this way also is involved: after all, it is the oldest Brahmin (thakāli) from their quarter of town who has become Agnihotrī. If this pattern is compared to the one created by the gods mentioned in the deśapāta offerings, it can be observed to form an overlay of the innermost sections of the letter oṃ: the new pattern circumscribes the diagram with oṃ at its center yet once again. <25> 4. The Agniśālā Similar patterns can be recognized if one now looks at the layout and the rituals of the Agniśālā itself: Though basically a Vedic setup of the three plus two sacred fires, there are a few additions, such as the Maṇināgakuṇḍa, the Mitrāvaruṇa, etc. More important in this context is their relative position. Again, they form overlapping triangles, of fire and water: <27> D.A. Maṇināga--------Samudra / \ \ / / \ \ / Āhav.------Gārh. Nāga The three most important Vedic fires thus form such a triangle, if re-interpreted in a Tantric fashion, as representing Brahma, Viṣṇu and Rudra (Śiva). One cannot be sure, yet there might also be a water triangle constituted by three wells and tanks. And, if one wishes to continue along these lines, it can be noted that there are three Gaṇeśa shrines, too, which are closely related to the Agniśālā. Well Cobhar 94 It is interesting to note that the present Agnihotrī still remembers the complicated pattern of exchange of offices, though he does it not get it right every time he attempts an enumeration.  (tunthi) Gaṇeśa ---------------- Jala Gaṇeśa / \ shrine \ / / \ (Kāryavinayaka) / Nāga tank ------ Tank \ / (on platform) AGNIŚĀLĀ \ / Gaṇeśa stone *Sūryavinayaka At any rate, we can safely establish an innermost triangle of Fire (the three main fires, Brahmā, Viṣṇu, Rudra Kuṇḍa) which is intersected by the one connected with Water (Maṇināga, Samudra, Nāga). These two designs, in turn, might perhaps be regarded as being surrounded by a triangle of the three Gaṇeśas, intersected by a Water triangle, constituted by the well and the two tanks. Such diagrams can be found to magically constitute the "plans" for many Nepalese (or Indian) towns.95 The sacred geography of Patan itself is based on a diagram involving the ten Mahāvidyās, eight Gaṇeśas, four Bhīmasenas, four Nārāyaṇas, and the eight Aṣṭamātṛkās beyond the rim of the settlement.96 It should be noted that a simple male/female opposition, as apparent at the innermost location of the Agnimaṭha diagram, is not a,ways in evidence in the "ground plans" of the towns of the Valley: There are, for example, three Gaṇeśas in the diagram underlying the city of Bhaktapur.97 To sum up: The Fires and other altars or figures of gods in the Agniśālā constitute Tantric triangles, which intersect and represent the powers which, in concert, are active in the Universe. The resulting Tantric diagram represents the gods and powers, which constitute and sustain the town of Patan and the Kingdom of Nepal. 5. The ritual 95 See N. Gutschow and B. Kölver, Ordered space. Concepts and functions in a town of Nepal, Wiesbaden 1975, and N. Gutschow, Stadtraum und Ritual der newarischen Stadte im Kathmandu-Tal. Eine architekturanthropologische Untersuchung, Stuttgart 1982. 96 See Niels Gutschow, Stadtraum und Ritual, p.165. 97 See B. Kölver, A ritual map from Nepal, in Folia Rara, Festschrift W. Voigt, ed. by H. Franke et al., Wiesbaden 1976, pp. 67.  Turning now to an interpretation of the ritual and its various aspects, it can be observed that the Agnihotra which is based on this maṇḍala and takes place at its center, the Agniśālā temple, provides stability. Of the three gods of the Hindu trinity, it is its sustaining aspect, represented by Viṣṇu, which is stressed by his identification with the main fire, the square Āhavanīya/Viṣṇukuṇḍa. Most offerings, and especially all those of the visitors to the temple, are made into this fire. As indicated above, it often is the Agni Purāṇa which seems to be closest in its concepts to the Tantric Agnihotra of Patan. The fire in the (Viṣṇu)kuṇḍa is identified with Viṣṇu himself by this Purāṇa:98 "The mighty, victorious Viṣṇu is manifest in the fire a.d in the sun. As fire, he is the mouth of the gods... As manifest in the form of the sacrifice, he is praised in the Vedas and Purāṇas." Other passages confirm this interpretation, notably Agni Purāṇa, ch. 34.32-35: "Having meditated on Lakṣmī in her menses, (who is situated) in the midst of the kuṇḍa, he should offer an oblation. Prakṛti (of the three guṇas or qualities) is called the Lakṣmī of the kuṇḍa. (32-34). Viṣṇu is the origin of all creatures... Fire is the instrument of emancipation...(35)"99 The central act of this Tantristic sacrifice therefore consists of a secret, mental identification of the priest with the fire in front of him, and with Viṣṇu who is the Fire and the Sun.100 Fire is male, the kuṇḍa is female, with a yoni-shaped 'mouth': Into this altar made of bricks (iṣṭakā, again a female element), some ghee (= sperm) is offered 98 Agni Purāṇa 270.15-16 (Calc. ed. 271.15 sqq.) viṣṇur jiṣṇur bhaviṣṇuś ca agnisūryādi rūpavān | agnirūpeṇa devāder mukhaḥ viṣṇuḥ parā gatiḥ ||15|| vedeṣu sapurāṇeṣu yajñamūrtiś ca gīyate. āgneyākhyaḥ purāṇan tu rūpaḥ viṣṇor mahattaram | āgneyākhyapurāṇasya karttā śrotā janārddanaḥ | cf. transl. Dutt, p. 998. 99 Transl. Dutt, p.136; kuṇḍamadhye ṛtumatīḥ lakṣmīḥ sañcintya homayet | kuṇḍalakṣmīḥ samākhyātā prakṛtis triguṇātmikā ||34|| sā yoniḥ sarvabhūtāṇāṃ vidyāmantragaṇasya ca | vimukteḥ kāraṇaḥ vahniḥ paramātmā ca muktidāḥ ||35|| cf. Agni Purāṇa 270.15-16 in the preceding note. -- An even more detailed description of the actual process of identification of the adept with the god and the fire is given by Agni Purāṇa 74-75, especially at 75.46: yāgāgniśivayoḥ kṛtvā nāḍīsandhānam ātmanā... This, however, deals with Śiva, not with Viṣṇu. 100 Note that this Tantristic identification recapitulates the old Vedic and Iranian one of Fire = Sun, which in fact can be traced much beyond the Indo-Iranian period and can be found in other Asian religions as well (see my work on Eurasian mythology, forthc.; and note Jpn. hi "fire, sun, light"). with the yoni-shaped ladle (female). Through these multiple male/female unions, Agni is then conceived and born from the womb of the earth, and has to undergo the various rituals of passage (saṃskāras).101 When the two forces of Agni and his śakti are brought into productive union e.g., by pouring ghee with the yoni-shaped ladle, (symbolized by the diagram in the form of a star of David), the world is sustained. Here, at this level, perhaps the one with the deepest meaning of this Tantric Agnihotra, lies its efficacy. It is now understandable that the Patan Brahmins --whether they remember these concepts (cf. n. 89) or not-- never wanted to give it up: this ritual stabilizes their families, the caste, the town and the entire kingdom. 6. The Priest and his wife. The central role of the Agnihotra and the Agniśālā within the geographical context of the Valley, or in other words, within the Tantric maṇḍala surrounding it, can again be observed with the actors of the ritual as well: One case in point is the restriction of movements of the Agnihotrī, as already referred to above, to which the priest, and to some extent also his wife, have to conform. * Restriction of movements. They are not allowed to cross the samudra situated between the Dakṣināgni (Rudrakuṇḍa) and the Āhavaṇīya (Viṣṇukuṇḍa) but must, as they indeed complain, always take the long clockwise route around the Āhavanīya and the Gārhapatya fires.102 <28> As the Āhavanīya or Viṣṇukuṇḍa is the most important fire of the ritual, one can again detect an opposition between this Fire and the Water of the samudra. The two elements are represented, better, they are personified by the priest and his wife. And their movements are restricted accordingly. 101 This process is described in detail, for example in the old Pañcarātra Saṃhitās such as the Jayākhyā Saṃhitā, paṭala 15 or in the Śaivaite Somaśambhu-Paddhati of Kashmir, ed. J.D. Zadoo, KSTS 73, Srinagar 1947, p. 29, vs. 261 = sect. IV, vs. 46 in H. Brunner, transl. p. 260 ; it is also found, for example, in the yearly re-consecration of the figure of the Patan god Macchendranāth, see J. Locke, Karunamaya, Kathmandu 1980, p. 208 sqq. 102 This reminds, of course, of the restriction put on the Agnihotrī, forbidding him to cross a river, see above n. 63-64 and below n.104.  Outside the Agniśālā, the priest, now Fire Incarnated, cannot cross Water as well: he must not be exposed to the danger of "extinction" posed by rivers. This again reminds of a Vedic prescription.103 The rule concerning the Patan Agnihotrī is based, apparently, on the fear of a loss of magic power, of heat (tapas), similar to that which could occur when crossing a river. The strict rule of celibacy, which is to be kept by the husband and wife104 who perform the service of Agnihotrī and Agnihotriṇī, is reinforced because the two elements must not unite outside the ritual but must be kept in a state of suspension. This is clear, as far as the two humans are concerned, even during the performance of the ritual. How far it was the case in the middle ages or whether it is a late development remains unclear: there is no overt kaula Tantra left in the Kathmandu Valley.105 The gods, however, do unite in the ritual. And both the priest and his wife help to do so when they churn the first fire, the baby Agni, of the newly established Agnihotra. Indeed, the last Agnihotrī, in an interview taken in 1979, overtly described himself as Agni. * Other aspects : Agastya/Arundhatī Yet, while he is Agni, his wife is called Arundhatī. According to the Brahmins' explanation, Arundhatī is the śakti of Mitrāvaruṇa. She carries a moj (mauñja) grass belt around her waist, as a reminder of the string of the fire drill (mithāsi). Indeed, she and the Agnihotrī start the first fire of their tenure in the ancient way of using a fire drill. Consequently, she is said to carry fire in or on her belly. Now, Mitra and Varuṇa procreate, according to old Ṛgvedic mythology, as Mitrāvaruṇā(u) their pot-born son Maitravaruṇa, i.e. Agastya.106 In the Agniśālā he is represented, however, by single male figure visible as the Mitrāvaruṇa figure, emerging 103 Kaṭha-Brāhmaṇa, as partially edited by W. Caland, 1920, p. 1, see above, n. 62 . 104 Just as in the Vedic Agnihotra, the person offering the ritual must be married (but sexually abstinent, cf. below n. 102 ). When his wife dies, he has to give up the ritual. 105 The medieval MSS of the Agnihotra do not mention overt sexual acts during the ritual. However, some quasi-kaula Tantra is said to be performed still by some Rājopadhyāyas, though only with their own wives. And some Tibetans, for example at Svayambhunath, were said to do so as well. 106 See RV 7.33.10,13; cf. Agastya = Maitrāvaruṇi, Kumbhayoni at Mbhār. 3.8776, 8797, 8807, 12.343 (13216), 13. 4471, 4786. from a vessel, --note the grammatically wrong form, instead of *Maitrāvaruṇa.107 He is clad in a red cloth, and always positioned in front of the Āhavanīya fire (Viṣṇukuṇḍa). Arundhatī actually is the wife of Vasiṣṭha, another 'pot-born' Ṛṣi, who is born at the same time as Agastya (= Māna), already according to RV 7.33.13. Obviously, there is a confusion between the two Ṛṣis.108 Note that the Mantra addressed to Mitrāvaruṇa during the Patan Agnihotra speaks of the seven Ṛṣis indeed. If Arundhatī is Mitrāvaruṇa's female partner, his śakti, then the Mitrāvaruṇa statue must represent Agastya. The Agnihotrī is, as we shall see, his descendant, i.e. another *Maitravaruṇa / Mitrāvaruṇa. This outcome is also reflected in the identification of the Agnihotrī with Agastya and Mitrāvaruṇa on one hand, and of his wife with Arundhatī on the other: ---------------------------------------------------------------------------------------------------------------- *Maitrāvaruṇa (Agastya) = Mitrāvaruṇa statue/fire drill = Agnihotrī <28> 109 Arundhatī (*Lopāmudrā) = moj string around the waist = wife of the and the plate of the fire drill Agnihotrī ---------------------------------------------------------------------------------------------------------------- 7. Mitrāvaruṇa <29> The statue of Mitrāvaruṇa is made, according to the last Agnihotrī, from the wood of the Varuṇa (correctly: *varaṇa) tree which grows in the temple courtyard. 107 The name of this "god", which at first is puzzling, thus is a simple grammatical mistake, a neglect in composing a proper vṛddhi form, *maitravaruṇa- "belonging to Mitra and Varuṇa." 108 Note, however, that, according to Mbhār. 3.232, Agni who had fallen in love with the Pleiades (kṛttikāḥ, the wives of the seven Ṛṣis), entered their Gārhapatya fire and eventually cohabited with all of them, except Arundhatī (who creates another pot-born being, Skanda). Note that the rule of sexual abstinence of the Agnihotrī (= fire, the sun) and his wife (thought to be Arundhatī) agrees with this myth. -- Note also that iḍā, one of the "arteries" inside the body of the Yogin (Agnihotrī) is regarded as the daughter of Mitrāvaruṇau. 109 Apparently representing the string of the fire drill used, on the drill plate, during the first churning of the fire of a new Agnihotrī. The fire drill represents the Agnihotrī, in Vedic fashion -- see the Purūravas and Urvaśī Mantras in YV rituals (VS 5.2, MS 1.27:16.7; KS 3.4, etc., and well described at ŚB 3.4.1.22: their son was Āyu); on this topic cf. also B. van den Hoek, elsewhere in this volume. Legend tells that this particular tree had sprung up from the staff (daṇḍa)110 of the "First Agnihotrī", who had descended from the Maṇicūṛ Mountain in the north-east of the Valley to take his bath at the confluence of the Nakhu and Bagmati rivers (Skt. Prabhāvatī, Vāgmatī). The First Agnihotrī necessarily has to be the clan ancestor of the present day Patan Agnihotrins. The staff became his only remnant after his death and cremation, as by that time it had already become the Varuṇa tree which remains at the Agniśālā to this day. The Mitrāvaruṇa figure made from it therefore represents the ancestor and the lineage god of the Agnihotrī clan of Patan.111 In this way, ritual produces a repetition of actions in 'primordial time',112 which in the present case, however, means only c. 1630 A.D. The God Fire and the Agnihotrī, coming from north-east113 (Mt. Maṇicūṛ) are represented by the Viṣṇukuṇḍa and the Mitrāvaruṇa figure in front of it. All of this is still visible even today in the setup of the Agniśālā and in its relation to the Valley surrounding it. <30> It can now be understood why Mitrāvaruṇa leaves the temple towards the end of the ritual: The figure, representing the first Agnihotrī and being carried by the present one, repeats the 'walk' of the first Agnihotrī. He begins114 at the confluence of the two 110 Note that the Varu/aṇa tree corresponds, in Tantric myth, to the tree at the confluence of the Gaṅgā and Yamunā at Prayāga (see above, n.90).The Agni temple is thus located at the center of the earth, where according to Agni Purāṇa 111.2) the three sacred fires are found. -- Cf. the similar legend about the founding of Patan (Wright, History, p. 135 sq., Hasrat, History of Nepal, p. 43 sq.): A thirsty grass cutter rammed his carrying pole into the ground, while looking for water. On his return, it had become immovable. 111 Note that the worship of lineage deities (digudyaḥ) and of the ancestors is firmly established among the Newārs of the Kathmandu Valley. Consequently, the worship of the Agnihotrī ancestor is one of the most solemn occasions in the course of the year. A special worship of this ancestor takes place on the Śrāvaṇa Pūrṇimā day (= Guhnu Puhnī, Ṛṣitarpaṇī, Kumbheśvaramelā, Rakṣabandhana, Nep. Janaī Pūrṇimā) in August when all priests as well as the Mitrāvarūna figure, the Agnikuṇḍas, etc., receive, in the pavitrārohaṇa rite, a new yajñopavīta-like cord, made of many strands (apparently 27, 3 x 9; cf. Agni-Pur. 37); each strand again has 3 strings. 112 This is also indicated by Wright's Vaṃśāvali (p. 159) which mentions a Prajñāpāramitā MS. of 188 A.D. in this context. 113 Cf. above, Agni-Purāṇa, ch. 34, 69.1 and 40.2: "Īśa (the Lokapāla of the north-east)" (and above, notes 82, 84). 114 There are other parts that have already taken place before, at the initiation of the new Agnihotrī, where he repeats the walks "down from the north-east", from the ghāṭ at Śaṅkhamūla, on the Bāgmati, north of Patan. The later part takes place at his death: from the Agniśālā towards the south- rivers Bāgmati (Vāgmatī) and Nakhu (Prabhāvatī) at Cobhar. First, he 'greets' the Varuṇa tree, that is the staff he put into the sands at the confluence while taking his bath at the the Nakhu/Bagmatī confluence near Cobhar. He then proceeds to the Gaṇeśa which represents the Jalavinayaka of Cobhar, and then to the Bhairava slab in the courtyard, that is the Bāgh Bhairav of Kīrtipur, and Siddhilakṣmī, Kumāra. Finally he approaches the Nāga next to the door of the Agniśālā, that is the famous Kārkoṭaka Nāga who lives in the small pond of Taudāha. This is situated just next to Cobhar and is regarded, in myth, as the last remnant of the beings in the lake that once filled the Kathmandu Valley,115 (cf. below, section 10). 8. The Agnihotrī In spite of his identification with Agni, with the Sun, and with Viṣṇu, the Agnihotrī, on an everyday level still has very personal wishes. They are voiced at the beginning of the ritual, before the gurumaṇḍala rite.116 " Oṃ, homage to Brahmā-Viṣṇu-Maheśa (Śiva) ... and to the Maṇika-King of Nāgas (snakes)117 in the form of sacrifice. Oṃ, may there be success at the beginning of the ritual. May there be growth and increase of crops. May there be well-being in (my) body, and peace in my house. (May there be) abolishment of all hindrances and (may there be) good luck which creates complete tranquility (śānti). (May there be) a son of long life and desire (kāma), and increase of incessant (santati) riches."118 west, to Nakhu ghāṭ at Cobhar, (followed later on by a return towards the north-east, to the pitṛtīrtha at Gokarṇa). 115 See the account by M. Slusser, Nepal Mandala, p. 353 116 For the Buddhist version of this see J. Locke, Karunamaya and D. Gellner, IIJ (forthc. 1992); cf. also the booklet Gurumaṇḍalārcaṇa-pustakam, ed. by Amoghavajra Vajrācārya, publ. by the Bajracāryya Saṅgha of Kathmandu (Saṃkoṭā Pres, Àebahā¿, Kathmandu, NS 1092), republished in NS 1101 by Naṃdakumārī Vajrācārya. 117 Cf. the foundation of a subterranean Nāga pond, Maṇi-talāva, according to Wright's Vaṃśāvalī, p. 135 sq. in the center of the town of Patan; cf. Gutschow, Stadtraum, p. 148. 118 Oṃ Brahma-viṣṇu-maheśāya savitreṣaṇmuṣ(kh)āya ca | Maṇiko-nāgarājāya yajñarūpāya te namaḥ | oṃ siddhir astu kriyārambhe vṛddhir astu dhanāyaśe | puṣṭir astu śari(ī)re (')stu śāntir astu gṛhe mama | sarvavighna-pra(ṇā)śanaṃ sarvaśāntikaraṃ śubhaṃ | āyu(ḥ)-putraḥ ca kāmaḥ ca lakṣmī-santati- varddhanaḥ | tatra deva vidhātānāḥ śāntir bhavatu vāraṇaḥ |  It is obvious that the usual wishes of an average Hindu are also those of this priest: prosperity, peace or tranquility, a son, more surprisingly, gaining entry into Heaven. Heaven, it is true, is not mentioned in the saṃkalpa-like text quoted just now, but is it stated to be the Agnihotrī's aim even today, just as it had been in the middle ages. In fact, there is a fragment from the Śatapatha-Brāhmaṇa, which is inserted into the Agnihotra MS of 1433 A.D., after a short description of the daily Agnihotra, which refers to the identification of the (agnicayana) sacrifice and the sun.119 In the 15th century, this statement apparently was important enough to be dug out from a, by then, quite obscure ritual, the Vedic Agnicayana, and from a bulky ancient Vedic text, the Śatapatha-Brāhmaṇa, which was no longer studied to any degree in medieval Northern India and Nepal. 9. Offerings, ash If the Agnihotrī himself exhibits personal wishes an goals in this ritual that is intended for the benefit of king and country, other mundane goals can be expected to be 119 The MS reads (uncorrected!): oṃ namo pañcāgnaye namaḥ || pradakṣinā | (A) athādhiyajñam: yad etan maṇḍalan tapati || ayaguṃ śa rugmā. (B) -tha yādatad arccidāyātmahāvratan nāni śāmāni sa śāmnālor (C) tha yeṣu ye tasmin maṇḍala śa vo yajṇaṃ hiraṇmaya | puruṣan tatheti devetraguṃ saṃskṛtohopapadyate tad yajuṣevāna guṃsthām ati (D) evādhiyajñam || This is a bad transcription of ŚB 10.5.2.6: (A) athādhiyajñam | yad etaṃ maṇḍalan tapaty ayaṃ sa rukmo. 'tha yad etad arcir dīpyata idam tat puṣkararaparṇam āpo hy etā āpaḥ puṣkaraparṇam (C) atha ya eṣa maṇḍale puruṣo 'yam eva sa yo 'yaṃ hiraṇmayaḥ puruṣas tad etad evaitat trayaḥ saḥskṛtyehopapadhatte tad yajñasyaivānu saṃsthām ūrdhvam utkrāmati tad etam apy eti ya eṣa tapati tasmād agniṃ nādriyeta parihantum amutra hy eṣa tadā bhavatīty u (D) evādhiyajñam | 6 | Cf. 10.5.2.1: (A) yad etaṃ maṇḍalan tapati| tan mahadukthaṃ tā ṛcaḥ sa ṛcāḥ loko (B) 'tha yad etad arcir dīpyata, tan mahāvratan tāni sāmāni sa sāmnāṃ loko 'tha ya eṣa etasmin maṇḍale puruṣaḥ so 'gnis tāni yajuṃṣi sa yajuṣām lokaḥ | 1 | Cf. also ŚB 4.1.1.25 amuṣmin vā etaṃ maṇḍale 'hauṣīt | ya eṣa tapati tasya ye raśmayas, te devā marīcipās, tān evaitat prīṇāti, ta enaṃ devāḥ prītāḥ svargaṃ lokam abhivahanti || "Now (the explanation) with regard to the sacrifice: For he has offered this into that circle which burns up there, and the rays of the circle are the gods sipping light sparks: it is these gods that he thereby gratifies, and gratified in this way, the gods make him pass to the heavenly world." -- verse 7: athādhyātmam ... yad etan maṇḍalaṃ tapati yaścaiṣa rukma idaṃ tachuklam akṣann atha yad etad arcir dīpyate yaccaitat puṣkaraparṇam idaṃ tat kṛṣṇam akṣann atha ya eṣa etasmin maṇḍale puruṣo yaścaiṣa hiraṇmayaṃ puruṣo'yam eva sa yo 'yaṃ dakṣiṇe 'kṣan puruṣaḥ \7\ observable as well. Taking even a brief look at the materials offered during the ritual such aims are visible indeed. During the Agnihotra it is not just ghee which is offered. In most cases it is also grain, a product of the earth which rather unexpectedly also is a symbol of seed.120 Ghee is offered as well, and this has always been a clear symbol of human sperm.121 These offerings to the gods are transformed during the ritual by being offered, with the help of a mediator, often a god such as Agni, but in this case Agni personified, the Agnihotrī, into the God Fire, in the Āhavanīya/Viṣṇukuṇḍa fire. One can compare this with the old Ṛgvedic `paradox' yajñena yajñam ayajanta devāḥ (RV 1.164.50). But what do the Patan people and the priests receive as the remnants of the gods' food, as ucchiṣṭa, or in modern Hindu terms, as prasāda? Is it rain as spittle of the gods who have touched and eaten part of the offerings? Rain, next to children, would be most important to them as most of them are rice farmers. Rather, rain comes from megha, the clouds, and these are holding their urine/sperm, falling (devo mehati) down to earth as varṣa.122 Do the people also receive sperm from the gods? We have seen already that this was one of the possible interpretations of the Vedic Agnihotra, in which they send down milk and sperm. And this may be another reason why women throng at the temple daily. Indeed, "ghee = sperm" is already an old Vedic equation. The gods, however, also leave other remnants of their food for mankind, especially the Agnihotra ash. This ash is so very 'sacred' and, certainly, auspicious, that it is not thrown away but given to the visitors as tilaka mark on the forehead or, in small 120 Most cleary expressed already in the Yajurveda, e.g. KS 6.3:51.9: Sūrya and Agni were in same yoni. Sūrya rose upwards, lost his seed. Agni received it with metal pan, made it stick to it, and transferred it to the cow where it became milk. - "Rice grains are the milk of the bull; the bull is that sun. If one performs the Agnihotra with rice gruel, then one offers that sun alone." (... anaḍuho vai payas taṇḍulā. asā ādityo 'naḍvān. yad yavāgnihotraṃ juhoty, amum eva tad ādityaṃ juhoti.) -- Cf. also the story about the origin of rice and barley, as stemming from the asu and medha of the offered Agni, kimpuruṣa, aśva etc. down to rice and barley (VādhB, ed. Caland, AO VI p. 116 sqq. = Kleine Schriften p. 416 sqq.); similarly AB 2.8. -- Compare also the Upaniṣadic cycles of water/food/seed ChU 5.10.6, cf. 6.8.3 sqq., Mait.Up. 6.10. 121 For example reto ghṛtam KS 26.7, KpS 41.5, reto vai ghṛtam ŚB 9.2.3.44, retaḥ siktir vai ghṛtam KB 16.5. 122 Cf. the co-existence of the words vṛṣ "to rain", vṛṣan "male, man, bull", vṛṣabha "manly, bull" :: ṛṣabha "bull". portions, as medicine. (Note the character of the tilaka as return of a gift to the gods of grain (tila), as a kind of prasāda.123) Otherwise, the ash is stored upstairs in the temple, to be immersed into the river only at the cremation of the Agnihotrī, which takes place, not at the Patan ghāṭs, but at the confluence of the Bāgmati and Nakhu near Cobhar. The ash, or rather the transformed rice offerings, touched and tasted by the gods, are thus returned, by the means of the Triveṇī confluence, to their own heavenly realm, that is to the place from where they had come through the medium of rain.124 It is apparently because of this reason that the samudra (representing the Vedic vedi) is actually filled by rice grains during the Dārśapaurṇamāsa ritual at the Agnimaṭh.125 Will they become the new body of the Agnihotrī in the next world? Both the ash of the cremated body and that of the rice offerings are immersed into the river together at the Triveṇī of the Nakhu, Bāgmatī, and the heavenly viz. subterranean Gaṅgā/Sarasvatī. The probability of this interpretation can be supported by other evidence. The cremated bodies of other people, however, are reconstituted by immersing their ashes into various Tīrthas, at confluences of rivers, which perhaps may be regarded as having been arranged roughly in the form of a body.126 <15> At the Agnihotrī's cremation, both his ashes and those of the grains that had been offered in the Agniśālā are united again to form a new body for him. This new, divine body is, as it should indeed be the case for a representative of Agni on Earth, superior to the one of normal people, the ashes of whose bodies are only 'offered' into the rivers. In this way, by joining the sacred Agnihotra ashes with his own cremation ashes, the 123 Other varieties are the common one of red sindūra paste and, sometimes, of blood as well; note the substitution of sindūra for blood, and grains for the life-sustaining blood. For details, see a separate investigation, (forthc.) This has no connection with the use of sindūra in pūjā and its supposed etymology (see Mayrhofer, KEWA). 124 This is an old Vedic idea as well, cf. the well-known cycle of water/rain found already at RV 1.164.51 and in the Yajurveda Saṃhitās, (cf. E. Frauwallner, Geschichte der indischen Philosophie); for the origin of grain in VādhB (above, n. 120) . 125 But not at other Agnihotra rituals in the Kathmandu Valley which always follow the Vedic norm of the Kātyāyana Śrautatasūtra. For details see author, Agnihotra-Rituale in Nepal, 1986, pp. 157-187 and a forthcoming monograph. 126 See R. Herdick, in Heritage of the Kathmandu Valley, 235-275; cf. Pott, Yoga en Tantra. Agnihotrī will receive a new body for his life in heaven.127 He symbolically also receives sperm because of the ghee that is offered into the fire (and has reached heaven as smoke and smell).128 This will ensure his getting offspring even after death, through his son and his grandchildren.129 On a personal level, therefore, the Agnihotrī effects a continuation of his line through his grandchildren, and the acquisition of a place in heaven, with a new superior body. This is, incidentally, exactly the same aim still as that of a Vedic sacrificer. For, unlike what we regard as the common attitude to life after death in the Indian subcontinent, many if not most people, whether Buddhist or Hindu, believe that they can go to Heaven130 and they wish to stay there as long as possible. Rebirth is their 'second choice' only. It is precisely for this reason that the yearly ancestor rituals (śrāddha) still are performed for one's deceased parents, grandparents, and great- grandparents. 10. Body and ritual topography. If life, in the Valley of Kathmandu, ends with the return of the ashes of the cremated to the rivers, and thus, to the gods, and if this is done at various locations in 127 Cf. the ŚB quotation (ŚB 4.1.1.25) mentioned above, n. 119 . In medieval and modern Hinduism, the body of a deceased is re-created in heaven by offering the śrāddha rice balls during the first ten days after death (see P. Gaeffke, JAOS 105, 275). -- In my experience, in Nepal as elsewhere in the subcontinent people, whether Hindu or Buddhist, usually wish for a life in heaven and worry little or not at all about rebirth. 128 On these terms, see author, on rebirth (forthc.). 129 Note that the Agnihotrī, in the above saṃkalpa-like text, wishes his son a long life. 130 Often not understood, as how can rebirth agree with śrāddha offerings, cf. the Buddhists' criticism, who nevertheless have their own śrāddha ceremonies, see the edition by Amoghabajra Bajrācārya, Piṇḍa Vidhānam. Śrāddha vidhiḥ ṣoḍaśapiṇḍa sahitam, Kāntipur (Kathmandu : Pūrva Seva Khalaḥ), NS 1024. - Note also the term divaṃgata "deceased". the Valley, arranged according to the old ritual vāstupuruṣa pattern,131 this again indicates an identification of the natural surroundings with one's own body.132 Ritual stresses this in several ways: in the Agnihotra by identification of the Agniśālā and its mythical origin with several locations of the Valley. Likewise, for the Agnihotrī: his origin is the north-east, his final dissolution will take place in the south- west, at the ghāṭ of Cobhar. The identification of the rivers with the 'arteries' of the body (in Yoga and Tantra) lead to the same conclusion: external topography is identical with the internal structure of one's body. Ritual makes all of this visible, - very visible in case of the immersion of the cremation ash into various confluences, less visible in the Agnihotra ritual itself, yet obvious in the legends about its origin, and very visible if not immediately obvious, in the corresponding walk of the Agnihotrī through the courtyard. Ritual, in this way, not only identifies man and his surroundings, it also makes both manageable (and in so far is, again, similar to Vedic ritual). It serves as mesocosm to connect the microcosmic aspect of man with the macrocosmic one of the deities. The Patan Agnihotra thus produces wealth, children, rain, and general well-being for king and country. Gurujvālānand, the last Agnihotrī, confirmed this by a sentence which was, at the time, still incomprehensible if not nonsensical to me: "If we pour water here, it wets everything in Nepal... What is being done here, is done everywhere". Both the Vedic Agnihotra and the Tantric Agnihotra of Patan therefore are multivocal; they have not just one but many meanings, depending on the various levels of understanding and religious experience; these are quite different ones to its performers, local participants, onlookers and the foreign observer.133 131 According to the interpretation of R. Herdick, (Neue Kulte, in Heritage of the Kathmandu Valley, and Kirtipur, p. 144 sqq., see above, n. 29 @), the various tīrthas involved are arranged in the shape of a body. 132 For the identification of the inner Tantric and the outer geography, see Sanjukta Gupta et al., Hindu Tantrism, Leiden 1979, p. 60. 133 Even the local people are aware of it. To some "pūjā is beautiful", but they may a,so say that the ritual is effective in driving away the influence of the bad planets; others may stress the worship of Viṣṇu and Lakṣmī; still others may stake it for one of Śiva and Ūmā, see M. Slusser, Nepala Mandala, p. 226 (incidentally, only M. Slusser and Aiśvaryadhar Śarmā --on my instigation--, see note 35, have written about the ritual at all); the priests still know some of the Tantric background. -- The ritual is meaningless in the sense of Staal only if we isolate small sections of it and regard them out of context,  But the Tantric ritual one includes all of the following: psychic and religious experience, projection of it in terms of the priest's body, a.d transposition of it outside his body in terms of the geographical surroundings, the Kathmandu Valley. - At the same time, the "coming of the first Agnihotrī", i.e. the Yogic experience of the "original" Tantric who installed the ritual, is repeated daily by taking his representation and the present lineage god, the Mitrāvaruṇa figure, on a tour of the courtyard of the temple. Performing the Tantric Agnihotra ritual every morning, the Agnihotrī identifies himself with God Fire, the Sun, and with Viṣṇu, the preserving aspect of the Hindu Trinity. - By offering to Agni/Viṣṇu, he sustains his own community, that of Patan and that of the Kingdom of Nepal. Yet all of this still takes place in the garb of the ancient Vedic, some 3000 year old Agnihotra sacrifice. NOTES * This paper is based on research financed by the German Research Association (Deutsche Forschungsgemeinschaft) in 1985; it was first given as lecture at Harvard University and at the University of Pennsylvania in February 1986, and subsequently at the Collège de France, Paris in 1989 and at Kyoto in 1990. I present it here with some modifications, additions and changes. such as he does with bits and pieces of language. Of course, a/an, taken out of the context of a sentence and of the English language are meaninglessness (as even small children discover), but within a phrase or a sentence, the appearance or non-appearance of -n obviously is meaningful. -- On the so-called "meaninglessness of ritual", see now Scharfe, and note, A. W. MacDonald in L'Ethnographie, LXXXIII, numero spécial: Rituel Himalayens, p. 9.