^{page 269 note 1} When reconstructing archaic Chinese words in my “Word equations” I printed the vowels in italics in order to indicate that I did not wish to suggest anything about their archaic sounds.

^{page 270 note 1} In the page references given by K. in his W.F. there is a typographical error. The figure 8 must be added in order to make them correct.

^{page 270 note 2} At an earlier occasion Li Fang-kuei had already assented to my reconstruction of dentals (Ch'ieh Yüln â ti lai-yüan, p. 18). As for some dentals newly reconstructed by me after -a and -ie see below, p. 280, etc.

^{page 271 note 1} To this supposition of a considerable number of “hedge rhymes” in the Shi Ching Li Fang-kuei had objected (Ch'ieh Yüln a ti lai-yüan, p. 26), pointing to the strictness of the Shih Ching rhymes.

^{page 271 note 2} Cf. for instance, loc. cit., p. 132, n. 1, where he says of the words ,, , below referred back to an original final dental by me: “None of these finals have any connection, in script or rhymes, with ‘ju sheng’ words, see JRAS., 1928, p. 783. It is clear beyond doubt that they have never had any final consonant.”

^{page 272 note 1} I cannot discuss here the similar conditions of finals in Burmese. For the phonetic side cf. Latitzeichen und ihre Anwendung in verschiedenen Sprachgebieten (Berlin, 1928), p. 108, and n. 3.Google Scholar

^{page 272 note 2} Cf. below, p. 275.

^{page 273 note 1} My assumption of an uvular pronunciation of ancient Tibetan final r (v. Worlgleichungen, p. 28, and n. 1) has been assented to by A. Dragunov (Or. Litztg., 1931, col. 1087 and n. 3), who adduced in confirmation his own observation in modern Tibetan dialects. Cf. also ibid. Dragunov's very interesting remarks on a final d resembling an l, observed by him in Chinese and Tibetan dialects.

^{page 273 note 2} The appearance of -r for -t in certain documents of T'ang time, transcribing Chinese words-for instance, to quote the most recent publication, in the fragment of the Vajracchedika, edited by Professor F. W. Thomas (“A Buddhist Chinese Text in Brāhmī Script,” Zeitschr. d. Deutsch. Morgenl. Ges., vol. 91, pp. 1–48)—points to a dental fricative in the Chinese dialect transcribed, but not necessarily to an -r. (Cf. the transcription δ chosen by Professor P. Pelliot, but also his remarks in T'oung Pao, vol. 16 (1915), p. 3, n. 2.) It is very interesting to note that the Vajracchedika Text shows also a final guttural fricative (transcribed as h) for -k.

^{page 273 note 3} Cf. a number of similar examples in Jäschke, Tibetan Grammar (Berlin, 1929), “Addenda,” pp. 120–1.Google Scholar

^{page 274 note 1} Cf. also Mă šit (Ling. Survey of India, vol. i, part 2, p. 247, No. 263).

^{page 274 note 2} Cf. above, n. 1, and for example Măru pyit “to give”, belonging to Chin. (W.F., p. 37, reconstructed as ). I hope to come back to this point at another occasion.

^{page 275 note 1} In his Dictionary (p. 29) K. assumed a difference between an ordinary g vanished without a trace in opposition to a palatal g vocalized into i and a velar g vocalized into u. He has, however, never come back to this differentiation. If a consonant vanishes after i and u, this may mostly not be the case of a vanishing without a trace, but of a contraction after a previous vocalization.

^{page 276 note 1} Alternations between and the endings occur besides in the following series which are not connected with the rhyme words contained in Table XV: 48 49 (cf. below, p. 278), 603 , 819 .

^{page 277 note 1} On account of the guttural ending also “to wash” may now perhaps be added to W.F., p. 69, B. 235, etc.

^{page 277 note 2} I am indebted to Professor W. Percival Yetts for having incidentally drawn my attention to these words. Cf. his paper, “A Datable Pair of Chinese Bronzes,” Burlington Magazine, 1937, p. 8 and n. 3.

^{page 278 note 1} The Kuang Yün even indicates the reading b'u∂k besides b'uo for .

^{page 278 note 2} The exact title is: Hsü Tzŭ Hui Pu . The work (in 14 chüan) was compiled by Wu Jên-ch'ên (No- 2326 in Giles's Biogr. Diet.).

^{page 278 note 3} Professor G. Haloun kindly indicated to me some ju-shêng'a of this series: .

^{page 279 note 1} A. Dragunov (Or. Litztg., 1931, col. 1086, n. 1) points to the dialectal nasalized form occurring in Swatow.

^{page 279 note 2} The rhyme (Series 1244), adduced by me, is, however, to be eliminated as it belongs to the group dealt with above. In the case of K. reconstructs in table xiii (No. 58) an original , although it had an original dental initial. occurs in table xiii (No. 59) with the ending -io in spite of a reconstructed dental initial (for which K fails to give evidence), but appears in table xv (No. 16) with the ending and without any initial.

^{page 279 note 3} To this, I think, the rhyme 614 of table xiii should be added.

^{page 280 note 1} In the case of the endings -uo () some evidence for an original guttural may further be found in the following etymological connections: “street” related to “id.” (W.F., p. 60, M. 24: kêg); “to call” related to and yáu < -g “id.”; “let go” related to “id.”; “to snore” related to “id.” The latter word had in my opinion an original final guttural (see above, p. 277, series 484, and Tib.-Chin. Wortgl., No. 325).

^{page 280 note 2} The assumption of an original vocalic final of these endings is also contradicted by modern dialect forms in -ai (), esp. in the Min Dialects, as Fuchow muai , , nguai , tuai , etc. (cf. Karlgren, Phonol. Chin., pp. 714 seq.). I must thank A. Dragunov for having drawn my attention to this dialectal vocalization of finals generally vanished without a trace (Or. Litztg., 1931, col. 1086, n. 1).

^{page 281 note 1} In accordance with this suggestion and with the inclusion of some other corrections a G or a D? is to be replaced by a D, or a D is to be added in the case of the following Series (Nos. of K.'s Dictionary) in the table annexed to Relconstr., pt. ii (p. 28): 1 70 , 72 , 93 , 117 , 132 , 196 , 204 , 223 , 337 , 342 , 343 , 410, 413 , 414 , 417 , 437 , 533 , 534 , 569 , 593 , 647 , 679 , 684 , 721 , 753 , 776 , 846 , 1006 , 1008 , 1009 , 1011 , 1099 , 1157 , 1313 . At the same time the D appended to Series 704 is to be corrected in G, and a D? is to be added to Series 1096 and 1110 .

^{page 282 note 1} On account of its dental final should be added to W.F., p. 94, Nos. I l l seq.

^{page 283 note 1} Cf. also t'iem “to lick” for which the double reading t'im, lira in Cantonese suggests an original initial t‘l.

^{page 283 note 2} Cf. also Tibetan “horse”.

^{page 284 note 1} Professor Haloun suggests an original labial final also for Series 97 , as he considers this a phonetic subseries of d'âp, which is not contained in K.'s Dictionary. He hopes soon to deal himself with the difficulty in principle arising from the assumed alternation of initials differing so much in the place of articulation.

^{page 284 note 2} Also the development ĺěg > ia > , assumed in W.F., p. 55, makes one hesitate and rather think that the assumption of an archaic ie is the “principal error”.

^{page 285 note 1} This indirectly confirms my reconstruction of an archaic for in Tib.-Chin. Wortgl., § 97, where the development < , is assumed in order to account for the initial m found in most of the Indo-Sinitic words for “fire” (cf. Tibetan me, Arakanese mēn:, etc.). K. now reconstructs as (W.F., p. 80).

^{page 286 note 1} K. in his Dictionary leaves the question undecided whether we have to deal here with a hui i o r a phonetic compound.

^{page 286 note 2} For reasons unknown to me K. reconstructs in his Dictionary and in his Études (p. 834, No. 72) instead of the forms or .

^{page 286 note 3} Evidence for an alternation between t- and n- may also be derived from the modern reading niao for ancient tieu in (Series 662). Cf. further niep and “to pinch”, of which the former occurs in the same series as (Series 670), the latter in the same series as niep, , and (Series 667). Therefore also and “id.” (Series 1057) may be etymologically related.

^{page 287 note 1} Professor Gustav Haloun has very kindly provided me with the following supplementary remarks, for which I should like to extend to him my heartfelt thanks:

“Serie 13 : wird aber alt, noch im Suo-, geschrieben (mit Nebenform ; phonetisch). und , gehören zur gleichen Unterserie, die Lesung hat aber auch ‘Kröte’, die Nebenlesung , sonst niei.

“Serie 651 mit dem Lesungen nân, tân: ist alte Nebenform zu , zu , Serie 946 (mehrere Lesungen).

“Serie 654 mit den Lesungen nuai, usw.: ist alte Nebenform zu t'uâi, Serie 1141.

“Serie 668: , s. u. Serie 928 u. 1193 (wechselt übrigens auch mit , Serie 1164).

“Serie 788 : selbst Unterserie von nan.

“Serie 832 . mit den Lesungen : enthält ,, alte Nebenform zu .

“Serie 928 : enthālt neben , t'âm die Doppellesungen t'âm, , , , , , 'iem, niem; beachte auch , mit alter Nebenform Serie 668.

“Serie 930: ist phonetisch in , Serie 1076, dieses in 4p , Serie 669, und , einer alten Form für , , das wiederum eine alte Schreibung für < , Serie 186, ist und mit eine Nebenform zu < , Serie 784, bildet; of. auch sien, Serie 797 (Determinativ ‘gehen’)

“Serie 936: ist phonetisch in , , phonet. in und vielleicht in , Serie 664, dieses in , Serie 1222.

“Serie 941: , nicht bloss phonetisch in und , sondern auch in (Serie 1130), , Serie 221, und samt Ableitungen (nicht im Analytic Dictionary).

“Serie 1193 , usw.: enthält , , “Lederhose’, cf. ‘dass.’, und die Doppellesung , für das Determinativ ‘weiches Leder’ (Suo- Nr. 91, Phonetikum ist tatsächlich , Serie 668, das auch etymologiseh zugehört, alte Form wohl mit Phonetikum , ‘dass’.

“Es bleibt anzuführen dass das Suo-, dem fast die meisten Verknüpfungen ohnedies entnommen sind, auch in ts'i phonetisch erklärt.”

^{page 288 note 1} More exactly a “supradental” fricative.

^{page 288 note 2} Cf. the reconstruction suggested by E. Chavannes and P. Pelliot in Journ. As., xviii (1911), p. 538, and K.'s remarks in his Études, pp. 461–2.

^{page 288 note 3} Cf. above, p. 283, n. 1. See some more examples in Worlgl., p. 43 and n. 2.