^{page 666 note 1} For Ancient Japanese sounds see my little book entitled The Phonetic System of Acient Japanese, 1934Google Scholar.

^{page 666 note 2} Rahder (p. 180) is undoubtedly right in regarding akmafω “price, trade “as a cognate of aganaFω.

^{page 667 note 1} Various inerpretations have been put forward concerning the derivation of the word agata. Among others may be quoted: (Mabuchi) akati-ta “allotted field”, (Norinaga) aga-ta “raised field” and (Matsuoka) aga-ta “Our Imperial) field”. K. Shiratori interprets the word as meaning “fortress, city”, identifying it with Ainu kotan “island, border”, Mongol qota(n) “fortress, citadel, castle, town”, Manchu hoton “citadel, fortress” and Korean kot “a place”, but without explaining the function of a–, which he simply calls a “prefix”(cf. TN, pp. 119–0). Of all these interpretations Norinaga's is most accurate, for not only does this great philologist explain the precise meaning of the word agata, but its stem aga is in all probability related to agω “raise”(cf. MZ, pp. 1529–33) For Turkish aq- “rise” and its Mongol counterparts, see VS, pp. 267, 278.

^{page 667 note 2} Cf. TI, pp. 167–8; VA, p. 97.

^{page 667 note 3} Cf. AK, pp. 196, 217–18; TN, p. 175.

^{page 667 note 4} Cf. TJ, pp. 67–72; RA, pp. 473–5; RV, p. 115.

^{page 667 note 5} Cf. TN, p. 175.

^{page 667 note 6} SN, pp. 5, 28.

^{page 667 note 7} SE, pp. 1–2.

^{page 668 note 1} BGT, pp. 4–5.

^{page 668 note 2} Mention may here be made of Old-Turkish ïraq, Tel. ïrak, Yak. ïrāx “far, distant”, etc., which, according to W. Bang, are related not to ïrq but to Uig. jïraq, Tar.jiraq, Osm. iraq, Shor, Koib., etc., raq “far, distant” (cf. BGT, p. 5, footnote 1). He is of the opinion that these words have been derived from *jïr which has developed from, or is identical with, ar (BA, pp. 20–22). It is, however, plain that ar, arï “on the other side, thither “are built upon *a (cf. DG, p. 616) which is also found in Old-Osmanli an-, the stem of all the oblique cases of the personal pronoun of the 3rd person (cf. DG, pp. 201–2), with the directional suffix -rï > -r (cf. ZS, p. 47). If, therefore, *jïr were really derived from, or identical with, ar (< arï), then its final -r must also be a directional suffix. That such is probably not the case can be seen from Old-Turkish jïrgaru “to the left, to the north”, where -ru is a directional suffix (cf. ZS, pp. 48–9). Thus it is less difficult to assume an alternance between *jïr and *ïr than to suppose that in jïrgaru the older function of -r had been forgotten and the suffix -ru was therefore added.

^{page 668 note 3} Vladimirtsov, B. Y. derives iruwa from *ïrŭωa (DAN, pp. 294–5Google Scholar; VS, pp. 181, 204, 302).

^{page 668 note 4} The stem ωra “inner” of ωramω. “resent”, etc., seems entirely of different origin. In all probability it is related to Mongol (written) yry (or, according to Ramstedt, œry) “the interior, the inside”, which Z. Gombocz connects with Osm., etc., osz “the best part of a thing, the inside, the heart”, etc. (GL, p. 4).

^{page 669 note 1} Cf. TN, pp. 130–2. C. Haguenauer suspects a possible relationship between imω, and “Hades”, “darkness”, etc. (HC, p. 181), but I rather think these words with back stem vowels have been derived from a root with a dental initial.

^{page 669 note 2} “Vorschlag” (prothese, prothèse) is the development of a non-etymological element at the beginning of a word. If often occurs in both Turkish and Mongol, and is also found in early Chinese loan-words contained in Japanese, as e.g. Arch. chin. mp) “horse”, “plum”. Rahder (p. 181) prefers to consider to be related to Indonesian ivak.

^{page 669 note 3} Cf. TI, p. 27, note 3; p. 144. K. Shiratori cites joh-t'ung “fortunate” without disclosing its source (SH, p. 46).

^{page 669 note 4} Possibly also isa, iza (in the names of gods and places), isasa “sacred”, isωzω (the name of a river), “courtsey” “revere, honour”.

^{page 669 note 5} Cf. GL, p. 36. G. J. Ramstedt seems to consider this *δ to have been a palatalized r (cf. RF, pp. 26–32), while N. Poppe appears to regard it as a fricative r (cf. PT, pp. 78–9).

^{page 670 note 1} Cf. TI, p. 27, note 3; GF, § 75.

^{page 670 note 2} GF, § 101.

^{page 670 note 3} GF, § 123; according to Pedersen *ïët > ḯt (PL, p. 556).

^{page 670 note 4} Cf. PT, p. 71; also RF, pp. 12, 14–15. A similar change as this may also have taken place in Japanese, thus “send” “drive away”.

^{page 670 note 5} RF, pp. 14–15.

^{page 670 note 6} GF, § 101, § 142.

^{page 670 note 7} PL, p. 559.

^{page 671 note 1} Cf. MV, p. 63.

^{page 671 note 2} JO, p. 124.

^{page 671 note 3} Cf. JO, p. 124; BA, p. 28.

^{page 671 note 4} JO, pp. 124–5. The assibilated i as spoken of by Jarring is, according to G. Raquette, heard in the following words: it “dog”, pit “louse”, il, jil “year”, kī “soiled linen”, kīm “garment”, 'iki “two”, and sometime also in bir “one”, kim “who”, and ki'iāb “book” (RG, p. 117).

^{page 671 note 5} SB, pp. 183–7.

^{page 672 note 1} Cf. PT, p. 69.

^{page 672 note 2} Schriefl derives this word thus: bir > *b₃r > b₃ (SB, p. 185).

^{page 672 note 3} Cf. TS, pp. 54–5.

^{page 672 note 4} HL, pp. 91, 94.

^{page 672 note 5} Cf. BN, pp. xviii, 105.

^{page 672 note 6} Both Ramstedt and Poppe derive the Chuvash forms from *ïta (cf. RF, p. 12;PT, p. 71).

^{page 672 note 7} Does the word jëtf ^I in jëtti “a powerfully built dog” (where means “a hound”) preserve an older form of ? How are we to explain the relationship between Chuv. and the Old-Bulgarian word etx (or jetx) “dog”? (Cf. JSFOu., xxx, pp. 14–15). These problems must be reserved for future investigation.

^{page 673 note 1} Cf. KE, p. 165.

^{page 673 note 2} These may have developed from ikki and εkki.

^{page 673 note 3} Cf. KE, p. 164.

^{page 673 note 4} Cf. GF, § 107; RL, pp. 71–2.

^{page 673 note 5} Cf. PL, p. 558.

^{page 673 note 6} K. Shiratori cites from the Liao-shih the passage “niche-hê is a dog” (SH, p. 46).

^{page 673 note 7} These Mongol words are compared by Ramstedt with Uralian loga, loka, lokaā, lokka, lai, lei, etc., “fox” (cf. SR, p. 118), while Vladimircov quotes in comparison Manchu iuqa “young lynx, Russian puppy” and Samoyed loka, etc., “fox” (VS, p. 350).

^{page 673 note 8} P. Schmidt considers these Tungusic words as loans from the Palaco-Asiatics and draws attention to Japanese inω “dog” (SN, pp. 5, 16). He does not tell us, however, from what language they have been borrowed. It would probably not have been from Gilyak, for in this language the word for “dog” is kan, which Ramstedt compares with Turkish qant∫uk “bitch” (where -t∫uq is apparently a diminutive suffix) and Korean kε “dog”, written kai < *kaṅi (RK, § 23).

^{page 674 note 1} The initial vowel does not fall off in Khalkha, because in this dialect the first syllable is always stressed (cf. RU, p. 50). But in the Monguor dialect spoken in Kansu the falling off of the initial vowel is quite common, as e.g. dur (< Mong. Written edør) “day”, ndë (< Mong. written ende) “here”, mugæn (<Mong. written emegen) “old woman” (cf. MSM., p. 807).

^{page 674 note 2} That the Old-Turkish may correspond to Mongol n can be seen from Old-Turkish “sheep” = Mongol (written) qonin, Kalmuk (< *xoi) “sheep”.

^{page 674 note 3} The diminutive suffix -qai is found in Khalkha thāx^aflc (< *tai-qai > *tei-xa > Chuvash tixa) “foal” = Turkish taj, } “foal” (cf. RF, p. 24; VS, p. 293). It also occurs in the form -kai ∼ -kεi in Uryankhai, Kazak-Kirghiz, Kazan, etc. (cf. KO, pp. 163–4; NT, p. 570).

^{page 674 note 4} These must be related to the diminutive suffixes -kān ∼ -kōn in Barguzinsk Tungus (cf. PM, p. 6), -kan ∼ -ken ∼ -kon in Manchu (cf. ZG, pp. 81–2), -qan ∼ -ken in Mongol written (cf. PN, p. 114) and -qïna ∼ -xïna ∼ -gïna ∼ -kïna in Turkish (cf. ZS, p. 20). A diminutive (hence augmentative) suffix of the same origin is also found in the Japanese word kari-gane “wild goose”.

^{page 674 note 5} Poppe, like Ramstedt, believes that the Chuvash forms and preserve the original final vowel which has been lost in all other dialects (PT, p. 72). Their belief seems to be based on the assumed change *ïta (> Buryat ite “voracious”) > which, as we have seen, is probably not what actually happened.

^{page 674 note 6} A comparison of Japanese inω, Tungus inakin and Manchu indahun has already been made by H. Winkler (WU, p. 279).

^{page 675 note 1} WU, pp. 261, 264, 266.

^{page 675 note 2} PS, p. 181.

^{page 675 note 3} WU, pp. 261, 264, 265, 267.

^{page 675 note 4} PS, p. 179.

^{page 675 note 5} DW, p. 99.

^{page 675 note 6} PB, pp. 36–7.

^{page 675 note 7} SR, pp. 118–19.

^{page 675 note 8} Cf. PT, p. 71; VS, p. 145.

^{page 675 note 9} These three sets of words may well be compared with Ancient Japanese ţi ‘blood”, kiri “mist, haze” and siro “white”, respectively.