तिङन्ततुदादिप्रकरणम्

Covers SK 2534–2542 (9 sūtras): all present. Every unit has its own kaumudī. The spine is not an Aṣṭādhyāyī run. The SK follows the Dhātupāṭha order of the tudādi roots (1281 tuda … 1437 piśa, our numbering) and teaches each rule at the root where it first bites, drawing on AS books 1, 3, 6, 7 and 8 (3.1.77 → 6.4.47 → 1.2.2 → 8.3.76 → 1.3.61 → 6.1.140–141 → 8.2.21 → 7.1.59). Every AS-ref below is checked against references/ashtadhyayi/sutraani-data.json.

• The unit sizes are uneven. sk2536 (vija iṭ) carries the Dhātupāṭha run from olajī to sphula: 101 roots, ours 1290–1390. The outline groups that list by the kinds of forms it teaches and names examples. It does not list every root.
• The Dhātupāṭha's own sub-lists. The gaṇa closes three sub-lists with "vṛt":
• the kuṭādi (1366 kuṭa … 1401 kuṅ), whose affixes are ṅit by gāṅkuṭādibhyo 'ñṇin ṅit (1.2.1, SK 2461), closed at 2537;
• the kirādi (kṝ, gṝ, dṛṅ, dhṛṅ, pracch), closed at 2541;
• the mucādi (muc … piś), which take num before śa by 2542 and close the tudādi at 2542.
• Anuvṛtti from elsewhere in the SK. Most of the sūtras here carry words from rules taught far away:
• kartari from kartari śap (3.1.68, SK 2167) at 2534;
• ārdhadhātuke from ārdhadhātuke (6.4.46, SK 2307) at 2535;
• ṅit from gāṅkuṭādibhyaḥ (SK 2461) at 2536;
• mūrdhanyaḥ from apadāntasya mūrdhanyaḥ (8.3.55, SK 210) and vā from sivādīnāṃ vā… (8.3.71, SK 2359) at 2537;
• ātmanepadam from anudāttaṅita… (1.3.12, SK 2157) and, by ca, śitaḥ from śadeḥ śitaḥ (1.3.60, SK 2362) at 2538;
• upāt from upāt pratiyatna… (6.1.139, SK 2552) and suṭ kāt pūrvaḥ (6.1.135, SK 2553) at 2539–2540. Both are taught later, in the saṃhitā-bound sections;
• graḥ from gro yaṅi (8.2.20, SK 2639, later) and ro laḥ from kṛpo ro laḥ (8.2.18, SK 2350) at 2541;
• num from idito num dhātoḥ (7.1.58, SK 2262) at 2542.
• The outline names them where BM does.

tiṅantatudādiprakaraṇam — outline (p49, SK 2534–2542, all 9 sūtras)

p48 (tiṅantasvādiprakaraṇam) closed at 2533 (aśnoteś ca, 7.4.72) with dāśnoti and "vṛt", and the print adds "iti svādayaḥ".

• Where the prakaraṇa begins. The print heads it "tiṅante tudādayaḥ 46" (NSP p. 408). Our prakaraṇa number 49 differs systemically. sk2534 opens with the Dhātupāṭha entry 1281 tuda vyathane and "itaḥ ṣaṭ svaritetaḥ" (six svaritet roots, tud … kṛṣ, 1281–1286; BM), then tudādibhyaḥ śaḥ (3.1.77).
• Where it ends. At 2542 with 1437 piśa and "vṛt: mucādayo vṛttāḥ tudādayaś ca". The print closes "iti tudādayaḥ" ("Here end the tudādi roots"). Ours lacks it, so the English supplies it in brackets (NSP p. 413). p50 opens with rudhādibhyaḥ śnam (3.1.78, SK 2543) and 1438 rudhir.
• Dhātu numbers. Ours runs one lower than the print throughout (print 1282 tuda … 1438 piśa). The English, and this outline, give ours.
• The pada-sets as the kaumudī announces them.
• Six svaritet: tud, ṇud, diś, bhrasj, kṣip, kṛṣ (1281–1286).
• ṛṣī is parasmaipadin (1287).
• Four ātmanepadin: juṣī, ovijī, olajī, olasjī (1288–1291).
• Parasmaipadin from ovraścū (1292) on.
• gurī is anudāttet (1396).
• Four parasmaipadin: ṇū, dhū, gu, dhru (1397–1400).
• kuṅ, pṛṅ, mṛṅ are ṅit.
• Seven parasmaipadin: ri … gṝ (1404–1410).
• dṛṅ and dhṛṅ are ṅit.
• Sixteen parasmaipadin: pracch … śadḷ (1413–1428).
• Six svaritet: mila, mucḷ, lupḷ, vidḷ, lipa, ṣica (1429–1434).
• Three parasmaipadin: kṛtī, khida, piśa (1435–1437).
• Contents in outline.
• 2534: tud and the svaritets. śa for śap. It is apit and therefore ṅit: tudati, nudati, bhṛjjati.
• 2535: bhrasj, then kṣip, kṛṣ, ṛṣī, juṣī, ovijī. Optional ram for the r and upadhā in the ārdhadhātuka, and saṃprasāraṇa before kṅit by the vārttika: babharja / babhrajja, bharṣṭā / bhraṣṭā, bhṛjjyāt; kraṣṭā / karṣṭā, akrākṣīt / akārkṣīt / akṛkṣat.
• 2536: vij, then the 101-root run olajī … sphula. An iṭ-initial affix after vij is ṅit: udvijitā. The run adds saṃprasāraṇa (vṛścati, vavraśca, vivyāca), nuṭ (ānarccha), num before śa (tṛmphati), optional iṭ (lobhitā / lobdhā, cartsyati), the kuṭādi (cukuṭitha, kuṭitā) and more.
• 2537: sphur, sphul; then the kuṭādi close and pṛṅ. Optional ṣ after nir, ni and vi: niḥṣphurati / niḥsphurati; jugutha, gutā; vyāpriyate.
• 2538: mṛṅ, then ri … kṝ. Ātmanepada only in luṅ, liṅ and before a śit affix: mriyate, mṛṣīṣṭa, amṛta, but mamāra, mariṣyati; riyati; kirati, cakaratuḥ.
• 2539–2540: kṝ after upa and prati. suṭ before the k: upaskirati, upāskirat, upacaskāra, pratiskirati.
• 2541: gṝ, then dṛṅ … mucḷ. Optional l for the r before a vowel: girati / gilati. Then ādriyate, pṛcchati, sasraṣṭha, majjati / mamaṅktha, spraṣṭā / sparṣṭā, vicchāyati, sīdatī / sīdantī.
• 2542: muc and the rest. num before śa: muñcati, lumpati, vindati, limpati, siñcati, kṛntati, khindati, piṃśati.

Threads that run through the prakaraṇa:

• śa is apit, so it is ṅit. sārvadhātukam apit (1.2.4, SK 2234) makes it ṅit, and every ṅit-conditioned operation fires before it in the sārvadhātuka. This is the class-mark.
• No guṇa: tudati, not *todati. śa is nitya and is made first, and being ṅit it bars the laghūpadha-guṇa (BM). TB gives an alternative ground: the paribhāṣā 'akṛtavyūhāḥ'.
• saṃprasāraṇa by grahijyā… (6.1.16, SK 2412): bhṛjjati (bhrasj), vṛścati (vraśc), vicati (vyac), pṛcchati (pracch).
• Nasal deletion by aniditām… (6.4.24, SK 415), undone for the tṛmphādi by a vārttika (2536) and for the mucādi by 2542.
• riṅ and iyaṅ (7.4.28, SK 2367; 6.4.77, SK 271) for the ṛ-final ṅit roots: mriyate, ādriyate, vyāpriyate.
• iyaṅ and uvaṅ for i- and u-final roots: riyati, piyati, nuvati, dhuvati, suvati.
• ir for ṝ (ṝta id dhātoḥ, 7.1.100, SK 2390): kirati, girati.
• Outside the sārvadhātuka, the root goes back to its own rules. Before a conjunct-final root the liṭ is not kit (asaṃyogāl liṭ kit, 1.2.5), so there is no saṃprasāraṇa: babhrajja, vavraścatuḥ, papracchatuḥ. It returns before a kit affix: bhṛjjyāt, vṛścyāt, vicyāt, vivicatuḥ.
• The am augment and the optional iṭ in thal. Optional am comes by anudāttasya cardupadhasyānyatarasyām (6.1.59, SK 2402): kraṣṭā / karṣṭā, spraṣṭā / sparṣṭā, amrākṣīt / amārkṣīt. For sṛj it is obligatory before a non-kit jhal (sṛjidṛśor jhaly amakiti, 6.1.58, SK 2405): sraṣṭā, asrākṣīt. In thal:
• optional iṭ by Bhāradvāja's restriction (ṛto bhāradvājasya, 7.2.63, SK 2296): babharjitha / babharṣṭha, papracchitha / papraṣṭha;
• optional iṭ by vibhāṣā sṛjidṛśoḥ (7.2.65, SK 2404): sasarjitha / sasraṣṭha;
• optional iṭ for ū-marked roots (7.2.44, SK 2279): vavraścitha / vavraṣṭha.
• sic, ksa or aṅ in luṅ. The vārttika spṛśamṛśakṛṣa… (under na dṛśaḥ, 3.1.47, SK 2407) makes sic optional, and ksa (śala igupadhād aniṭaḥ ksaḥ, 3.1.45, SK 2336) comes otherwise: akrākṣīt / akārkṣīt / akṛkṣat, asprākṣīt / aspārkṣīt / aspṛkṣat, amrākṣīt / amārkṣīt / amṛkṣat. For tṛp here sic is obligatory (atarpīt), because the vārttika is an apavāda of aṅ and so belongs to the puṣādi tṛp (BM). The ḷ-marked roots take aṅ (puṣādi…, 3.1.55, SK 2343): amucat, alupat. So do lip and sic (lipisicihvaś ca, 3.1.53, SK 2418), optionally in taṅ (3.1.54, SK 2419): alipata / alipta, asicata / asikta.
• Why a root is read twice. ucchī, kaḍa and ṣadḷ are also read in bhvādi. The tudādi reading gives the feminine śatṛ optional num by ācchīnadyor num (7.1.80, SK 445): uñchatī / uñchantī, sīdatī / sīdantī. With śap alone śapśyanor nityam (7.1.81, SK 446) would make num obligatory. The two readings also differ in accent: śap is pit and so anudātta, while śa is udātta (BM, TB). For śadḷ the re-reading is for accent only. ṇud is re-read for parasmaipada even when the fruit goes to the agent, and mila for taṅ.
• The kaumudī lists forms bare. Lakāra, pada and person labels are the translator's, given in Pāṇinian terms. Derivation stages follow BM and TB as recorded in the unit notes.
• Where the commentary lies. Only the kaumudī blocks are translated (SK-TRANSLATE §25).
• TB on 2534–2535 is printed at the tail of sk.2533.t1 (after "iti svādayaḥ"). It runs on in sk.2533.n1, which is running TB text on NSP p. 408, not a footnote. The OCR of sk.2533.t1 then resumes at the top of p. 409. sk.2535.t1 carries only TB on the vārttika.
• TB on 2542 (and on 2543, for p50) is at the tail of sk.2541.t1. sk2542 has no t1.
• BM. sk.2542.b1 duplicates sk.2541.b1 word for word. The real BM on 2542, and the tail of BM on 2541 ("śatā tu nāsti…", mila, "paṭhitasya", "taṅartha"), are read from Wikisource uttara-2/0064.
• NSP footnotes in no source.json or only in garbled form. The note on "antaraṅgatvād iyaṅ" (p. 412; read on the scan) is transcribed in a textual note on sk2538. The note on tṛpa / tṛmpha (p. 410, under sk.2536.n1) cites the Bhāṣya's "śe tṛmpādīnām": tṛmpati, tṛmphati.
• A stray kaumudī line inside TB. The ṛpha / ṛmpha forms have intruded into the TB at sk.2536.t1 ("vivyācet | ṛphati…"). The print has the same.
• A BM text slip. BM's "tanādigaṇa eva" under ṇuda (2541) is in Wikisource too. Read tudādi.

Where the p47–53 audit (.work/sk-run/p47-53-audit/sk2534-2542.md) corrects a reading, this outline follows the corrected reading. The fixer is applying these to en.yml in parallel.

• 2534: the sūtra. "After tud and the rest, śa comes, in its own sense, when a sārvadhātuka denoting the agent follows; it is an apavāda of śap." That BM gives the setting belongs in the notes. The attribution of kartari to SK 2167 is the translator's, since BM gives the sense but names no source.
• 2535, TB. The others carry lopa from sanaḥ ktici lopaś cāsyānyatarasyām (6.4.45, SK 3315), the sūtra just before ārdhadhātuke. TB's five faults of the "rasor" view are apit liṭ, san, liṅ, sic and caṅ after ṇic. yaṅluk is the extra fault of the bhṛj-substitute view ("doṣo yaṅluki cādhikaḥ").
• 2535, akṛkṣata. It is the sic plural and the ksa singular (BM glosses it twice).
• 2536, kṣura vilekhane. TB glosses vilekhana as chedana: "in cutting" ("keśān kṣurati"), not "scratching".
• 2536, vṛhū. The others read a root beginning with b, the third of the p-class: bṛhū (BM "pavargatṛtīyādiḥ").
• 2536, ucchī and hila. "vivāsa" (ending, banishing) and "bhāvakaraṇa" (expressing feeling) are translator's glosses. No commentator glosses either.
• 2540, gṝ nigaraṇe. BM glosses nigaraṇa as eating (bhakṣaṇa).
• 2542, piśa avayave. BM: "in an action on parts" (avayavakriyāyām).

Print readings (NSP, scan-confirmed) that the outline follows:

• mittvād (2535);
• mimiccha, katthanayuddhanindā-, dvitīya-, cartsyati, śabdopakaraṇayoḥ, starhitā, atārṅkṣīt and vā bhrāśa… (2536);
• hrasvād aṅgāt and agutām (2537);
• kīryāt (2538);
• upāt prateś ca (2540);
• choptā, tālavyāntau, -avasādaneṣu and jvalādau (2541);
• the colophon "iti tudādayaḥ" (2542).

Still open for Ajai (mūla). The English emends two forms and keeps two misprints:

• 2537, niḥṣphurati. Ours and the print read "niḥṣphusphurati", with the syllable doubled. Wikisource has niṣphurati. The English gives niḥṣphurati / niḥsphurati.
• 2541, sīdatī. Ours and the print read "sīdati | sīdantī". BM (śatṛ, then ṅīp, then optional num) and Wikisource support sīdatī / sīdantī, and the English gives sīdatī.
• Kept as printed: 2536 maḍati under mṛḍa (mṛḍati expected; Wikisource mṛḍati) and truḍitā under truṭa (truṭitā expected; Wikisource truṭitā).
• Low priority: 1393 vruḍa (the print glyph may be vṛḍa; Wikisource vruḍa). "na bhakurchuram" under kura (the sūtra reads na bhakurchurām).

Ordering: SK order throughout.

A. tud and the svaritets: śa for śap, apit and therefore ṅit — running example: tud + śa + ti → tudati (no guṇa); bhrasj + śa + ti → bhṛsj-a-ti → bhṛjjati

• 2534 tudādibhyaḥ śaḥ 3.1.77 — after tud and the roots that follow it, śa comes in its own sense before a sārvadhātuka denoting the agent. It is an apavāda of śap (BM).
• kartari comes from kartari śap (3.1.68, SK 2167). The attribution is the translator's.
• Root entry 1281 tuda vyathane. tud + śa + ti: śa is apit, so it is ṅit by sārvadhātukam apit (SK 2234), and the light-penultimate guṇa does not come: tudati, tudate.
• liṭ tutoda, and tutoditha with iṭ, which is obligatory in thal by the krādi restriction because the root neither ends in ac nor has a (BM). tutude.
• luṭ tottā (aniṭ). luṅ atautsīt (halanta vṛddhi) and atutta.
• 1282 ṇuda preraṇe (ṇopadeśa, aniṭ): nudati, nunoda, nottā.
• 1283 diśa atisarjane (atisarjana = giving):
• deṣṭā (ṣ by vraśca…, 8.2.36, SK 294; ṣṭutva);
• dikṣīṣṭa (kit by liṅsicāv ātmanepadeṣu, 1.2.11, SK 2300);
• adikṣat, adikṣata (ksa by SK 2336).
• 1284 bhrasja pāke. The cooking here is roasting (bharjana), not the cooking of rice (TB). bhrasj + a + ti:
• śa is ṅit, so grahijyā… (SK 2412) gives ṛ for the r, with pūrvarūpa: bhṛsj-a-ti;
• s → ś by ścutva, then ś → j by jaśtva: bhṛjjati, bhṛjjate.
• BM closes by setting up ṇal, which leads into 2535.

B. bhrasj in the ārdhadhātuka: optional ram, and saṃprasāraṇa before kṅit; then kṣip, kṛṣ, ṛṣī, juṣī, ovijī — running example: bhrasj + ṇal → babharja / babhrajja; bhrasj + tās → bharṣṭā / bhraṣṭā; bhrasj + yāt → bhṛjjyāt

• 2535 bhrasjo ropadhayoḥ ram anyatarasyām 6.4.47 — for bhrasj, the augment ram optionally comes in place of the r and the upadhā before an ārdhadhātuka.
• ārdhadhātuke is under adhikāra from SK 2307. bhrasjaḥ is a sixth case of the part, and ropadhayoḥ a sixth case of place (BM).
• ram is mit, so it comes after the last vowel (mid aco 'ntyāt paraḥ, 1.1.47, SK 37). By the sthānaṣaṣṭhī, the r and the upadhā cease: bh-a-r-j.
• BM reads "sthāna" as occasion: where r and s would be pronounced, r after the a is used instead. TB gives ram the character of both substitute and augment, like śnam. Others carry lopa from SK 3315, the view the audit corrected. The English stays neutral (§16).
• liṭ: atus follows a conjunct, so it is not kit and there is no saṃprasāraṇa.
• With ram: babharja, babharjatuḥ; babharjitha / babharṣṭha (iṭ optional in thal by Bhāradvāja's restriction; without iṭ, j → ṣ by vraśca, th → ṭh); babharje.
• Without ram: babhrajja (s → ś → j after doubling), babhrajjatuḥ, babhrajjitha.
• babhraṣṭha: s lost by skoḥ… (8.2.29, SK 380), j → ṣ by vraśca (SK 294). babhrajje.
• luṭ bhraṣṭā / bharṣṭā, lṛṭ bhrakṣyati / bharkṣyati (k by ṣaḍhoḥ kaḥ si, 8.2.41, SK 295).
• Vārttika kṅiti ramāgamaṃ bādhitvā saṃprasāraṇaṃ pūrvavipratiṣedhena. Before a kit or ṅit affix, saṃprasāraṇa sets ram aside. ram, the later rule, would otherwise give *bharjyāt (BM). TB adds that saṃprasāraṇa could also win as nitya.
• āśīrliṅ parasmaipada: bhṛjjyāt, bhṛjjyāstām (yāsuṭ is kit).
• Ātmanepada: bharkṣīṣṭa / bhrakṣīṣṭa. The affix is not kit, so there is no saṃprasāraṇa.
• luṅ abhārkṣīt / abhrākṣīt; abharṣṭa / abhraṣṭa (sic lost by jhalo jhali, 8.2.26, SK 2281).
• 1285 kṣipa preraṇe (aniṭ): kṣipati, kṣeptā, akṣaipsīt, akṣipta.
• 1286 kṛṣa vilekhane (aniṭ):
• kṛṣati;
• luṭ kraṣṭā / karṣṭā: am optional by SK 2402, and without am, guṇa with r;
• kṛṣyāt and kṛkṣīṣṭa take neither am nor guṇa, being kit.
• luṅ: sic optional by spṛśamṛśakṛṣa… (under SK 2407), otherwise ksa.
• With sic, am is optional: akrākṣīt / akārkṣīt. With ksa: akṛkṣat.
• In taṅ, sic is kit by SK 2300, so there is no am: akṛṣṭa, akṛkṣātām.
• akṛkṣata is the sic plural and also the ksa singular (audit). akṛkṣātām, akṛkṣanta (ksa; ant, since jh follows a; NSP footnote).
• 1287 ṛṣī gatau, parasmaipadin: ṛṣati, ānarṣa.
• 1288 juṣī, with four ātmanepadin: juṣate.
• 1289 ovijī bhayacalanayoḥ, usually with ud: udvijate. In udvijitā the light-penultimate guṇa would be due, which leads into 2536 (BM).

C. vij: an iṭ-initial affix is ṅit; then the 101-root run olajī … sphula — running example: ud-vij + iṭ + tā → udvijitā (not *udvejitā)

• 2536 vija iṭ 1.2.2 — an affix beginning with iṭ (iḍādi) that follows vij is treated as ṅit.
• ṅit comes from gāṅkuṭādibhyo 'ñṇin ṅit (1.2.1, SK 2461; BM).
• How iṭ comes to mean "an affix beginning with iṭ":
• BM carries "yasya ādiḥ" by frog's leap (maṇḍūkapluti) from vṛddhir yasyācām ādis tad vṛddham (1.1.73, SK 1335);
• TB's moderns get it by implication (lakṣaṇā), since iṭ itself being ṅit would do nothing.
• The vij meant is ovijī of the tudādi and rudhādi, not juhotyādi vijir, so vivejitha keeps guṇa (TB; cf. p46 2503).
• ud-vij + iṭ-tā: the iṭ-initial affix is ṅit, so there is no light-penultimate guṇa: udvijitā, udvijiṣyate. BM adds udvijiṣīṣṭa, udavijiṣṭa.
• 1290 olajī and 1291 olasjī vrīḍāyām: lajate, leje; lajjate, lalajje.
• The rest of sk2536: the run 1292 ovraścū … 1390 sphula. From ovraścū on the roots are parasmaipadin. Most entries give only a meaning and the laṭ form (śubhati, śumbhati; dṛbhati; kilati…). Where the kaumudī gives more, it teaches these kinds of forms:
• saṃprasāraṇa before ṅit śa and in the liṭ abhyāsa: vraśc, vyac.
• 1292 ovraścū: vṛścati (grahijyā…, SK 2412). liṭ vavraśca:
• in the abhyāsa, liṭy abhyāsasyobhayeṣām (6.1.17, SK 2408) gives ṛ for the r, and urat (7.4.66, SK 2244) turns that ṛ to a;
• that a is sthānivat by acaḥ parasmin pūrvavidhau (1.1.57, SK 50), so the v takes no saṃprasāraṇa, by na saṃprasāraṇe saṃprasāraṇam (6.1.37, SK 363).
• BM derives "r first" from that prohibition as a jñāpaka. TB's moderns call this laboured.
• vavraścatuḥ (not kit, being after a conjunct). vavraścitha / vavraṣṭha: the root is ū-marked, so iṭ is optional.
• vraścitā / vraṣṭā, vraściṣyati / vrakṣyati.
• vṛścyāt (kit). avraścīt / avrākṣīt: s lost by skoḥ, ṣ by vraśca, k by ṣaḍhoḥ.
• 1293 vyaca: vicati; vivyāca (the y of the abhyāsa takes saṃprasāraṇa, not the v, by SK 363); vivicatuḥ (kit); vyacitā, vyaciṣyati; vicyāt; avyācīt / avyacīt (vṛddhi optional by ato halāder laghoḥ, 7.2.7, SK 2284).
• Vārttika vyaceḥ kuṭāditvam anasi. vyac counts as kuṭādi except before as. It "does not operate here": anasi is a paryudāsa, so the vārttika covers kṛt affixes only (udvicitā, udvicitum; TB). So there is no saṃprasāraṇa in vyacitā (BM, TB). The vārttika's wording ends at anasi (print and Wikisource boundary).
• Guṇa and nuṭ for ṛcch in liṭ.
• 1296 ṛccha. ṇal gives no light-penultimate guṇa, so ṛcchaty ṝtām (7.4.11, SK 2383) supplies it.
• The "dvihal" of tasmān nuḍ dvihalaḥ (7.4.71, SK 2288) stands for "more than one consonant", so nuṭ comes: ānarccha, ānarcchatuḥ.
• No ām, by the exclusion anṛcchaḥ in ijādeś ca gurumato 'nṛcchaḥ (3.1.36, SK 2237; BM).
• ṛcchitā has no guṇa, since ṛcchaty ṝtām carries liṭi.
• Likewise ānarca (1302 ṛca) and ānarpha (1315 ṛpha).
• num restored before śa: the tṛmphādi.
• 1308 tṛmpha: śa is ṅit, so aniditām… (6.4.24, SK 415) deletes the n. The vārttika śe tṛmphādīnāṃ num vācyaḥ restores it: tṛmphati.
• ādi means "of that kind": the roots here with an n (tumpati, tumphati, dṛmphati, ṛmphati, gumphati, umbhati, śumbhati).
• Outside śa the n stays deleted: tatṛmpha, tṛphyāt.
• The ij-ādi heavy roots take ām in liṭ (SK 2237): ṛmphāṃcakāra, umbhāṃcakāra. The NSP footnote cites the Bhāṣya's "śe tṛmpādīnām".
• Optional iṭ.
• 1305 lubha: lobhitā / lobdhā (tīṣasaha…, 7.2.48, SK 2340).
• The ū-marked 1348–1350 tṛhū, stṛhū, tṛṃhū: starhitā / starḍhā; luṅ atṛṃhīt / atārṅkṣīt, atārṇḍhām (h → ḍh → k, anusvāra → ṅ; BM).
• 1324 cṛtī: cartiṣyati / cartsyati by se 'sici… (7.2.57, SK 2506).
• 1351 iṣa: ch by iṣugamiyamāṃ chaḥ (7.3.77, SK 2400), giving icchati; eṣitā / eṣṭā; iṣyāt; aiṣīt.
• sic obligatory for tṛp. 1307 tṛpa: tatarpa, tarpitā, atarpīt. The sic option of the spṛśa… vārttika belongs only to the puṣādi tṛp, being an apavāda of aṅ.
• The kuṭādi (from 1366 kuṭa). Non-ñit, non-ṇit affixes are ṅit by gāṅkuṭādibhyaḥ (SK 2461), so there is no guṇa:
• cukuṭitha, kuṭitā;
• in the singular, cukoṭa / cukuṭa (ṇal is ṇit, or in the uttama optionally not);
• kṛḍitā, tuṭitā.
• 1375 truṭa also takes śyan optionally by vā bhrāśa… (3.1.70, SK 2321): truṭyati / truṭati; tutroṭa. Its luṭ truḍitā is kept as printed (truṭitā expected; for Ajai).
• Lengthening in āśīrliṅ, and its bar. 1340 ṣura: sūryāt, and 1341 kura: kūryāt, lengthened by hali ca (8.2.77, SK 354).
• na bhakurchurām (8.2.79, SK 1629) does not touch kura, since it means karoti only; "only the commentators' explanation is the refuge" (BM).
• It does bar 1372 chura: churyāt.
• Readings and re-readings.
• 1294–1295 uchi, ucchī are read here (as is kaḍa, 1380) for the optional num of the śatṛ and for the accent (see the threads).
• The variant readings the kaumudī records: riha for ripha ("śiśuṃ na viprā matibhī rihanti"); juna for juḍa ("maruto junanti"); p- or ph-final tṛp and dṛp; bṛhū for vṛhū (audit); khuḍa and chuḍa for thuḍa and sthuḍa; sphura sphuraṇe.
• 1327 mṛḍa: the kaumudī's maḍati is kept as printed (mṛḍati expected; for Ajai).
• 1365 likha: lilekha. It is not a kuṭādi root: likhitum is an error, and lekhitā stands (TB).

D. sphur, sphul: optional ṣ after nir, ni, vi; then the kuṭādi close and pṛṅ — running example: nis + sphurati → niḥṣphurati / niḥsphurati

• 2537 sphuratisphulatyor nirnivibhyaḥ 8.3.76 — after nir, ni and vi, a retroflex ṣ optionally replaces the s of sphur and sphul.
• mūrdhanyaḥ is under adhikāra from SK 210. vā comes from sivādīnāṃ vā… (SK 2359; BM, TB).
• niḥṣphurati / niḥsphurati. The mūla is emended from "niḥṣphusphurati" (for Ajai).
• Some read the root as sphara: pasphāra.
• 1391–1395 sphuḍa, cuḍa, vruḍa, kruḍa, bhṛḍa: meanings only (vruḍa is a low-priority reading).
• 1396 gurī udyamane, anudāttet: gurate, jugure, guritā.
• 1397 ṇū (ṇopadeśa, long-vowel final): pariṇūtaguṇodayaḥ. It takes no iṭ, by śryukaḥ kiti (7.2.11, SK 2381). Four parasmaipadin follow:
• ṇū: nuvati (uvaṅ), anuvīt;
• 1398 dhū: dhuvati;
• 1399 gu, a kuṭādi root, so ṅit and no guṇa:
• liṭ juguvitha / jugutha, luṭ gutā, lṛṭ guṣyati;
• luṅ aguṣīt, and agutām (s of sic lost by hrasvād aṅgāt, 8.2.27, SK 2369), aguṣuḥ;
• 1400 dhru / dhruva: dhruvati like guvati; the dhruva reading is seṭ (dudhruvitha, dhruvitā, dhrūvyāt, adhruvīt).
• 1401 kuṅ: seṭ on Kaiyaṭa's long-vowel reading (kuvitā, akuviṣṭa), aniṭ on the Nyāsa's (kutā, akuta). "vṛt": the kuṭādi end here.
• 1402 pṛṅ vyāyāme, usually with vi and āṅ:
• riṅ (SK 2367), then iyaṅ (SK 271): vyāpriyate;
• vyāpapre, vyāpaprāte;
• vyāpariṣyate, with iṭ by ṛddhanoḥ sye (7.2.70, SK 2366);
• vyāpṛta, vyāpṛṣātām.
• 1403 mṛṅ prāṇatyāge, which leads into 2538.

E. mṛṅ: ātmanepada only in luṅ, liṅ and before a śit affix; then ri … kṝ — running example: mṛ + śa + te → mri-y-a-te → mriyate; but mamāra, mariṣyati; and mṛṣīṣṭa, amṛta

• 2538 mriyater luṅliṅoś ca 1.3.61 — after mṛṅ, when it is the base of luṅ, liṅ, or a śit affix, ātmanepada comes, and not elsewhere.
• ātmanepadam comes from SK 2157. By ca, śitaḥ is drawn in from śadeḥ śitaḥ (SK 2362), and prakṛtibhūtāt is supplied (BM, TB). taṅ covers āna too: mriyamāṇa.
• The ṅ-marking alone would give ātmanepada everywhere, so the sūtra is a niyama. The ṅ is kept "for the accent", under tāsyanudātteṅid… (6.1.186, SK 3730): mā hi mṛta (TB).
• mṛ + śa + te: riṅ, then iyaṅ: mriyate.
• Parasmaipada elsewhere:
• mamāra;
• mamartha (no iṭ even on Bhāradvāja's view; BM);
• mamriva (iṭ by the krādi restriction);
• martāsi;
• mariṣyati (iṭ by SK 2366).
• Ātmanepada in liṅ and luṅ: mṛṣīṣṭa, amṛta.
• Seven parasmaipadin follow.
• 1404 ri and 1405 pi: riyati, piyati. iyaṅ comes because it is antaraṅga, and it prevails over guṇa. The NSP footnote says the two are due at once, so nājānantarye does not apply. retā, petā.
• 1406 dhi, 1407 kṣi.
• 1408 ṣū: suvati, savitā.
• 1409 kṝ vikṣepe, seṭ:
• kirati, kirataḥ: i for ṝ by ṝta id dhātoḥ (SK 2390), followed by r;
• cakāra, and cakaratuḥ: guṇa though kit, by ṛcchaty ṝtām (SK 2383);
• karitā / karītā: iṭ optionally long by vṝto vā (7.2.38, SK 2391);
• kīryāt (print), akārīt.

F. kṝ after upa and prati: suṭ before the k — running example: upa + kirati → upa-s-kirati → upaskirati; upa + a-kirat → upāskirat

• 2539 kiratau lavane 6.1.140 — in the sense of cutting, suṭ is added before the k of kirati after upa.
• upāt comes from SK 2552 and suṭ kāt pūrvaḥ from SK 2553 (BM, TB).
• upaskirati. When the sense is not cutting: upakirati (TB).
• Vārttika aḍabhyāsavyavāye 'pi suṭ kāt pūrva iti vaktavyam: suṭ comes even when aṭ or the abhyāsa intervenes. BM ends the vārttika's own words at api. Examples: laṅ upāskirat, liṭ upacaskāra.
• 2540 hiṃsāyāṃ prateś ca 6.1.141 — in the sense of injuring, suṭ is added to kirati after prati, and by ca after upa (BM, TB).
• upaskirati, pratiskirati.
• TB cites Māgha: "uro vidāraṃ praticaskare nakhaiḥ". It is liṭ in the karman, with guṇa by ṛcchaty ṝtām.
• That suṭ and kirati continue from 2539 is the translator's inference.
• 1410 gṝ nigaraṇe ("swallowing"; BM: eating), which leads into 2541.

G. gṝ: optional l for r before a vowel; then dṛṅ … mucḷ — running example: gṝ + śa + ti → gir-a-ti → girati / gilati

• 2541 aci vibhāṣā 8.2.21 — before a vowel-initial affix, l optionally replaces the r of gṝ.
• graḥ comes from gro yaṅi (8.2.20, SK 2639) and ro laḥ from kṛpo ro laḥ (8.2.18, SK 2350; BM).
• ac qualifies the affix (tadādividhi), so the rule does not reach girau and the like (BM).
• girati / gilati; jagāra / jagāla; jagaritha / jagalitha.
• luṭ garitā, garītā, galitā, galītā: the iṭ is optionally long by SK 2391. The print has all four; Wikisource lacks the long-ī forms.
• 1411 dṛṅ ādare (aniṭ, ṅit):
• ādriyate (riṅ, iyaṅ), ādadre, ādadriṣe, ādartā;
• ādariṣyate (SK 2366);
• ādṛṣīṣṭa: kit by uś ca (1.2.12, SK 2368), so no guṇa;
• ādṛta (hrasvād aṅgāt, SK 2369). TB cites Śrīharṣa.
• 1412 dhṛṅ: dhriyate.
• Sixteen parasmaipadin follow.
• 1413 praccha jñīpsāyām (aniṭ):
• pṛcchati (grahijyā…);
• papraccha, papracchatuḥ (not kit, after a conjunct);
• papracchitha / papraṣṭha: Bhāradvāja; ṣ for ch by vraśca, which takes ch with its tuk;
• praṣṭā, prakṣyati, aprākṣīt.
• "vṛt": the kirādi end here. That five kirādi are counted (kiraś ca pañcabhyaḥ) serves the anuvṛtti into rudādibhyaḥ (BM, TB).
• 1414 sṛja visarge:
• sasarjitha / sasraṣṭha (vibhāṣā sṛjidṛśoḥ, SK 2404);
• sraṣṭā, srakṣyati, asrākṣīt, with the augment am by sṛjidṛśor jhaly amakiti (SK 2405);
• sṛjet, sṛjyāt.
• 1415 ṭumasjo śuddhau:
• majjati (s → ś → j), mamajja;
• num by masjinaśor jhali (7.1.60, SK 2517).
• Vārttika masjer antyāt pūrvo num vācyaḥ: the num stands before the final.
• saṃyogādilopa of the s by skoḥ (SK 380): mamaṅktha. In BM's order: s lost, j → g by kutva, num, anusvāra, ṅ by parasavarṇa, g → k by cartva.
• mamajjitha, maṅktā, maṅkṣyati, amāṅkṣīt.
• The vārttika's real fruit is n-deletion in magna, where the n must be penultimate (BM).
• 1416 rujo: roktā, araukṣīt. 1417 bhujo: like ruj.
• 1418 chupa: choptā (print), acchaupsīt.
• 1419–1420 ruśa, riśa (palatal ś; print tālavyāntau): roṣṭā, rekṣyati.
• 1421 liśa: alikṣat (ksa).
• 1422 spṛśa: spraṣṭā / sparṣṭā (am by SK 2402), asprākṣīt / aspārkṣīt / aspṛkṣat.
• 1423 viccha: āya by gupūdhūpavicchi… (3.1.28, SK 2303), optional in the ārdhadhātuka.
• vicchāyati; vicchāyāṃcakāra / viviccha.
• TB records whether śa or śap follows the āya-final, undecided.
• 1424 viśa: viśati, veṣṭā.
• 1425 mṛśa: amrākṣīt / amārkṣīt / amṛkṣat.
• 1426 ṇuda is re-read for parasmaipada even when the fruit goes to the agent (BM's "tanādi" read tudādi).
• 1427 ṣadḷ: sīdati as in bhvādi.
• The reading here gives sīdatī / sīdantī (śatṛ, ṅīp, optional num by ācchīnadyoḥ). The mūla emendation of "sīdati" is for Ajai.
• The jvalādi reading is for ṇa (sādaḥ). The double reading also serves accent.
• 1428 śadḷ: re-read for accent only. It has no śatṛ, since śadeḥ śitaḥ gives ātmanepada.
• Six svaritet follow.
• 1429 mila saṅgame: re-read for taṅ (milati, milate; mimela, mimile).
• 1430 mucḷ mokṣaṇe, which leads into 2542.

H. muc and the rest: num before śa — running example: muc + śa + ti → mu-n-c-a-ti → muñcati

• 2542 śe mucādīnām 7.1.59 — before śa, num is added to muc and the roots that follow it.
• num comes from idito num dhātoḥ (SK 2262). BM (Wikisource only) calls the gloss śeṣapūraṇa.
• muñcati, muñcate; moktā; mucyāt.
• mukṣīṣṭa (kit by SK 2300).
• luṅ amucat (aṅ, the root being ḷ-marked, SK 2343); amukta, amukṣātām (sic lost by jhalo jhali).
• 1431 lupḷ: lumpati, lumpate; alupat, alupta.
• 1432 vidḷ lābhe:
• vindati, vindate; viveda, vivide;
• seṭ for Vyāghrabhūti (veditā), aniṭ for the Bhāṣya (vettā);
• parivettā: pari means passing over, so "one who marries before his elder brother" (tṛn or tṛc).
• 1433 lipa upadehe:
• limpati, leptā;
• aṅ by lipisicihvaś ca (SK 2418), optional in taṅ (SK 2419): alipat; alipata / alipta.
• 1434 ṣica kṣaraṇe:
• siñcati; asicat; asicata / asikta;
• abhiṣiñcati: ṣ by upasargāt sunoti… (8.3.65, SK 2270);
• abhyaṣiñcat: ṣ across aṭ by prāk sitād aḍvyavāye 'pi (8.3.63, SK 2276);
• abhiṣiṣeca: by sthādiṣv abhyāsena cābhyāsasya (8.3.64, SK 2277).
• TB dismisses the reading ṣicir as purposeless, since the aṅ of 3.1.53 is an aprāptavidhi.
• Three parasmaipadin follow.
• 1435 kṛtī chedane: kṛntati, cakarta, kartitā, kartiṣyati / kartsyati (SK 2506), akartīt.
• 1436 khida parighāte: khindati, cikheda, khettā. The root is also read in divādi and rudhādi in the sense of dejection.
• 1437 piśa avayave ("in an action on parts"; BM): piṃśati, peśitā. In the sense of lighting up: "tvaṣṭā rūpāṇi piṃśatu" (TB: dīpayatu).
• "vṛt": the mucādi end, and with them the tudādi.
• Close: "iti tudādayaḥ" (Here end the tudādi roots). It is supplied from the print; TB closes alike, and Wikisource's BM closes the śa-vikaraṇa section. p50 opens with rudhādi, rudhādibhyaḥ śnam (3.1.78, SK 2543).