Notes

n.1The Sanskrit title of this work on the title page of this translation reflects the emended reading from the Tohoku catalog.

n.2Mārīcī is referred to with this title in the opening line of the Sanskrit witness corresponding to Toh 565, but the reference is lost in the variant reading that survives in the Tibetan.

n.3Tarthang Tulku 1982, pp. 198–99.

n.4There are counterparts to many of the forms of Mārīcī described in these texts among the thirty-seven sādhanas for the goddess Mārīcī preserved in the Tengyur that were translated in the twelfth century, and it might be reasonable to assume that Toh 565 and Toh 566 were translated during the same period.

n.5Lewis R. Lancaster, K 1156, The Korean Buddhist Canon, accessed September 5, 2023. For etexts of this version, see Fo shuo da molizhi pusa jing 佛説大摩里支菩薩經 (Mārīcīdhāraṇī), Taishō 1257 (CBETA; SAT).

n.6This matter might be settled by a close comparison of the Chinese and Tibetan translations of these works, which we have unfortunately not been able to complete for this publication.

n.7Multiple Sanskrit witnesses are preserved among the Sanskrit manuscript collections at the University of Tokyo, Cambridge University, and the Kathmandu National Archive.

n.8The point at which the Sanskrit cuts off corresponds to Degé Kangyur vol. 90, folio 178.a7. See n.­101. The lacuna is followed immediately by the fragment of the scribal colophon.

n.9NE 1480/9 reads yāddṛstāpustakaṃ dṛstāstādṛsatvā [sic for yathādṛṣṭaṃ pustakaṃ tathā likhitaṃ?] mayā | yadi suddham asuddham vā mama dokho [sic for doṣo] na dīyate | śrī 3 śrī 3 vajrāccāryyaravṛndrabhadrare [sic for ravīndra­bhadreṇa?] lekhyāko yo postakaḥ śubham. This fragment of the scribal colophon might be tentatively translated, “I have copied this down exactly as it appeared in the book. If it is pure or corrupted, the fault should not be placed on me. This was copied down by Śrī (3) srī (3) Vajrācārya Ravīndrabhadra. May there be prosperity and good fortune.” The identification of Ravīndrabhadra as the scribe for this text is traced to the NGMCP catalog card. Given the high honorific prefixes that precede this name in the scribal colophon, it is possible that the text was transcribed for (and not by) the Vajrācārya Ravīndrabhadra.

n.10Mārīcīkalpa, NGMCP E 1480/9, Nepal National Archive, Kathmandu.

n.11Mārīcī­kalpa­tantra, IASWR MBB II 112.

n.12Following D and S lha mo ’od zer can la phyag ’tshal lo. NE 1480/9 reads namo vaj[ra]sattvāya, preserving an alternative reading that would be translated “Homage to Vajrasattva.”

n.13D reads tsan+dana dkar pos yan lag byugs la me tog gi phreng bas brgyan te, S reads can dana dkar pos yan lag byugs la me tog gi ’phreng bas brgyan te, and NE 1480/9 reads sitacandanaliptāṅgāḥ puṣpamālavibhuṣitāḥ. The Sanskrit witness makes it clear that the vases, and not the person performing the rite, are anointed with sandalwood. This is also supported by a similar line in The King of Ritual Manuals from the Tantra of Māyā Mārīcī’s Arising (Toh 565, 1.­89) that includes the step of anointing the vases with white sandalwood.

n.14D and S read de nas rnam par rgyal ba’i bum pa la brgad brgya sngags te, and NE 1480/9 reads vijayakalaśa[ṃ] aṣṭasatābhimaṇtritaṃ. It is not entirely clear, but this most likely refers to using the six-syllable Mārīcī mantra.

n.15D and S read gos zung gis gdung dkris la, and NE 1480/9 reads vastrayugena mukhaveṣṭitaḥ. This translation is tentative and interprets this line as a reference to the practice of a disciple being blindfolded prior to entering a consecration maṇḍala. It is not entirely clear that this is the case here, and this line might simply refer to the disciple’s head being wrapped with two pieces of cloth.

n.16Following NE 1480/9 oṃ arkamasi svāheti pūrvasyā diśi nyase[t]. D reads oM ark+ka ma si swA hA zhes bya ba ni shar phyogs su dgod do, and S reads oM ark+ka ma si swA hA/ zhes bya ba ni shar phyogs su dgod do. In this text, the components of Mārīcī’s dhāraṇī are explicitly identified as the vidyā goddesses surrounding Mārīcī in her maṇḍala. The text itself only refers to the vidyās for these goddesses being installed in the cardinal and ordinal directions, but it is clear that they are identified as goddesses with distinct iconographic features. This translation follows the syntax in NE 1480/9, where the vidyā stands alone at the beginning of each of these short sections describing the vidyā goddesses in Mārīcī’s maṇḍala.

n.17This is the only instance in this series of instructions that explicitly mentions “depicting” these goddesses.

n.18D and S read ge’u ri gser gyi mdog can gzhon nu ma’i rnam pa ’chang ba, Y, K, J, and C read dge’u ri gser gyi mdog can gzhon nu ma’i rnam pa ’chang, and NE 1480/9 reads gaurīdevī kaṇakavārṇā kumārākārarūpadhāriṇī. The Sanskrit text tells us that ge’u ri / dge’u ri is a Tibetan transliteration of gaurī. Although the Sanskrit indicates that this is the goddess Gaurī (gaurī devī), this translation follows the Tibetan, which treats the term as a modifier for gser (“yellow”) to give us the color “pale yellow.” There is no equivalent of the term devī here in the Tibetan text.

n.19Following NE 1480/9 aṣṭasu [sic for aṣṭeṣu] sṭhāneṣu sā [sic for tāṃ?] nyased devīm anupūrvataḥ. D and S read go rims bzhin du byang gi phyogs su gnas brgyad du lha mo dgod par bya’o. The readings in all witnesses are a bit corrupted. The alternate reading in the Tibetan witnesses is “install a goddess in the eight directions in the north in the proper order.” NE 1480/9 also reads “in the eight directions” (aṣṭasu sthāneṣu), but in both cases this line is followed only by a list of four goddesses who occupy the ordinal directions.

n.20Following S oM u da ya ma si swA hA. D reads oM ut+ta yama si swA hA, and NE 1480/9 reads oṃ uktayamasi svāhā.

n.21Following S oM ba na ma si swA hA and NE 1480/9 oṃ vanamasi svāhā. D reads oM ba nA ma si swA hA.

n.22Following NE 1480/9 oṃ cīvaramasi svāhā. D and S read oM tsi ba ra ma si swA hA.

n.23Following D sa ma sa mU ra tI. S reads sarb+ba mU rti, and NE 1480/9 omits. Although this transliteration follows the reading in D‍—the meaning of which is not clear and which is likely corrupt‍—this same phrase has been rendered as sarvasattvamudrati (“she who delights all beings”) in the Sarnath Sanskrit edition of The Mārīcī Dhāraṇī (Toh 564, 1.­8), and it is possible that this is the preferred reading here as well. The reading in S might tentatively be translated “she who is an embodiment of all.”

n.24D reads ba dA kra ma si, S reads pa dA kra ma si, and NE 1480/9 omits.

n.25D and S read snam bu, and NE 1480/9 omits. This translation is tentative. The Tibetan term snam bu likely translates the Sanskrit terms paṭa or paṭṭaka/paṭṭika here. This implies that the material on which these mantras are written is likely a small cloth canvas, or perhaps a small wooden, metal, or stone tablet.

n.26This renders the Tibetan thong zhing yid ches pa (Skt. dṛṣṭapratyaya). This term seems to occur most frequently in reference to enthralling a king, as it does here.

n.27Following NE 1480/9 dūrvāghṛtāktena. D reads dUr ba la smyugs pas bya’o, and H, J, K, C, and S read dur ba la bsnyugs pas bya’o.

n.28D reads gos thod du bcings pa yang rnyed do, and NE 1480/9 reads paṭabandhañca labhate. “Receiving a turban” should likely be understood as an idiomatic phrase that refers to receiving a rank at a king’s court.

n.29Following S aM tardA na ma si, which is supported by NE 1480/9 anta[r]dhānamasi. D reads An+d+hardA na ma si.

n.30D reads byi la nag mo’i dri ma dang / byi ma la dang mig sman la, S reads byi la nag mo’i dri ma dang / /byi la dang / byi mi la dang / mig sman la, and NE 1480/9 reads kṛṣṇabidālāmalaṃ. This translation is tentative. The only phrase in the Tibetan that appears in NE 1480/9 is byi la nag mo’i dri ma (kṛṣṇabidālāmalaṃ). It is possible that the seemingly redundant phrase byi ma la is actually a gloss for the phrase byi la nag mo’i dri ma, with the Tibetan dri ma rendered in the transliterated Sanskrit mala. On its own, the Sanskrit can be read either as “excrement (mala) of a female black cat (kṛṣṇabiḍālā)” or “a rheum (mala) of black collyrium (kṛṣṇabiḍālā).” The translation here reflects an attempt to capture both meanings, which is likely the intent of the Tibetan rendering.

n.31This is the same list of qualities that the Buddha Śākyamuni assigns to Mārīcī when he teaches her dhāraṇī in The Mārīcī Dhāraṇī (Toh 564, 1.­3). This set of qualities appears several times in this text, which is clearly conversant with The Mārīcī Dhāraṇī.

n.32Following S ga Na. D reads ga na, and NE 1480/9 reads gana.

n.33Following D and S triphatri. Y and K read tipatri, and NE 1480/9 reads dipyādipya [sic for dīvyadīpya?]. Although this transliteration follows the reading in D and S, it is likely that the amended reading dīvyadīpya in NE 1480/9 is the preferred one. We have not adopted it here, however, because it is somewhat different from the reading in the Tibetan witnesses.

n.34D and S omit this, and NE 1480/9 reads akṣasūtrapaśaṃ, adding a noose to the list (paśa [sic for pāśa]) for a total of five hand implements.

n.35D and S read me lha bzang po, and NE 1480/9 reads pāvakāgni. Note that the Sanskrit refers to this peaceful form of Agni by the name of a specific form of the deity, Pāvaka Agni. This peaceful form is also the subject of the Agni homa practice in The King of Ritual Manuals from the Tantra of Māyā Mārīcī’s Arising (Toh 565, 1.­107), where the Tibetan refers to the form as mang sa be da’o zhes bya ba’i me lha.

n.36D reads de nas phyis bkod pa, S reads de nas phyi bkod par, and NE 1480/9 reads tataḥ paścān nyaset. This translation is tentative. The term “deity” does not appear in the Tibetan or Sanskrit witnesses, but it can be infered based on the context. It is also not clear just where the deity should be installed, but it is presumably in the fire itself.

n.37D and S read bdag nyid gal te dngos grub ’dod na, and NE 1480/9 reads yadi cet[?] siddhim ātmanaḥ. This translation is tentative. It is not entirely clear to whom the term bdag nyid (ātmanaḥ) refers here. The reading in NE 1480/9 might translate, “if one possesses siddhi.”

n.38Following S gal te sa gzhi bdag por ’dod pa ’am yin na’o and Y and K gal te sa gzhi bdag por ’dod pa’am yin na’o, which are supported by NE 1480/9 yadi cchet bhupatitvaṃ vā. D reads gal te sa gzhi bdag por ’dod pa ma yin na’o.

n.39D reads sna tshogs rdo rjes bkang bas mnan pa bri bar byas na, S reads sna tshogs rdo rje rkang pas, and NE 1480/9 reads viśvavajrapādātrānya likhet. This translation is tentative.

n.40D and S read ’od zer can ’byung ba’i dkyil ’khor du dbang bskur ba bla ma, and NE 1480/9 reads mārīcyudayamaṇḍalābhiṣiktasya. The opening line of NE 1480/9 preserves a different spelling that reads mārīcyutbhave[sic udbhāve] tantre (Tib. ’od zer can ’byung ba’i rgyud), and the Tibetan transliteration of the Sanskrit title reads mārīcyai jāta. Unfortunately, the single Sanskrit manuscript witness at our disposal is incomplete and is missing its colophon title page. In light of these and other variations in the way it is spelled in our witnesses, this phrase is only tentatively presented here as the proper title of a work. This seems reasonable, however, because the current text begins with a set of instructions for Mārīcī’s consecration maṇḍala referred to as “the maṇḍala of Mārīcī’s arising,” and both the current text and The King of Ritual Manuals from the Tantra of Māyā Mārīcī’s Arising (Toh 565) claim to derive from a source text called The Tantra of Māyā Mārīcī’s Arising.

n.41Following S ’thug pa’i skra brdzes pa’i gzugs yid du ’ong ba. D reads thu ba brdzes pa’i gzugs yid du ’ong ba, and NE 1480/9 reads mānavārūrūpāṃ[?] manoramāṃ.

n.42Following NE 1480/9 mā me paśyantu satvā, meaning “may beings not see me.” The reading in the Tibetan witnesses appears to be a corruption of this phrase‍—D and S read ma me pA shAM tu sta swA hA, and Y and K read ma me pA shAM stu swA hA.

n.43Following the reading in NE 1480/9 (oṃ māṃ), which includes Mārīcī’s heart mantra. The Tibetan witnesses only provide the seed syllable oṃ.

n.44D reads ’dod pa bzhi sbyin pa bskyed pa ’di, C and S read ’dod pa bzhin sbyin pa bskyed pa ’di, and NE 1480/9 reads catuṣkāmapradaṃ. The precise identity of these four is not clear.

n.45D and S read ’od zer can ’byung ba zhes bya ba’i ’od zer can ’byung ba’i rgyud ’di gsungs pa ni, and NE 1480/9 reads mārīcyutpadayam nāma mārīcyudbhave tantre nīgaditaṃ. These witnesses do not include any term that would indicate that this is an excerpt from a “section” of the root text, but it seems reasonable to assume that this is the case.

n.46D and S read pa tA kra ma si, and NE 1480/9 reads padākramasi. This transliteration follows the reading in the Sanskrit witness.

n.47Following S ba na ma si and NE 1480/9 vanamasi. D reads ba nA ya ma si.

n.48Following D and S bi pa ri ma si. NE 1480/9 reads cīramasi[?]. Though this transliteration follows the reading in D and S, this is the only time that the name of this vidyā goddess appears in any of the works devoted to Mārīcī in the Kangyur witnesses. If we follow the (somewhat) standard sequence of these vidyās that is established in The Mārīcī Dhāraṇī (Toh 564), The King of Ritual Manuals from the Tantra of Māyā Mārīcī’s Arising (Toh 565), and elsewhere in this text, it is likely that this vidyā should read cīvaramasi. NE 1480/9 is unfortunately not helpful here.

n.49D reads g.yon pa’i phyogs kyi zhal ni, and S reads g.yon pa phyogs kyi zhal ni. NE 1480/9 reads vāmamukhaṃ varāhaṃ, noting that the left face of this form of Mārīcī is that of a boar.

n.50Following H and S mchod rten kyi rkad ba la. D reads mchod rten gyi ske ba la, and NE 1480/9 omits.

n.51D and S read g.yas brkyang ba’i zhabs kyis bzhugs pa. NE 1480/9 reads pratyālidhasthāsthitaṃ, providing a variant here that has Mārīcī standing with the left leg extended (g.yon brgyang, pratyālīḍha), not the right.

n.52Following NE 1480/9 rāhu. D reads rA hu la, and Y, K, and S read ra hu la.

n.53S reads khab dang skud pa’i shing rta’i mchog mas, and D reads khab dang skud par byed pa’i shing rta’i mchog las.

n.54Following S va rA ha mu khi and NE 1480/9 varāhamukhi. D reads bA rA hA mu khi.

n.55The reading sarvaduṣṭapraduṣṭānāṃ is emended following precedent for how this phrase is rendered in this text and Toh 564 and 565, which reflects the proper Sanskrit declension. D and S read sarba duSh+Ta pra duSh+tA cak+Shur mu khaM. NE 1480/9 reads sarvaduṣṭānāṃ mukha, reflecting the proper declension but omitting some material preserved in the Tibetan witnesses.

n.56Following S ba rA ha mu khi. D reads bA rA hA mu khi, and NE 1480/9 reads valāhamukhi [sic for varāhamukhi].

n.57Following S ba rA ha mu khi. D reads bA rA hA mu khi, and NE 1480/9 reads varāhamuṣi [sic for varāhamukhi].

n.58D reads sarba duShTAM me ba sha ma nA ya swA hA, S reads sarb+ba du Sh+ta n me ba sha ma nA ya swA hA, and NE 1480/9 reads sarvasattvān meḥ [sic for me] vaśam ānaya svāhā. This transliteration emends the reading me ba sha ma nA ya, based on the reading in NE 1480/9, to me vaśam ānaya, which would be translated “bring under my control.” The variant “all wicked beings” (Tib. sar ba duShTAM) in the Tibetan witnesses is retained here instead of the reading “all beings” (Skt. sarvasattvān) that appears in NE 1480/9.

n.59It is not clear in the text whether this refers to performing the rite for seven consecutive days or simply performing the rite seven times. It is also not clear whether the entire rite is performed seven times or just the mantra recitation. This translation assumes that it is the mantra recitation.

n.60The terms “siddhis” and “ritual powers” are not in the Tibetan text but are added here for clarity.

n.61D and S read steng g.yogs yod pa’i phyogs su, and NE 1480/9 reads pracchanne guptadeśe. The Tibetan term steng g.yogs translates the Sanskrit pracchanna.

n.62D and S read chos ’byung du, and NE 1480/9 reads bhage. The Sanskrit text indicates that this “source of phenomena” might be a code word for the female genitalia. The target of this enthralling rite is clearly feminine in the Sanskrit, but this translation preserves the gender ambiguity in the Tibetan.

n.63D reads sin d+hu ra’i rdul gyis mdzes pa’i snod du bstan pa’i sman chen pos, S reads sin+d+hu ra’o rdul gyis mdzas pa’i snod du bstan pa’i sman chen pos, Y, K, and C read sin du ra’o rdul gyis mdzas pa’i snod du bstan pa’i sman chen pos, and NE 1480/9 reads pratikṛtvā sindureṇa rajitā purvoktai mahoṣadhai. The Sanskrit clarifies a problem in all the Tibetan translations here, which render the phrase sngon du bstan pa (purvokta) as snod du bstan pa.

n.64Following NE 1480/9 ṣadagniveṣṭitai, which suggests that the Tibetan me dug gis bskor ba (in D and S) should be emended to me drug.

n.65D and S read g.yas brkyang ba’i zhabs kyis bzhugs pa, and NE 1480/9 reads pratyālīḍhasthānasthā.

n.66D and S read dge ba’i skar ma bzang ba rgyal la, Y reads dge ba’i sgar ma bzang ba rgyal la, and NE 1480/9 reads śubhanakṣatre puṣye. Although the term for “day” is missing in both the Tibetan and the Sanskrit, this phrase is read as a variant of the phrase dge ba’i nyin skar ma rgyal la, which appears below in this same group of instructions.

n.67Following NE 1480/9 aṅkasamāropya. D and S read lus kyis khur nas.

n.68Following the emended reading in NE 1480/9 caityālaṅkṛtamūrdhajāṃ (for caityānaṇkṛtamurddhajā). D and S read mchod rten gyi dpal la brgyan pa’i mgo skyes can, and C reads mchod rten gyis dbul la brgyan pa’i mgo skyes can.

n.69This translation is tentative and follows D zla ba dang / nyi ma’i rang bzhin gyi lha ma yin dang / phag pa’i shing rta dang nye bar sbyar ba and S zla ba dang nyi ma’i rang bzhin gyi lha ma yin dang / phag pa’i shing rta dang nye bar sbyar ba. NE 1480/9 reads śasibhālupamaṃ [sic for śasibālusamanvitaṃ?] caityasurarathaniyojitaṃ [sic for caityasurathaniyojitaṃ?], providing an alternate reading that, when it is emended following the corrections suggested above, would be translated, “She is accompanied by the sun and moon, and her chariot is furnished with a caitya.” It is also possible to leave the compound caityasurarathaniyojitaṃ as it stands, which would be translated, “a chariot of the gods that has a caitya.” The Sanskrit may also support a number of valid readings.

n.70Following D and S g.yon brkyang ba’i zhabs. NE 1480/9 reads āliḍha, which says that Mārīcī stands with her right leg forward, not her left.

n.71Following NE 1480/9 ha, which is consistent with the depiction of Rāhu and his whirlwind that has already been provided in the Tibetan and Sanskrit witnesses above. D and S read haṁ, Y reads hū, and K reads hūṁ.

n.72D and S read rāhula, and NE 1480/9 reads rāhu.

n.73Following D and S khab yod pa’i sa phyogs gcig yangs pa’i gnas su. NE 1480/9 reads pracchane bhupradeśe gṛhe (“in a house on a plot of land that is hidden”).

n.74Following NE 1480/9 kanakarajatatāmra athavā mṛnmayakalaśa [sic]. D and S read gser dngul dang / zangs ma’o/ /yang na ’dab chags rang bzhin gyi bum pa la bya’o. The Tibetan ’dab chags rang bzhin gyi bum pa for mṛṇmayakalaśa is obscure.

n.75Following NE 1480/9 vairocanamahāmudrāṃ samādhāya atmānaṃ. D reads rnam par snang mdzad phyag rgya rgya chen po’i gzung ba la bsam par bya’o, and Y, K, C, H, and S read rnam par snang mdzad phyag rgya chen po’i gzung ba la bsam par bya’o.

n.76D reads de nas sngon du gsungs pa’i bum pa tsan+dana dkar pos byugs la me tog phreng ba ’dzin pa la rig pa ’dzin pas brgya phrag brgyad bzlas te/ arg+ha byin la bum pa de gzhag par bya, S reads de nas sngon du gsungs pa’i bum pa tsan dana dkar pos byugs pa me tog gi ’phreng ba ’dzin pas brgya phrag brgyad bzlas te arga byin la bum pa de gzhag par bya’o, and NE 1480/9 reads evaṃ purvoktena kalaśena sitacandrena liptena śṛṇgāmālāvibhuṣitena aṣṭaśatajaptena vidyādhareṇa.

n.77D and S read bcom ldan ’das ma bdag nyid la dbang bskur zhing, and NE 1480/9 reads bhagavatīm ātmānam abhiṣiñcayet. Both the Tibetan and Sanskrit indicate that Mārīcī is the object of the verb here, which means that the Sanskrit term abhisiñcayet (Tib. dbang bskur) likely refers not to the practitioner receiving a consecration or empowerment but to the act of sprinkling the statue of Mārīcī with the offering water.

n.78Following NE 1480/9 tato mārīcyā sāhaṃkāragarvitena japabhāvanā ca kāryā. D and S read de nas ’od zer can bdag la bsnyen pas bzlas pa dang bsgom pa bya.

n.79D and S read tsan+dana dmar po dang / gi’u wang gis, and NE 1480/9 reads raktacandanakuṁkama[sic]rocanayā (“with red sandalwood, saffron, and bovine bezoar”).

n.80Following NE 1480/9 gugguludhupitaṁca deyaṃ. D reads gur gum gyi bdug spos sbyin, and S reads gur kum gyi bdug spos sbyin. The reading in D and S prescribes making the incense offering with saffron.

n.81Following NE 1480/9 māṃ. The Tibetan witnesses omit this.

n.82D reads skroṃ, Y, K, and S read strom, C reads sgrom, and NE 1480/9 reads strīṃ strīṃ strīṃ.

n.83Following NE 1480/9 aśvagharma. D and S read rta’i rdul.

n.84D reads byi la nag po’i dri ma dang / khyi nag po’i mig dang / rta nag po’i mig dang / khwa nag po’i mig dang / phag nag po’i rna ba g.yon pa dang / khrag dang, S reads byi ma nag po’i dri ma dang khyi nag po’i mig dang / rta nag po’i mig dang / khwa nag po’i mig dang / phag nag po’i rna ba g.yon pa dang / khrag dang, and NE 1480/9 reads kṛṣṇamājārāmālalecanaṅ[?]kṛṣṇākṣikṛṣṇavarāhavāmekarṇarudhireṇa marditaṃ. The verb “ground up” (Skt. marditaṃ) is adopted from NE 1480/9 and seems to be omitted in the Tibetan witnesses.

n.85D and S read rdo rje ’od mas mnyes la, and NE 1480/9 reads vajrīkṣīreṇa [sic?] veṣṭitā. The Sanskrit is corrupt here, but it is still clear that the Tibetan should read rdo rje’i ’o ma.

n.86Following NE 1480/9 prāsādikān. D reads bong ring ba, Y reads bong ba ring ba, K reads bong ge ring ba, and S reads bongs ring ba.

n.87Following D and S bi pu la ba re na ge. NE 1480/9 reads vipule pravare nāge, which is most likely the preferred reading.

n.88Following D and S nA ga ku la ya va ti swa ni. NE 1480/9 reads nāgakulavidhvaṁsani, which is most likely the preferred reading.

n.89Following Y, K, and S sarba duSh+Ta, which is supported by NE 1480/9 sarvaduṣṭa. D reads sarba duSh+TAM.

n.90Following S sarba duShTa, which is supported by NE 1480/9 sarvaduṣṭa. D reads sarba duSh+TAM.

n.91Following NE 1480/9 sarvaduṣṭa. D reads sarba duSh+TAM, and S reads sarba duShTaM.

n.92Following H phye ma dang ’o ma’i klu. D and S read phye ba dang ’o ma’i klu, and NE 1480/9 reads śaktapāyaśena.

n.93D and S read rnam pa sna tshogs pa ji ltar rnyed pa’i chu, and NE 1480/9 reads nānāvidharasā yathāsaṃvidyamānā. The Sanskrit witness indicates that the Tibetan term chu translates rasa, which can mean “water” but often also refers to a nectar, liquor, drink, or juice and has many more potential interpretations. The term nānāvidharasa might refer to substances that contain the various “flavors,” even though this is not the way that the Tibetan translator has interpreted the passage.

n.94D and S read lus rnams la sha dkar gyi mdze ’byung bar ’gyur, and NE 1480/9 reads sarvaśarītvāṇi [sic for śarīrāṇi?] citrā [sic for citrāṇi?] bhavanti. The Sanskrit witness uses the term for spots (citrāṇi) as a synonym for leprosy (kuṣṭha), while the Tibetan includes a more specific reference to the condition vitiligo, rendered here as a variety of “white leprosy” (dkar gyi mdze).

n.95D and S read se’u ’bru’i yal ga la lan nyi shu rtsa gcig bzlas pa bya zhing brgya rtsa brgyad kyang bzlas pa byas la/ klu’i gzugs brnyan gyi gzugs rdzogs par bya’o, and NE 1480/9 reads dāḍimalatayā[ṃ] ekaviṁśati vā rambari [sic?] japtayā[ṃ] aṣtottarasataparijaptayā[ṃ] vā mantrena nāgapratitūpakam aya[ṃ] mārjāyitavyaṃ. This translation is tentative and adopts the participle form mārjāyitavyaṃ from the Sanskrit in place of the Tibetan rdzogs par bya.

n.96D and S read klu de’i gzugs ’gul zhing sbrul mgo’i gar yang byed do. This translation is informed by NE 1480/9 sa nagapratirūpaka[ḥ] prakrāmati phaṇaṃ ca karoti, where we see that the Tibetan phrase sbrul mgo’i gar yang byed (“performs a snake-hood dance”) translates the Sanskrit phrase phaṇam ca karoti, which simply means “opens its hood.”

n.97D reads lhag ma’i klu’i rus pa’i gong po rnams yongs su bskam par bya’o, and NE 1480/9 reads tata sarvanāgabhavanāni pradipyante nāgā | asthisakalān na vaśiṣṭāḥ parimucyante.

n.98Following S ma lus pa’i yul dang / phyogs kyi nam mkha’, which is supported by NE 1480/9 saka[la]digdiśāvyomu [sic for vyomā?]. D reads ma lus pa’i yum dang / phyogs kyi nam kha’.

n.99Following D and S zhabs gnyis rol pas gnas shing. NE 1480/9 reads krodhālīḍhapadam ubhayasṭhitā, according to which Mārīcī should be visualized in a wrathful (krodha) and not a playful (kriḍa) posture.

n.100D and S read mi mnyam pa’i spyan klu’i dbang po ser po’i g.yon pa’i phyag tu, and NE 1480/9 reads kaṭakṣe[na?] pīḍayan[ti?] nagendrāṃ | vāme kare. This translation is informed by the reading in NE 1480/9, where it is clear that the nāga lord Pīta’s name has erroneously made its way into the Tibetan witnesses. The Sanskrit also makes it very clear that the text is describing what this goddess holds in her left hands, not the implements “in the left hands of the nāga lord Pīta.”

n.101At this point the Sanskrit witness (NE 1480/9) cuts off. The lacuna continues until the fragment of the manuscript’s scribal colophon. For the colophon, see n.­9.

n.102Following D and S gul+ma ya. Y, K, J, and C read gu la ma ya. It is likely, however, that the correct reading is gulmamasi.

n.103D reads sarba nA ga, and S reads sarb+ba nA ga. This transliteration emends the reading in the Tibetan witnesses to the grammatically correct second-case plural declension sarvanāgān.

n.104D reads sarba nA ga Na, and S reads sarb+ba nA ga na. This transliteration emends the reading in the Tibetan witnesses to the grammatically correct second-case plural declension sarvanāgān.

n.105D reads sarba nA ga na, and S reads sarb+ba nA ga na. This transliteration emends the reading in the Tibetan witnesses to the grammatically correct second-case plural declension sarvanāgān.

n.106D reads sarba nA ga na, and S reads sarb+ba nA. This transliteration emends the reading in the Tibetan witnesses to the grammatically correct second-case plural declension sarvanāgān.

n.107This reading emends the reading in D ba sha ba na ya to the expected Sanskrit phrase vaśaṃ ānaya, or “bring [them] under [my] control.” S reads ba sha na ya and NE 1489/9 omits.

n.108D and S read zhes gnas pa ni go rims sngon bzhin du klu brgyad po’i ming bri bar bya’o zhes bya’o. The translation “the place where this rite is performed” is a tentative interpretation of the Tibetan phrase zhes gnas pa.

n.109D and S read dge slong rnams dang gzhon nu ma rnams la ston mo bya’o. Although it is not made explicit here, the fact that this line instructs the officiant to hold a feast for monastics and young women (Tib. gzhon nu ma) most likely means that we should read the term “monastics” (Tib. dge slong rnams) as a reference to the saṅgha of fully ordained nuns (dge slong ma rnams). This reading agrees with the instructions for holding a preliminary feast for performing rites to the goddess Mārīcī that we have already seen above.

n.110D and S read ma lo ka’i me tog, and J and C read ma le ka’i me tog. The identity of this flower is unknown.

n.111At this point the text appears to omit Takṣaka, who we have already learned from the sequence above is typically in the western quarter.

n.112D and S read srin po’i phyogs su, literally “in the direction of the rākṣasas.” This is a somewhat obscure term for the southwest ordinal direction.

n.113Following C phyag rgyas bkug. D and S read phyag rgya bkug.

n.114Following S ba rA ha mu khi. D reads bA rA hA mu khi.

n.115D and S read sngags dang bcas pa’i ming bcug ste. The text does not actually specify whether this ritual is performed solely for oneself or if it can be performed for someone else. This passage is translated as if the instructions are for performing this rite for oneself, but it seems possible that it might be performed for oneself or for someone else.

n.116Following Y, K, and S pu tra ja ri. D reads pu tra dza re.

n.117Following Y and K a va na ci ti ka. D and S read a va na ca ti ka.

n.118D and S read a sha pa tsa ti ka dang. This translation is tentative and emends the Tibetan transliteration to āśapacatika.

n.119D reads kran+ta, and S reads kran ta. This transliteration is tentative and emends the reading to the Sanskrit krānta. The identity of this substance is not clear.

n.120D reads sran ti, and S reads sra na ti. The identity of this substance is not clear.

n.121D and S read lag smin. This transliteration emends the reading in the Tibetan witnesses to the correct Sanskrit spelling of the consonant cluster kṣmī. The term lakṣmī is used for a number of plants, so the precise identity of this substance is not clear.

n.122D and S read tran+ta ni khyab ’jug rnam par gnon pa la bya’o, and J and C read tran+tu ni khyab ’jug rnam par gnon pa la bya’o. This verse is obscure and remains open to several interpretations. We suspect that tran+ta is actually a corruption of krānta, which would suggest a relationship between the plant krānta and vikrama (rnam par gnon pa), a common name of the deity Viṣṇu. The fact that tran+ta refers to a plant here is assumed by distributing the term ldum bu from the next two lines, and the statement is read as a gloss for this plant by distributing the complete verb shes par bya’o from the third line in this series of three glosses.

n.123D and S read pra tsan+ta ni ldum bu rtag tu sngun ma la bya’o, Y and K read pra tsan ti na sum bu rtag tu sngum ma la bya’o, and J and C read pra tsan+ta ni sum bu rtag tu san ma la bya’o. This line is obscure but, like the line that precedes it, appears to be a gloss providing the more common name of a plant. The term pracanta may correspond to the substance that was transliterated above as sranti.

n.124D and S read a na pa tsi ta ka. This may correspond to the substance a sha pa tsa ti ka that was mentioned above.

n.125D and S read ha ri ti b+hu ba ha na. The term is read here as a corrupted transliteration of haritabhuvana, because it is grouped together with the other mineral components of this powder. It is possible, however, that this is a term for the turmeric rhizome, or the “conveyance” (vahana, i.e., one of the ways that the plant spreads or travels) produced by (bhu) the turmeric plant (harita).

n.126D reads spangs, S reads sbangs, and Y, K, J, and C read spags. This translation is tentative.

n.127D reads bdag nyid kyi myos byed dang bcas pa, and Y, K, J, C, and S read bdag nyid kyis myos byed dang bcas pa. This translation is tentative.

n.128Following Y and K yang nges par dran pa’o. D and S read yang nges par dran pa’i.

n.129D reads lang ba’i bye’u, and S reads lang pa’u bye’u. This translation is tentative.

n.130Following Y, K, and S sida sa ra. D and J read sat+si ra, and C reads satmi. The full phrase in Y, K, and S reads sida sa ra ma sa ra mong dkar po, where the Tibetan mong dkar po is a gloss of the transliterated Sanskrit sida sa ra ma sa ra. This transliteration is read as a corrupted form of the Sanskrit sitasāramāṣa or “the pulses” (māṣa) of the sitasāra plant. The Tibetan gloss mong might be read as a clarification of the type of pulse, identifying it as a “white mung bean” (mong dkar po).

n.131D reads sngon ma kShi ba ki, Y and K read sngon ma kSha ba ki, and S reads sngon ma kShi bi ka. This is interpreted as a corrupted transliteration of the Sanskrit kṣavaka, which is identified as a number of plants. It is possible that this line also indicates that the Tibetan sngon ma, rendered above as sngun ma, is identified with this same plant.

n.132D and S read sum cu pa’i bdag po yis nges par rang gi chung ma dbang du byed do, Y, K, and J read sum bcu pa’i bdag po yis nges par rang gi chung ma dbang du byed do, and H reads sum cu pa’i bdag pos nges par rang gi chung ma dbang du byed do. The Lord of the Thirty is likely an epithet for Śakra/Indra. It is also very likely that this passage is corrupted and should instead match previous instances where this text emphasizes the potency of the ritual action by mentioning that it even works on a divine being.

n.133D and S read khyab ’jug gis gnon pa. This name is similar to the explanation of the medicinal plant tranta earlier, in the Tibetan phrase tran+ta ni khyab ’jug rnam par gnon pa la bya’o, where the terms khyab ’jug (viṣṇu) and rnam par gnon pa (vikrama) were taken as alternate names for the medicinal plant tranta (krānta?). See n.­122.

n.134D and S read kha ga ba ti’i ’khor lo. This translation emends the Tibetan transliteration of this Sanskrit term to *kagapati both here and below, which literally means “lord of birds” and is an epithet for Garuḍa.

n.135D and S read lha’i dbang po sbas pa. This translation is tentative. The Tibetan might be a translation of the Sanskrit *indragupta, but the identity of this substance is not clear.

n.136This translation is tentative. D and S read dbang po sbas pa, which might be a translation of the Sanskrit *indragupta, though the identity of this substance is not clear.

n.137D and S read dbang po’i chung ma shi tsa ’ong te ma thag tu dbang du byed do. Presumably, this means that when the husband’s time of death comes, his wife will be enthralled.

n.138Following D and S de’i dbus su mchod rten dang. There is a chance that the word order is a bit off in the Tibetan and that Mārīcī is visualized in the middle of a caitya.

n.139Following D and S A barta yi Sh+ya mi, although the grammatically correct Sanskrit reading is *āvartayiṣyāmi.

n.140Following S ba rA ha mu khi. D reads bA rA hA mu khi.

n.141Following D ban d+ha ni. S reads baM d+ha mi, which suggests amending this to *bandhāmi.

n.142D reads bA rA hA mu khi, Y, K, J, and C read ba rA hA mu khi, and S reads ba ra ha mu khi. This transliteration emends the Tibetan witnesses to varāhamukhi, which is used as the standard spelling for this term in this translation.

n.143Following D and S gran+thi ban d+hA mo. This transliteration should be amended to the grammatically correct reading *granthaṃ bandhāmi.

n.144D reads btus shing myur du ’grub pa’i sngags, and S reads btus shing myur du ’grub pa’i bsngags. These lines refer to the process of “mantra extraction” (Skt. mantroddhara; Tib. sngags btu ba), the extraction of individual syllables from any arrangement of the vowels, consonants, semivowels, etc. of the Sanskrit alphabet and their arrangement in a mantra. This text appears to follow the arapacana ordering of the Sanskrit alphabet.

n.145D and S read dbyangs yig dang po. This is most likely a term for the first vowel of the Sanskrit alphabet, a.

n.146This is most likely a reference to Mārīcī’s seed syllable māṃ.

n.147D and S read sa bon sde tshan gnyis dang po/ /dbyangs yig bzhi po dbye byas la. This mantra formation refers to the arapacana alphabet, where “the initial consonant of the second class and the fourth vowel” would give us the mantra syllable rī.

n.148D and S read ’og ma’i steng nas, and Y and K read ’og ma’i steng na. This translation is tentative.

n.149Following Y, K, and S ma hA tsI va ra ma si. D reads ma hA tsI ba rA ma si.

n.150D and S read de ba tA yai, Y and K read de wa ta ye, U reads da ba tA ye, and H reads de bA tA ye. Note that only three of the four goddesses who usually attend Mārīcī in this mantra appear here. It is likely that the term devatāyai has worked its way into this mantra through a scribal error for the goddess Vattāli, who would usually appear here in the mantra.

n.151Following S ba rA ha mu khi. D reads bA rA hA mu khi.

n.152D and S read thams cad du yi ge’i mgo’i nang du byas te. This translation is tentative. Presumably, this refers to the location for depicting the eight vidyādharas mentioned in the previous line.

n.153Following S ban d+ha bin. D reads ban d+ha bi na. Both readings appear to be corrupted, and it is possible that this transliteration should simply read bandha.

n.154Following S gsad pa. D reads gsang ba, and C reads gsod pa.

n.155D and S read ’bar ’khrugs pa’i me tog. The identity of this flower is not clear.

n.156D reads mtshon gyis bcad pa’i ras la, J and C read mtshon gyis ma bcad pa’i ras la, and S reads tshon gyis bcas pa’i ras la. The line is not explicit on this point, but the reference here is likely to a piece of cloth from someone who has died in battle, a writing medium that is also juxtaposed with the medium of a “piece of cloth from a charnel ground” in The King of Ritual Manuals from the Tantra of Māyā Mārīcī’s Arising (Toh 565, 1.­79).

n.157D and S read ’khor lo la sogs pa bris. For the last set of five ritual instructions it has been assumed that the drawing is of Mārīcī because this is explicitly mentioned in the very first set of instructions in this grouping. Here it is explicitly noted that one should draw the mantra wheel, and it is assumed that this is the ritual object for the instructions that follow from this point forward.

n.158D and S read de nyid kyi sol bas. This translation assumes that the Tibetan de nyid kyi refers to charcoal from a charnel ground.

n.159D and S read rje btsun ma. Bhaṭṭārikā is usually an epithet for Tārā, but it is possible that this is a reference to the goddess Durgā, as the rite also incorporates an image of Gaṇapati. It is of course also possible that the term is used here as an epithet for Mārīcī.

n.160D and S read de gnyis kyi gos la. This translation is tentative. The referent of de gnyis kyi is not clear, but it is possible that this is a corruption of the phrase de nyid kyi and that it indicates that the same type of cloth can be used for this rite.

n.161D and S read phyugs kyi rwas. The term phyugs (Skt. paśu) can signify any domesticated beast or livestock.

n.162D and S read stobs po che’i rnal ’byor gyis, and Y and K read stobs po che’i rnal ’byor gyi. Here the term stobs po che might refer to a mantra related to the wrathful being (krodha) Mahābala.

n.163D reads chus khyer ba’i shing gi smyu gus, and S reads chus khyer ba’i shing gi snyu gus. This translation is tentative. The Tibetan chus khyer ba’i shing refers to either driftwood or a type of tree that is “carried/conveyed by water” (chus khyer ba).

n.164D and S read sa la sogs pa la sgrub pa byas la, and Y, K, J, and C read sa la sogs pa la bsgrub pa byas la. This translation is tentative.

n.165D and S read de’i nang du bcug ste. This translation is tentative. The object of this phrase is not clear, but it is interpreted here as a reference to the specific command that one places inside this mantra wheel to bind a specific ritual target.

n.166D and S read phan tshun ’bye zhing mi mthun pa chen por ’gyur ro. This phrase indicates that this rite might actually be a rite for sowing discord, not a binding rite.

n.167Following S and H thod le kor. D reads thod lo kor, Y and K read thod lo bskor, J reads thod lo skor, and C reads the lo kor. This translation interprets the Tibetan thod le kor as a translation of the Sanskrit term for chalk, kaṭikā.

n.168This is a reference to the protection spell from The Mārīcī Dhāraṇī (Toh 564, 1.­7).

n.169D reads g.yon pa ser po cung zhig bzhad pa’i zhal sangs rgyas kyi zhal g.yon pa’o, and S reads g.yon pa ser po cung zhig bzhad pa’i zhal/ sangs rgyas kyi zhal g.yon pa’o. This translation is tentative. There is no indicator that this is the second tier of heads above (steng du) the first tier, and both faces are described as the “left” (g.yon pa) face, when one clearly must be the left and the other the right. This translation assumes that the first face is the right and the second the left, following the convention established in the description of the first tier of Mārīcī’s six faces.

n.170Following S ba rA ha mu khi. D reads ba rA hA mu khi.

n.171Following S ba rA ha mu khi. D reads bA rA hA mu khi. This translation emends the reading in the Stok Palace Kangyur to *ba rA ha mu khI (Skt. varāhamukhī) to reflect the correct Sanskrit spelling for the nominative singular feminine.