Notes

n.1Among the four Tibetan translations included in the Degé Kangyur, only Toh 592 lacks the introductory narrative. It shares this feature with the earliest extant Sanskrit version reported in the Khotanese manuscript and with a number of Tibetan translations preserved among the Dunhuang manuscripts.

n.2In most but not all citations of the spell’s name‍—either as the title of a text or as a reference to the spell within a text‍—the term “name” (nāman; zhes bya ba) immediately follows the compound sarvatathāgatoṣṇīṣasitātapatrā, thereby marking it as the primary title of the deity/spell/text. In a small number of instances the word nāman is omitted altogether, and in rare cases it is shifted to a different point in the title. The instability of the title both across and within the texts contributes to the challenge of interpreting it properly.

n.3Though all Tibetan translations of the title consistently insert “born from,” they are anything but consistent in their use of “all tathāgatas” (sarvatathāgataº; de bzhin gshegs pa thams cad). Many of the Tibetan translations of the title omit “all,” thus reading “the uṣṇīṣa of the Tathāgata,” even when the Sanskrit title provided reads sarvatathāgataº. All Sanskrit sources consulted are unanimous in reading sarvatathāgataº.

n.4A challenge in accurately translating the title of the texts and the spell they contain is the fact that the terms pratyaṅgirā and aparājitā are used in other contexts as names of female deities and their spells. Here the terms aparājitā and pratyaṅgirā are taken as adjectives of māhāvidyārajñī or dhāraṇī , depending on the version of title used in a specific text. The decision to render it in this way is not meant to be definitive.

n.5For a synopsis of these forms, see Porció 2000, pp. 14–16.

n.6The manuscript containing the Sitātapatrā spell discovered by Aurel Stein at Dunhuang is written in a “cursive Gupta script” that Hoernle (1911, pp. 448–49 and 472–77) argues was in use beginning in the sixth century in Central Asia. He has also identified loan words used in the manuscript that would suggest it dates to the eight century. Additionally, a male deity named Sitātapatra is mentioned in The Root Manual of the Rites of Mañjuśrī (Toh 543, Mañjuśrīmūla­kalpa), where he is counted among the eight uṣṇīṣa kings (uṣṇīṣarāja; gtsug tor kyi rgyal po). See Dharmachakra Translation Committee, trans. (84000: Translating the Words of the Buddha, 2020), g.1525. The Root Manual of the Rites of Mañjuśrī can be dated to at least the eight century (see ibid., i.2).

n.7On the circulation of apotropaic Buddhist literature in Inner and Central Asia, see White 2021, pp. 45-84.

n.8About which see below.

n.9Orgyen Lingpa 2001, folio 207.b.

n.10See for example the relatively recent work by Dudjom Jikdral Yeshé Dorjé (bdud ’joms ’jigs bral ye shes rdo rje), the gtsug tor gdugs dkar mo’i rgyun khyer ’bar ba’i thog brtsegs (1997), in which the author incorporates phrases, words, and spell formulas from the canonical material into an otherwise distinctive practice manual.

n.11Chökyi Jungné, dkar chag , folio 149.b.

n.12The colophon reads in full, “This was translated and finalized by the great scholar from Jagaddala [Monastery] in eastern Tibet, who is favored by the illustrious Cakrasaṃvara, by the paṇḍita Vibhūticandra, and by the monk-lotsāwa Sherap Rinchen. It surpasses the earlier versions for being translated in coordination with the commentarial literature and carefully edited by checking it against various manuscripts from Magadha in India.” (Folio 193.a: rgya gar shar phyogs ’dza’ gata ta la’i mkhas pa chen po/ dpal bde mchog gis rjes su bzung ba’i paN+Di ta b+hi b+hu ti tsan+tra dang / lo tsa+tsha ba dge slong shes rab rin chen gyis bsgyur cing zhus te gtan la phab pa’o/ ’di ni sngar gyi dpe rnams las khyad par du ’grel pa dang bstun zhing bsgyur ba dang / rgya gar yul dbus kyi dpe du ma dang gtugs te/ shin tu dag par byas pa yin no.)

n.13Toh 591, c.­1: paN+Di ta pa ra hi ta b+ha dra dang / lo ts+tsha ba gzu dga’ rdor gyis kha che’i bdud rtsi ’byung gnas kyi gtsug lag khang gi dpe rnying la gtugs nas gtan la phab pa.

n.14’phags pa de bzhin gshegs pa gtsug tor nas byung ba’i gdugs dkar mo can gzhan gyis mi thub pa. Denkarma F.302.a; see also Herrmann-Pfandt 2008, pp. 197–98.

n.15Toh 593, c.­1: gtsug tor lha yul ma chung ba kha che’i paN+Di ta ma hA dz+nyA na nas rang ’gyur du mzad do. This colophon identifies the translator as Mahājñāna, but this is certainly a reference to the Kashmiri paṇḍita Mahājana, who was active in Tibet in the eleventh century and translated a number of other works. The colophon to the same translation in the Phukdrak Kangyur correctly identifies him as Mahājana. For a synopsis of the life of Mahājana, see Kano 2016, pp. 5–8.

n.16A Catalogue of the Urga Kanjur, p. 280, folio 54.a. The same source also identifies Toh 592 (Urga no. 593) as the version “known as the Uṣṇīṣa of the human realm, or the longer of those of the heavenly realm” (mi yul ma’am lha yul ma che bar grags pa), but the precise meaning of this statement is uncertain, as Toh 592/Urga 593 is slightly shorter than Toh 593/Urga 594. This way of referring to Toh 592 could not be confirmed in any other sources.

n.17See the full citation below.

n.18Sönam Nampar Gyalwa, gtsug gtor gdugs dkar rgyas pa rig sngags kyi rgyal mo chen po, p. 736: gzungs kyi rtog pa ’di la/ rgyas pa rig sngags kyi rgyal mo chen mo zhes bya ba sngon ’gyur byang med pa ’di dang / ’bring po mchog grub mar grags pa zu dga’ ba’i rdo rjes bsgyur ba dang / bsdus pa lha yul ma chung bar grags pa kha che ma hA dza nas bsgyur ba dang gsum du zad kyi/ chung ba gzhan zhig snang ba ni lha yul ma chung bar grags pa’i gleng gzhi dor ba tsam du zad pas zur du bgrang mi ’tshal lo.

n.19Kawagoe 2005, p. 19.

n.20Herrmann-Pfandt (2008, p. 198) also considers it likely that the Sitātapatrā text recorded in the Denkarma (ldan dkar ma) is a version of Toh 592.

n.21Both Mahājana, the translator of Toh 593, and Parahitabhadra, the translator of Toh 591, were from Kashmir. The colophon to Toh 591 also indicates that it was prepared on the basis of a manuscript found in Amṛtabhavana monastery.

n.22Toh 590 may also be linked to Kashmir, as Kashmir is directly referenced in the body of the text. The deity Mahākāla is described as “residing in a great charnel ground in Kashmir” (kha che’i dur khrod chen po na gnas pa). Among the canonical translations, this reference is unique to Toh 590, but it is also attested in the more recent Sanskrit witnesses.

n.23See Hoernle 1911, pp. 448–49.

n.24This manuscript is available digitally from the University of Cambridge Digital Library and has also been edited and published in Hidas 2021.

n.25Three other Nepalese versions of the Sitātapatrā spell were consulted for this translation. Royal Asiatic Society Hodgson 77, dated to 1894, preserves a unique witness of the spell that is generally aligned with Toh 590 but contains a number of variants. It is also the most corrupt of the Nepalese sources consulted. University of Tokyo Library no. 441-01, dated to 1828, is generally similar to Cambridge Add. Ms. 1326 and thus Toh 590. Finally, a version of the spell from an undated Nepalese dhāraṇī­saṅgraha was edited by Ngawang Samten and Janardan Pandey and published in volume 33 of the journal Dhīḥ. It too correlates with Toh 590, perhaps more so than the other Nepalese sources.

n.26Kiliç Cengiz and Turanskaya 2019, p. 20.

n.27About these versions of the spell, see Kiliç Cengiz 2020, and Kiliç Cengiz and Turanskaya 2019 and 2021.

n.28A summary of these texts and conjectures about their relationship to the canonical materials can be found in Lalou 1936 and in Porció 2000, pp. 19–24.

n.29Reading legs ldan here and throughout as bhagavat following the Sanskrit.

n.30It is not clear who the “five great mudrās” (mahāpañcamudrā; phyag rgya chen po lnga) are in the context of Nārāyaṇa.

n.31In the corresponding passages in the Sanskrit sources, this final clause is treated as a separate object of homage: “Homage to the one accompanied by the horde of mātṛs” (namo mātṛgaṇasahitāya Dh33, CL1326, UTM 441-01; oṁ namo mātṛgaṇavāndena sāhitāya RASH 77).

n.32This Sanskrit rendering is tentative, as the Tibetan term used here differs slightly from the Sanskrit sources. All Sanskrit sources consulted have Vikasitakamalotpalagandhaketurāja, a term that aligns with the Tibetan translation of the name given in Toh 590, pad+ma rgyas pa dang ut+pa la’i dri’i tog gi rgyal po.

n.33In the Sanskrit versions aligned with this text, as well as in Toh 590, the verbal statement “I will teach…” (pravakṣyāmi; rab tu brjod par bya) is given here, with the title of the spell as the object of the verb. Toh 591 lacks this verb and does not provide a finite verb until the end of the verse section that follows.

n.34In this series of verses it is difficult to determine what is a descriptive phrase and what is a proper name. Both the Tibetan and Sanskrit sources are ambiguous in places, and some of the names/descriptive phrases are repeated. Thus, the parsing of this sequence of proper names, epithets, and adjectives that follows is tentative. The Sanskrit sources clarify that all the terms are in the feminine nominative singular, apart from the first few lines, about which see the following note.

n.35This translation follows the Tibetan translations in rendering the preceding lines in verse. It is in prose in all Sanskrit version consulted apart from KT728, in which the original structure is unclear from the published edition. According to the Sanskrit syntax, each of these descriptive phrases is in the accusative case (apart from KT728), marking them as adjectival phrases that construe with pratyaṅgirāṃ (the “averting [spell]”) as the object of the verb “teach,” which is absent in Toh 591, 592, and 593. Beginning with the next verse, the syntax changes to render the epithets in the feminine nominative singular. From that point the text is rendered in verse in the Sanskrit as well as the Tibetan sources.

n.36Reading rgyal ba’i rdo rje ’phreng as rgyal ba rdo rje phreng.

n.37The term aparājitā (gzhan gyis mi thub), “invincible,” is frequently used in this text as an adjective describing Sitātapatrā and her spell. Aparājitā is also the proper name of a protective deity, which is how the term seems to be used here.

n.38Tib. rnam sgeg ma, which is the equivalent of the Sanskrit vilāsinī. All Sanskrit versions consulted, except KT728, read viśālī (“vast one”). KT728 reads vajradaṇḍī (em. vajradhaṇḍī), “one with a vajra staff,” which agrees with the Tibetan of Toh 593 (rdo rje mkhar bsnams).

n.39Tib. zhi ba’i lha rnams kyi mchod pa. Most of the Sanskrit versions consulted read, “Peaceful, she is worshiped by vaidehas” (śāntā vaidehapujitā CL1326, Dh33, UTM 441-01).

n.40Reading rigs as rig based on the attested Sanskrit vajravidyā (CL1326, Dh33, UTM 441-01).

n.41This translation follows the Sanskrit syntax vajrahastā vajravidyā kāñcanamālikā . The word vajra has been omitted because it does not have an equivalent in the Tibetan translation, and the Tibetan term rigs is interpreted as rig following the Sanskrit º vidyā .

n.42The Tibetan reads le brgan rtsi dang rin chen ma, which could be interpreted as two names/epithets, but it is clear from the Sanskrit that this should be read as the single compound kusumbharatnā .

n.43Tib. sngags kyi tshogs, which would translate the Sanskrit compound mantragaṇa. The Sanskrit sources and Toh 590 read mātṛgaṇa (ma mo’i tshogs), “host of mātṛs,” suggesting a variant in the Sanskrit manuscript used by the translators of Toh 591, or a misreading of it.

n.44This can be tentatively translated as “Oṁ, Sitātapatrā born from the uṣṇīṣa of all tathāgatas, hūṁ drūṃ hrīḥ ṣṭoṃ! Crusher, hūṁ drūṃ hrīḥ ṣṭoṃ! Paralyzer, hūṁ drūṃ hrīḥ ṣṭoṃ! Stupefier, hūṁ drūṃ hrīḥ ṣṭoṃ! Devourer, hūṁ drūṃ hrīḥ ṣṭoṃ! Binder, hūṁ drūṃ hrīḥ ṣṭoṃ! Consumer of the poisons of adversaries, hūṁ drūṃ hrīḥ ṣṭoṃ! Paralyzer of all evil and wicked ones, hūṁ drūṃ hrīḥ ṣṭoṃ! Destroyer of all yakṣas, rākṣasas, and grahas, hūṁ drūṃ hrīḥ ṣṭoṃ! Disrupter of all spells, hūṁ drūṃ hrīḥ ṣṭoṃ! Destroyer of the eighty-four thousand types of grahas, hūṁ drūṃ hrīḥ ṣṭoṃ! Appeaser of the twenty-eight lunar mansions, hūṁ drūṃ hrīḥ ṣṭoṃ! Destroyer of the eight great celestial bodies, hūṁ drūṃ hrīḥ ṣṭoṃ! Protect, protect me, and all sentient beings!”

n.45Reading rnam par bsgyings ma’i rdo rje dang as rnam par bsgyings ma dang rdo rje following the Sanskrit sources, and specifically and Dh33, which reads ºvijṛmbhamānikā || vajrā kanakaprabhā locanā º.

n.46This translation follows the syntax of the Sanskrit (śrībuddhalocanā tathā UTM 441-01; śrībuddhalocanīti ca CL1326). The Tibetan reads sangs rgyas dpal dang de bzhin spyan (“Buddhaśrī and likewise Locanā”).

n.47In Toh 590, these verses are included the verse passage that precedes the spell above.

n.48Emended based on the Sanskrit, D, F, and S: pra shas+ta.

n.49Emended based on the Sanskrit and variants reported in N and S. D reads pra sA d+ha na.

n.50This can be tentatively translated as “Oṁ, Sitātapatrā born from the uṣṇīṣa of all tathāgatas, praised by the host of ṛṣis, hūṁ drūṃ! Crusher, hūṁ drūṃ! Paralyzer, hūṁ drūṃ! Devourer of great spells, hūṁ drūṃ! Devourer of opposing spells, hūṁ drūṃ! Paralyzer of all evil ones, hūṁ drūṃ! Destroyer of all yakṣas, rākṣasas, and grahas, hūṁ drūṃ! Destroyer of the eighty-four thousand types of grahas, hūṁ drūṃ! Appeaser of the twenty-eight lunar mansions, hūṁ drūṃ! Destroyer of the eight great celestial bodies, hūṁ drūṃ! Hūṁ hūṁ phaṭ phaṭ, protect, protect me, svāhā!”

n.51This translation follows the syntax of the Sanskrit sources, in which this and each of the subsequent phrases are declined in the ablative case.

n.52Tib. las kyi ’jigs pa, which could also be interpreted as “danger from karma.” This line is omitted in F, K, Y, and S, as well as all Sanskrit sources consulted.

n.53Here the initial supplication, “Grant auspiciousness...” has been inserted for the sake of clarity in the English translation.

n.54Tib. gshin rje’i gdon. Among the Sanskrit sources consulted, CL1326, RASH 77, and UTM 441-01 read jāmakīgraha. It is possible that similarities in the sound of ja- and ya- in Indic vernaculars resulted in the reading of jāmakīº instead of yāmakīº, but the term yama does appear elsewhere in the Sanskrit sources. Whatever the case may be, the Tibetan translators either read yāmakīgraha or interpreted jāmakīgraha as such.

n.55The Sanskrit versions confirm that the following sentences are in the first person singular present indicative voice. Additionally, the Tibetan phrase phur bus gdab bo suggests that the pinning is done by a kīla (“dagger”), but the Sanskrit texts suggest that the Tibetan phrase is a translation of kīlayāmi (√kīl), meaning “to pin down.” The Sanskrit versions consulted, as well as Toh 590, read kilayāmi vajrena (rdo rje phur bus gdab bo), “pin down with a vajra.”

n.56The Sanskrit sources as well as Toh 590, 592, and 593 read Mahāpaśupati (phyugs bdag chen po) instead of Maheśvara (dbang phyug chen po).

n.57Following Nārāyaṇa (Viṣṇu), the ambiguous phrase nam mkha’ lding yang dag pa dang lhan cig pa has been interpreted to be a specific reference to the garuḍa who serves as Viṣṇu’s mount. An equivalent to yang dag pa dang lhan cig pa does not appear in any of the Sanskrit sources consulted.

n.58The precise identity of these figures is uncertain, and this translation is conjectural. In his Nāma­mantrārthāvalokinī, a commentary on the Mañjuśrī­nāmasaṅgīti, Vilāsavajra refers to three brothers named Jayakara, Madhukara, and Sarvārthasiddhikara (Tribe 2016, p. 226: jayakaramadhukara­sarvārthasiddhikarās trayo bhrātaras). In all available sources the first two names are consistently given as Jayakara and Madhukara, but the third name varies in the Sanskrit witnesses consulted. The oldest, KT728, reads ºsarvārtha(sā)dhanaº; CL1326, Dh33, and UTM 441-01 have ºsiddhikara­sarvārtha­sādhanaº; and RASH 77 has ºsarvārthāsiddhi­sādhakaº. It is possible to interpret the Sanskrit reported in these sources to read “[the spells] of Jayakara, Madhukara, and Siddhikara used to accomplish all aims (sarvārtha­sādhana).”

n.59Emended based on the Sanskrit sources as well as Toh 590, wherein this Sanskrit line is translated into Tibetan. D, F, and S read, in transliteration, asitānalārkaprabhāspuṭavikā­sitātapatre.

n.60This can be tentatively translated as “Oṁ, White Umbrella ( sitātapatrā ) opened broadly and shining with the white fire of the sun! Oṁ, blaze, blaze! Devour, devour! Kill, kill! Break, break! Destroy, destroy! Cut, cut! Cleave, cleave! Bind, bind! Hūṁ hūṁ phaṭ phaṭ svāhā!”

n.61This line is absent in Toh 590, 592, and 593, as well as all Sanskrit sources consulted.

n.62There is wide variation in the rendering of this compound across Sanskrit and Tibetan sources. The reading given here is reported in D; F reads (in transliteration), sarvakaṭaṇebhyaḥ; H, K, Y, N, and S read (in transliteration), sarvakatadanebhyaḥ. As none of these variants offers a clear alternative, the reading in D has been preserved. The sequence in the Sanskrit sources do not align closely enough with the Tibetan sources to identify an obvious equivalent, but the most plausible reading of what appears to be a corrupt line is sarvakaṭapūtanebhyaḥ, a phrase already used above.

n.63This line also appears corrupt and is preserved here as it appears in D and S. The line is absent in F, and there is no clear equivalent in the Sanskrit sources.

n.64There are several points of possible corruption in this passage, but it can be tentatively translated as “He he phaṭ! Ho ho phaṭ! To the unfailing one, phaṭ! To the unobstructed one, phaṭ! To the boon granter, phaṭ! To the boon bestower, phaṭ! To the averter, phaṭ! To the disperser of asuras, phaṭ! To the disperser of the enemy, phaṭ! To all devas, phaṭ! To all nāgas, phaṭ! To all yakṣas, phaṭ! To all rākṣasas, phaṭ! To all bhūtas, phaṭ! To all pretas, phaṭ! To all piśācas, phaṭ! To all kumbhāṇḍas, phaṭ! To all pūtanas, phaṭ! To all kaṭapūtanas, phaṭ! To all skandas, phaṭ! To all unmādas, phaṭ! To all chāyās, phaṭ! To all apasmāras, phaṭ! To all ostārakas, phaṭ! To all ḍākinīs, phaṭ! To all revatīs, phaṭ! To all yāmakas, phaṭ! To all śakunis, phaṭ! To all hosts of mātṛs, phaṭ! To all kambukāminīs, phaṭ! To all ālambakas, phaṭ! To all kaṭaṭaṇas [?], phaṭ! To all kaṇḍas [?], phaṭ! To all gandharvas, phaṭ! To all asuras, phaṭ! To all kinnaras, phaṭ! To all garuḍas, phaṭ! To all mahoragas, phaṭ! To all those difficult to violate, phaṭ! To all unsightly spirits, phaṭ! To all fevers, phaṭ! To all fears, phaṭ! To all calamities, phaṭ! To all infectious diseases, phaṭ! To all kākhordas and kṛtyā rites, phaṭ! To all kiraṇas and vetālas (vaitaḍa), phaṭ! To all ciccas (cicha), preṣakas, and spirits of vomiting, phaṭ! To all indigestion spirits, phaṭ! To all non-Buddhists, phaṭ! To all ascetics, phaṭ! To downfalls, phaṭ! To all vidyādharas, phaṭ! To Jayakara, Madhukara, and Siddhikara, the accomplishers of all aims, phaṭ! To all masters of spells, phaṭ! To the four bhaginīs, phaṭ! To the Vajrakaumārī, the scion of her family and master of spells! To the masters of spells, phaṭ! To all great averters, phaṭ! To Varjaśṛṅkhala, king of averting, phaṭ! To Mahākāla, who is honored by the host of mātṛs, phaṭ! To Brahmaṇī, phaṭ! To Vaiṣṇavī, phaṭ! To Māheśvarī, phaṭ! To Raudrī, phaṭ! To Mahākālī, phaṭ! To Cāmuṇḍī, phaṭ! To Kaumārī, phaṭ! To Vārāhī, phaṭ! To Indrā, phaṭ! To Agni, phaṭ! To Yamā, phaṭ! To Nairṛti, phaṭ! To Varuṇā, phaṭ! To Mārutī, phaṭ! To Soma, phaṭ! To Aiśānī, phaṭ! To Kāladaṇḍī, phaṭ! To Kālarātri, phaṭ! To Yamadaṇḍī, phaṭ! To Yamarātri, phaṭ! To Kapāli, phaṭ! To those who prefer to dwell in charnel grounds, phaṭ! Oṁ ṣṭoṃ, bind, bind! Protect, protect me, svāhā!”

n.65Tib. re lde ’don pa. Associating this Tibetan term with the attested Sanskrit term kaṭakamālinī is tentative. The Tibetan term re lde (“felt”) is a known equivalent of kaṭa, while ’don pa is used translate mālinī in this text, as attested in the next term in the list, kaṇṭakamālinī (tsher ma ’don pa).

n.66Tib. mi bzad pa; Skt. viṣama. While viṣama can be interpreted as “unbearable,” as the Tibetan translators did, in the context of the duration or recurrence of fever it means “irregular.”

n.67This translation follows the attested Sanskrit term ardhāvabhedaka. The Tibetan term, gzhogs phyed na ba, could also be interpreted as a translation of pakṣavadha, referring to hemiplegia.

n.68Tib. chu so, which translates the Sanskrit vasti. The same term is translated with mdoms in Toh 590.

n.69The “major” appendages would include the head, arms, legs, etc. The “minor” appendages include the nose, ears, fingers, and toes.

n.70Tib. bas bldags. The Tibetan term, for which there is no Sanskrit equivalent in the sources consulted, means “cow licked” (Skt. golīḍha?) and refers to a type of skin irritation with a sensation similar to that of being licked by a cow.

n.71It is clear from the Sanskrit sources that these verbs are in the first person.

n.72Toh 590 and all Sanskrit sources consulted read khakhame khakhame, the meaning of which is ambiguous.

n.73Viṣada (“poisoner”) is attested in the majority of sources but should perhaps be emended to viśada (“brilliant”). The confusion of sibilants is a consistent feature of Sanskrit manuscripts, thus the reading viśada is perhaps preferable. However, none of the Sanskrit and Tibetan sources consulted attest to viśada.

n.74All Sanskrit sources consulted read vajrapāśe.

n.75This can be tentatively translated as “It is like this: Oṁ, O fire, fire! O you who are equal to the sky, equal to the sky! O hero, hero! O gentle one, gentle one! O peaceful one, peaceful one! O tamed one, tamed one! O poisoner, poisoner! O vengeful one, vengeful one! O vajra-holding goddess! Bind, O you who bind! Vajrapāṇi, phaṭ! Oṁ hūṁ drūṃ hrīḥ ṣṭoṃ phaṭ svāhā! Oṁ, Vajrapāṇi! Bind all evil obstructors and obstacle makers with your vajra noose, hūṁ hūṁ phaṭ phaṭ svāhā!”

n.76This line is absent in Toh 590, 592, and 593 and all Sanskrit sources consulted.

n.77The syntax of the Tibetan translation presents some interpretive challenges, thus this translation is tentative. The term mthun par appears to translate satatasamitam as attested in the Sanskrit sources, and so it has been translated here as “perpetually.”

n.78Following the Sanskrit sources in reading maunin for the Tibetan thub pa.

n.79Tib. nad. The Sanskrit sources all read ºmāra, which could be understood as “obstacles,” or perhaps “fatalities.”

n.80“City” (nagara; grong khyer) is repeated in all sources consulted. The difference appears to be that the banner can be planted either at the gateway to a city or generally in the city.

n.81All Sanskrit version consulted, as well as Toh 590, 592, and 593, read dara dara, “Break, break!”

n.82This can be tentatively translated as “It is like this: Oṁ ṣṭom, bind, bind! Protect me, protect, svāhā! Oṁ ṣṭoṃ, bind, bind, vajra! Protect me, protect, Vajrapāṇi, hūṁ phaṭ svāhā! Oṁ, O fire of the uṣṇīṣa of all tathāgatas who gaze down! Brilliant splendor reduces to dust! Oṁ, blaze, blaze! Burn, burn! Hold, hold! Destroy, destroy! Cut, cut! Cleave, cleave! Hūṁ hūṁ phaṭ phaṭ svāhā! Oṁ, the uṣṇīṣa of all tathāgatas, hūṁ phaṭ svāhā! It is like this: Oṁ, O fire, fire! O one equal to the sky, one equal to the sky! O vengeful one, vengeful one! O gentle one, gentle one! O you who are empowered by the blessing of all buddhas, Sitātapatrā born from the uṣṇīṣa of all tathāgatas, hūṁ phaṭ! Hūṁ mama hūṁ ni hūṁ svāhā!”

n.83This enigmatic statement perhaps indicates that one should recite this formula while meditating on or contemplating the Buddha, or perhaps in the presence of an image of a buddha.