^{page 91 note 1} Pelliot, P., “Les mots à H initiale, aujourd'hui amuie, dans le mongol des XIII^e et XIV^e siécles,” JA., avril-juin, 1925, pp. 193–263; cf. pp. 230–4.Google Scholar

^{page 91 note 2} Shirokogoroff, S. M., “Northern Tungus terms of orientation,” RO., torn IV (1926). Lwow, 1933, pp. 167–187.Google Scholar

^{page 91 note 3} Kotwicz, W., “Mongol terms of orientation,” RO., torn IV (1926). Lwow, 1933, pp. 188–9.Google Scholar

^{page 91 note 4} Kotwicz, W., “Sur les modes d'orientation en Asie Centrale,” RO., torn V. Lwow, 1933, pp. 68–91.Google Scholar

^{page 91 note 5} Cf. Ifa, F., Koryūkyū, Tōkyō, 1922, pp. 59–62Google Scholar; Andō, M., Kodai Kokugo no Kenkyū, Tokyo, 1924, pp. 128–9.Google Scholar

^{page 92 note 1} Matsuoka, S., Kogo Daijiten, Tōkyō, 1930, pp. 501, 1064–5.Google Scholar

^{page 92 note 2} Kogo Daijiten, op. cit., p. 90.Google Scholar

^{page 92 note 3} Kogo Daijiten, op. cit., pp. 9–10, 489.Google Scholar

^{page 93 note 1} Kogo Daijiten, op. cit., pp. 651–2Google Scholar; Matsuoka, S., Nihon Gengogaku, Tōkyō, 1933, p. 61.Google Scholar

^{page 93 note 5} Shimmura, I., Tōa Gogenshi, Tōkyō, 1930, p. 237.Google Scholar

^{page 94 note 1} Nikon Oengogaka, op. cit., p. 4.Google Scholar

^{page 94 note 2} Sur les modes d'orientation, op. cit., pp. 84–5.Google Scholar

^{page 95 note 1} Sur les modes d'orientation, op. cit., pp. 85–8.Google Scholar

^{page 95 note 2} Ramstedt, G. J., “Über die Zahlwörter der altaischen Sprachen,” JSFOu., xxiv, 1. Helsingissa, 1907, p. 5Google Scholar; cf. also Kotwicz, W., “Contributions aux études altaïques,” RO., torn VII (1929–1930). Lwow, 1931, pp. 159–160, 216–17.Google Scholar

^{page 96 note 1} B. Y. Vladimirtsov, 1929, pp. 151–7.

^{page 96 note 2} Matsuoka derives this word from mu- “body” + *na (suffix), stating that the earlier meaning of the word is “body” (Kogo Daijiten, op. cit., p. 1236).Google Scholar

^{page 96 note 3} Sur les modes d'orientation, op. eit., p. 87, n. 68Google Scholar; also Contributions aux études altaīques, op. cit., p. 161, n. 35.Google Scholar

^{page 96 note 4} Cf. Yoshizawa, Y., Kokugoshi Gaisetsu, Tōkyō, 1931, p. 18.Google Scholar

^{page 97 note 1} Cf. Sur hs modes d'orientation, op. cit., pp. 88–9.Google Scholar

^{page 97 note 2} Cf. Kiyono, K., “Minzokuron,” Kōkogaku Kōza, vol. x, Tōkyō, 1929, p. 31.Google Scholar

^{page 97 note 3} As an example of Turk. -ä- ∽ Jap. -a-, -i- may be cited kirn (stem ki < *kiyi-) “to put on, wear”; Turk. *käδ Orkh., Uig. käδ- “to put on, wear”, Koib. kes “id.,” Osm. giy- “id.,” Yak. kät- “id.,” etc. The words kata “shoulder” and kadzuku (stem *kad-uk) “to put on one's head” seem also to have come from the Turkish *käδ.

^{page 98 note 1} Kogo Daijiten, op. cit., p. 501.Google Scholar Matsuoka does not explain the function of the -no- in kinovu. For the preterite suffix -ki see my article entitled “An Analytical Study of the Conjugations of Japanese Verbs and Adjectives”, BSOS., vol. vi, part 3, 1931, pp. 657–8.Google Scholar

^{page 98 note 2} Matsuoka connects these words with koyu “to cross over” (Kogo Daijiten, op. cit., pp. 570, 596).Google Scholar

^{page 98 note 3} According to Matsuoka the ka- of this word is of the same origin as the Chinese tšia (< ka, “to add to, join”, “good, admirable”, “good, beautiful”), whilst -tsu is a suffix (Kogo Daijiten, op. cit., p. 424). Thus he identifies this word with katsu “to add to” (pp. 432–3). However, the word katsu (stem kate) “to add to”, together with kata “side”, katsu “in addition”, etc., seems to be related to Osm. etc. qat “side, layer, -fold”, qat- “to add, join, mix”; Chuv., xut “side, layer, a time”, XUDÏŠ- “to be mixed”; Yak. Kïtar- “to unite, join”, etc., all derived from *qat.Google Scholar

^{page 99 note 1} Kogo Daijiten, op. cit., pp. 1048, 961.Google Scholar

^{page 99 note 2} Cf. my article, op. cit., pp. 642, 646.Google Scholar

^{page 100 note 1} Kotwicz, , Mongol terms of orientation, op. cit., p. 188.Google Scholar

^{page 100 note 2} Although Andō believes (Kodai Kokugo, op. cit., pp. 202–4)Google Scholar that homu and hogu have been derived from ho (< *fo) by the addition of the “formative suffixes” *-mu and *-gu, they, together with hafuri, may provide an example of the alternation *fom ∽ *fog ∽ *fav as here conjectured. M. Ueda quotes two current explanations of the word hafuri: (1) hafuru “to exorcize (evils)”, (2) hafuri (∽ haberi) “to attend on (gods)” (Dainihon Kokugo Jiten, vol. iv, Tōkyō, 1929, p. 334).Google Scholar

^{page 101 note 1} Yoshizawa thinks (Kolcugoshi, op. cit., p. 55)Google Scholar that the character came to represent m towards the end of the Kara period and denies the existence of the syllabic η in the Japanese language. Cf., however, Andō, , Kodai Kokvgo, op. cit., pp. 146–162Google Scholar; Matsuoka, , Nihon Gengogaku, op. cit., p. 289.Google Scholar

^{page 101 note 1} Cf. Matsuoka, , Kogo Daijiten, op. cit., pp. 652, 833.Google Scholar