^{page 231 note 1} The abbreviations indicate class and district—NA, Nagar Brahmin, Ahmedabad; NSK, Nagar Brahmin, South Kathiawar; NCK, Nagar Brahmin, Central Kathiawar; TA, Teli, Ahmedabad; Ch.Kh. Christian, Kheṛa.

^{page 232 note 1} See See Divatia, N.B., Gujarati language and literature, University of Bombay, 1921, especially p. 156.Google Scholar He notices opener ę in the 3rd person singular of the present and future, and also in the instrumental case, and the opener in the nominative singular, masculine terminations of nouns and adjectives, and in the imperative 3rd person singular and 2nd person plural. He also notices opener e and o with nasalization and when followed by nasals: ‘sign’, (a sort of vegetable), ‘a bat for games’, ‘corn from cob’, ‘to have a man locked up’, ‘a word’, cęn ‘signs’, sęn ‘army’), or nęn ‘eye’, ‘a milk pot’, ‘a show’, jęm, tęm, kęm. The distinction of close and open vowel quality does not seem to be distinctive in nasalized syllables. Dr. Dave has made similar suggestions in a more developed thesis.

^{page 234 note 1} See See Allen, W.S., Phonetics in ancient India, Oxford University Press, 1953, 38, 48 Google Scholar et seq., 65 et seq.

^{page 234 note 2} See Sir Ralph Turner's article ‘The Sindhī recursives’, BSOS, III, 2, 1924, 301–15.Google Scholar

^{page 235 note 1} A Brahmin informant from South Kathiawar rejected my pronunciation of bədi ‘vice’, presumably with English phonation though accurate otherwise, on the ground that the final syllable sounded like -dhi. A slight creak would have been safer. The same informant rejected my pronunciation of udər ‘belly’, remarking that the d sounded like Hindi or Marathi. A fully voiced consonant with a minimum of breath appeared to satisfy him. There may be something akin to the Panjabi and Sindhi processes of phonation in certain dialects of Gujarati.

^{page 235 note 2} I wondered whether the pronunciation khəmbat with unreleased final had led to the English version Cambay.

^{page 237 note 1} On the other hand, the i both in cita ‘pyre’ and citta ‘cheetah’ is also comparatively short in duration.

^{page 237 note 2} The informant commonly pronounced the English word ‘somewhat’ in a similar way,

^{page 237 note 3} For example, ləkhto (present), ləkhtyo (past), ləkhvano (future), ləkhvo, so that some of the verbal forms could be described as -t forms, -y forms, v-forms—and also perhaps -el forms and-∫ forms. To these are to be added the -a forms (ləkha-). The form ləkhauelo is generally pronounced ləkhaylo. Other interesting prosodic processes can be studied in what are sometimes called the double causatives: kha-, kha-vaɖ, khəvɖa-.

^{page 238 note 1} See preceding note.

^{page 238 note 2} The phonetic transcriptions are meant to suggest the usefulness of the phonological description of such words in terms of y-prosodies.

^{page 238 note 3} Some attention to these problems has been given both by European and Gujarati scholars. See ‘Stress accent in modern Gujarātī’, by Master, Alfred, Journal of the Bombay Branch of the Royal Asiatic Society, N.S., I, 1925, 76–94,Google Scholar and also the work of Dr. Dave.

^{page 239 note 1} In the matter of the ictus or the incidence of whatever stress may be felt by the native speaker, a Nagar Brahmin of Ahmedabad often remarked on the differences he felt in the ‘gesturestroke’ in the two words khətta and khadɦa in the sentence ‘he had to eat his own words’. He felt the ‘stroke’ ended before the end of the first syllable khət- of khətta, and was also in the vowel of khá in khádɦa. Other speakers sometimes felt that whereas in breathy voice syllables the ‘stroke’ was in the vowel, in non-breathy syllables with initial plosive, it was in the initial consonant.

^{page 239 note 2} Syllabic prominence, for example, is much clearer in than in