^{page 218 note 1} Or “with love her torch”.

^{page 218 note 2} For the translation see text a few lines below.

^{page 220 note 1} calaiṇa “foot” = caraṇa is found in Pkt. (Pisoh. PG 257).

^{page 220 note 2} It is interesting to find that the commentator Ratneśvara had consulted the commentary of Namisādhu on Rudraṭa, for he notes on the passage the fragment anangasyeyamānangī taken from Nami.'s tathânangasyeyamānangī dīpika/tayā kāmasya prakāśitatvāt.

^{page 221 note 1} I regret that I do not know the name of the author of this brilliant little piece of research and apologize to him or her for my ignorance.

^{page 222 note 1} Using ayukta like Var. and Hem., not ayuja (ps.-Var.) or asaṃyogyasya (Caṇḍa).

^{page 222 note 2} Cf. Caṇḍa, ii, 22, upamāne piva iva (? viya) viva vva va jahā rataḥ — kamalaṃ viva tujja (B tumba) muhaṃ — and Hem., ii, 182, miva piva viva wa va via ivârthe vā—kamalaṃ via—.

^{page 223 note 1} Cf. also Mid. Kan, Nāki = Naga (above); Aśoka, NW, Kamboca = Kamboja; Old Sinh. Naka = Nāga (Geiger, Gram, Sinh., 1938, § 41), and other proper nouns.

^{page 224 note 1} The Hāthigumphā inscription has padhamam, Pāli paṭhama.

^{page 224 note 2} But the single ñ is found in inscriptions.

^{page 226 note 1} Aśoka's Gīrnār inscription has the equivalent kacam < kṛtya.

^{page 226 note 2} These are Iws direct from Skt. The dental is found breathed in Tel. only. In Tam. it is voiced and a fricative, but can have been regarded as a breathed stop, just as Kumārila Bhaṭṭa (seventh century) reproduces Tam. atar “road”, as atar and not adar, adhar or athar, which he might have used.

^{page 226 note 3} Hem.'s error of reading nādiyujyor as nādi and the root yuj is rejected here.

^{page 227 note 1} tumhätiso, the reading of one MS., is to be preferred to umhātiso or yumhātiso. Pāli has tumhādiso, etc.

^{page 227 note 2} But the Hathigumpha inscription c. second century b.c. has gananā = gaṇanā and pāpunati = Pāli pāpuṇati (Woolner, , Intro, to Pkt., p. 192)Google Scholar.

^{page 227 note 3} Taye and tāye are Aśokan forms (All. Qu).

^{page 228 note 1} Cf. i, 100, ii, 60, in which he mentions Kaśmīra.

^{page 228 note 2} mutinga, of. Drav. *murrukai, Kan. muttige, Tarn, murrukai “cover”, Tel. muta “parchment of drum”. Is Skt. mṛdanga “then”, a product of popular etymology = clay-body, alluding to the old earthenware pot drum ?

^{page 229 note 1} Pāli has other instances of a r/1 alternation, e.g. cattārīsa and cattālīsa = catvāriṃśat, possibly via cattaḷīsa, cf. parigha, paḷigha, paligha.

^{page 229 note 2} Grammar of Sinhalese Language, 1938.

^{page 229 note 3} Also in Hem.'s Kumārapāla-naritra exemplifying CP. (viii, 13).

^{page 230 note 1} See also Grierson, , EPG., p. 119Google Scholar, where he mentions Barth's identification of the Sthaviras' Paiśācī with Pāli.

^{page 232 note 1} i.e. n'ādi-yujyor for ayujor-anādyor.

^{page 232 note 2} To illustrate the errors in which it was possible to fall, I give here two examples from Rāmaśarman.

(i) Stanza 10 (EPG., 128, 138). “In the case of abuse the termination of a noun, whose base is a, is ī in Paiśācika Śaurasena, as śiālī (= syāla ‘brother-in-law’).” The source of this is ps.-Var. xi, 17, Māgadhī—śiāla— from śṛāla “jackal”. Mārkaṇḍeya, xii, 12, has, correctly, śiāle as a Māgadhī word.

(ii) Stanza 20 (id. 131, 141), where the error is due to a defective text. jaai matam āruhanti giri-taṇayā paṇai-kappalaā

which Grierson translates “Victory to Pārvatī, daughter of the Himālayas, who ascends upon pride to destroy it, but who is a Wishing-tree to him, who bends low before her”.

This is taken from an author unknown to Rāma. It occurs in Saras, ii, 17, 9, but tamalliantī is read for matam āruhantī. The passage runs—

vaktṛ-viṣay' aucityādi-prayojyā miśrā yathā

jayati janatâbhivanchita-phala-pradaḥ kalpapādapo giriśaḥ

jaai a tamalliantī girī-tanayā paṇai-kappalaā

evaṃ bhāṣāntarāṇam api miśrī-bhāvo dṛṣṭavyaḥ

“Miśra (mixture of lines), which is to be used for suiting the meaning to the context and so forth, is, for example—

“‘Victorious is the Tree of Wishes, mountain-born, which yields the fruit desired by the world. Victorious to her worshippers is the Liana of Wishes, who tops the Tamāli tree, the Mountain-born (Pārvatī).’

“Thus should be regarded the trope Miśra in connection with different languages (or—translations).”

Bhojadeva's second line is clearly Prakrit (n in tanayā is irregular, but is sporadically found for n, PG224). Rāmaśarman's reading is inferior, and, as Grierson points out, matam (in the suspected passage) is the only Pai. feature.

L. Nitti-Dolci (GP 124) explains why it is unnecessary to trouble with the “mixed Paiśācis” of Rāmaśarman. He probably invented them, as she thinks, and used them to account for anomalies of language, which were, in fact, due to the inferiority of his texts.