page 104 note 1 Census of India (1931), vol. xiv, Madras, part ii, table xv (pp. 294–5)Google Scholar.

page 104 note 2 , Fleet, Gupta Inscriptions (1888). C.I.I., iii, p. 79Google Scholar.

page 104 note 3 Only 664 persons (in 1931) returned Telugu as their subsidiary language. The Madura District Gazetteer (1906, p. 110), however, states that most of the Patnūlkārans can still speak Telugu, and in an illustration of marriage decorations reproduced by Thurston the gōtra names are in Telugu script. As is stated below, Vēṅkaṭa Sāri wrote in all three languages.

page 105 note 1 LSI., ix, 2 (1908), p. 336Google Scholar, lists among authorities on Gujarati: “Rama Rao, T. M., Saurāshṭra Bodhan. Madras, 1900.” I have failed to trace publication, and suppose that 1900 may be a mistake for 1906, the reference being to the Saurāṣṭra-bōdhini referred to above. The book is listed among authorities on Gujarati because specimens of Paṭṇūlī received from Bombay appeared to be ordinary Gujarati. No specimens were received of Madras Paṭṇūlī. Two records of the language were included in the collection of Madras Gramophone Records of which transcriptions were published in 1927; but the transliteration makes the specimens difficult to use. I have failed to follow the spoken language of the records themselves.

page 106 note 1 Postscript. After this article had been written the Saurashtra Literary Society of Madura, through the good offices of Saurāṣṭra Viprabandhu K. V. Padmanābha Ayyar (above mentioned) presented several books to the India Office Library, including a copy of the only (1905) edition of Vēṅkaṭa Sūri's Saṃgīta Rāmāyaṇu, as well as Mr. Padmanābha Ayyar's life of the poet, in Tamil (Madura, 1942; with a brief sketch of Saurāshtran literary activity to date).—Vēṇkaṭa Sūri was the son of Nārāyaṇa Ṥarmā, a purōhit of Rāmacandrapuri near Ayyampēṭai in the Tanjore district, and was born there in the Bahudhānya year (= A.D. 1818–19; but the Kaliyug equivalent is given as 4919). At the age of twenty-five he went to his sister's house in Tanjore and taught Tamil and Telugu songs: but he also composed and taught some Saurāshtran songs. He received the patronage of Ṥivājī Mahārāj, the last prince (1832–1855) of the Tanjore ruling family, and was given a post in the Sarasvati Mahal (the palace, or the library). When the I Tanjore State ceased to exist he went to Madura and continued to teach “songs with music” (saṃgitun bhajana sara) for twenty years continuously. Then he made an extensive pilgrimage in South India, and on his return to Madura occupied himself in literary work (apiān kavita herē Kṛṣṇāḷi Rāmāḷi granthun sara Saṃgīta-Rāmāyaṇu kavita kerli prasaṃgu kerlētu dinnu gaṃḍes. “Along with his poetical works K. and B. he spent his days in writing and editing his S.R.).” The work might then be dated about 1878. Shortly afterwards he set out on a pilgrimage with his son to Benares (where he left the boy to be educated), and to “Nēpālādi divya dēsī”. In the course of this pilgrimage he died at the age of seventy-two, at Srīraṅgakṣētra on the Jumna. (Since I cannot read Tamil I have followed the account given in the introduction to the Saṃgita-Rāmāyaṇu. I believe it is in general agreement with Mr. Padmanābha Ayyar's Tamil account. I very much regret that I received a copy of the Saṃgita-Rāmāyaṇu too late for the purpose of the present article.)

page 107 note 1 He says that y between vowels has the full sound in nayana, sayana, and in the inserted y in such a case as gōy avai—gō avai. But v or y is laghu in such cases as kasuvurai (kasu + urai, the spring flows), avyāsi (they came). It is laghu in such plurals as menkyānu (from meniku, man). Final -o in kōno (what) is laghu. Rāma Rao occasionally writes, in such verb forms as avāsi (for avyāsi), a lāghava mark over -ā-. If Rāma Rao's analysis of kasuvurai were correct it would be necessary to suppose that the v, in this case, had attached itself permanently to (v)ur, flow, since vur- seems to be the established usage. The insertion of y for euphony, as in Telugu, is common. I have not found v inserted in this way; but it would be inevitable.

page 107 note 2 e.g. pani, water, akāsu (ākāṥa), rajo (rājā). Sometimes there is a doubling of the following consonant, as in ruppo, silver, sonno, gold.

page 111 note 1 Some cases are doubtful. It seems natural that a word collocated with another in the manner of a compound should not have the “nominative” inflection. Thus kaḷa mēgu (mēghu), black cloud, which is as frequent as kaḷo mēgu, may sometimes be felt to form a single compound word. However, in such a case as vatta kerosu there is no doubt that vatta alternates with vatto as the nominative (here accusative) case. And it would seem that the language does not object to compounds with an inflected first element, like mhaḷin-pogul “fishes-feet” (given as an example of the incredible, like “sky-flowers”). Mhaḷin is the plural of mhaḷi.

page 112 note 1 The y has the lāghava mark. In one Tamil print the plural is invariably menkēn (or meṇken).

page 113 note 1 The y has the lāghava mark.

page 113 note 2 aiḍe is a “ṥiṣṭa-grāmya-ṥabdu” for ariḍe. Similarly vaiḍe for variḍe. Compare laiḍo for lagḍo (stick), and laina for laguna (negative verb = it is not fixed).

page 114 note 1 Postpositions are normally added to the possessive stem: Morlenta, morahāli, morsara, mortokān, mossommaru; amralenta or auralenta, etc. (but amtokan). Toralenta, tortohan, tossommaru; turalenta or tumralenta, turtokan, etc. (but tumsommaru).

page 116 note 1 e.g. lovunum askāku sonno bheḷi bhārvasti among metals gold is the heaviest of all. (I cannot explain asti except as a Sanskrit loan.)

page 118 note 1 poḍuashas the termination of the first person. As Mr. Alfred Master has suggested to me, and as the evidence proves, Saurāshtran has confused the Indo-Aryan root paḍd- (proper to the karmaṇi prayōga construction) with a Dravidian root meaning “suffer”. Mĭ avni poḍusu is really a case of kartari prayōga— I suffer coming. Rāma Rao does less than justice to Saurāshtran in saying that its passive prayōga is confined to this cumbersome periphrasis with paḍ-. See next note.

page 118 note 2 These passives will be dealt with in a continuation of the present article.

page 118 note 3 See preceding note. The stem is no doubt passive in form.

page 118 note 4 It will be noticed that he chooses a causal stem to exemplify transitive action. But this must be accidental. Causatives in -da will be dealt with in the continuation.

page 119 note 1 Mr. Alfred Master pointed out to me that -su, -si “terminations” are really the auxiliary similarly added by other languages in such forms as āvũ chũ, I am coming: which is identical with Sau. avusu.

page 119 note 2 In Saurāshtran insentient plurals require a verb in the singular.

page 119 note 3 Rāma Rao observes later that roots in -ā (sā to see, dā to give, jā to go) are “just like rhāā but that siyesi, jiyesi—he might have added diyesi—have the sense of the past. Usage indicates that the latter part of the statement is true not only in these cases, but in the case of all verbs except the auxiliary rhiyesi: e.g. keresi, did; menesi, said; pusesi, asked; poḍesi, fell.

page 120 note 1 An absolutive hoti has not been noted in usage. Hoyi is always used.

page 120 note 2 In such examples as menāyi, dekkuāyi, the termination -vāy(i) is added to the unmodified stem. But rhavāyi looks like a passive or bhāvī prayōga formation. The negative form ends in -vāna.

page 120 note 3 rhavātte (and rhavānātte) connect with bhāvi prayōga forms of rhā.

page 121 note 1 This is the present participle with the dative inflection of a noun. But as will be shown later -tte verbals are usually pronominalized.

page 121 note 2 I have met with no example and cannot assign the function of this form. Hoyyoko may be comparable:—doḍo tē Manmantu-bāṇulu hoyyēti, tē kaḍo hoyeti moro monnu, bheḍi sompu hoyyoko(Aṣṭ., x, 19.4. If that eye be Love's arrows, though this my mind be dark it may well become fair [?].)