¹ I am indebted to Dr. A. E. Cowley and Dr. G. B. Gray, of Oxford, for remarks which I am glad to be able to use. The former has, however, some very revolutionary suggestions, which will be noticed at the proper place. My indebtedness to Mr. Buckler has been already mentioned (p. 82).

I should add that the ‘Louvre inscription’ (note 5, etc. ) is a Lydian text found by M. Bernard Haussoullier and shortly to be published by him and presented to the Louvre. He has very kindly allowed me to use a copy and photograph of it in preparing this paper.

² Hrozny's solution of Hittite (Mitt. Deut. Orient. Gesell. No. 56, Dec. 1915) is welcomed by Meyer, Ed. and by Bohl, (the latter in the Theol. Tijdschrift, Jan. 1916)Google Scholar. A brief and cautious summary is given by Vosté, in the Revue Biblique, 1917, pp. 315Google Scholarsqq. Among the identifications may be mentioned the Hittite ug(a)—Lat. ego; tuêl—tuus; kuiš, kuiški — quis, quid; kuvatka — quodque; danteš—dantes (plu. part.); ammug— ἐμοιγε āppa— ??? parā— παρα šipanti— απενδει

³ The above forms are of course highly schematic. Mr. Arkwright—as observed, p. 82 (above)—assigns to Littmann's ũ and ć, the values l and n respectively. Dr. Cowley, too, has other doubts.

⁴ The readings in L. lb, 11. 4–5 (for the references see the list above p. 82) on p. 42 and quoted on p. 13 are doubtful. The citation from L. 7, 1. 1 is inaccurate; read Hudānl Artimuũk. On p. 15, middle (the remarks on ũ), the words savũnt and akmũnt should presumably be savẽnt and akmũ or akmũit. Page 17, among the words where ć(?) occurs in the middle or beginning, references should have been given, fẽntamćić, for example, I cannot verify, unless it is fetamć idć, 7₁₈. The same applies to the words beginning with q (?) on p. 18, especially qashrlãć (? omit the s). qiśaũad (for ś read r), cf. 29₅. ₆. On p. 64 read dummüis for dummis (l.7 from foot) and apparently fẽllaũin (at foot). On p. 69, 1. 8 from foot, for 26 read 24. Page 84, third item, read kulumćak and 11₁₀ (for 10₁₀).

⁵ Cf. also the Louvre inscription forlũ xii oraũ … Aũiksantruũ aũmũuũ dᾶć.

⁶ No notice is taken of the sign in Sayce's Lydian inscription from Egypt (above, p. 77, n. 2).

⁷ For -valis, cf. katovalis (cited p. 84), which has some relation to katovaũ (16₂) and presumably to kotav (7₁₇). Katov- may be the Greek απενδει etc., see Buckler, W. H. and Robinson, D. M., American Journal of Archaeology, xvii., pp. 33Google Scholarseq. Katovalis occurs several times in 16, along the righthand side of which is a typical threat preceded by the obscure words ẽnaũt bakivalũ (see p. 65 seq.).

⁸ Amer. J. Arch. xvii. 362, 366.

⁹ Further light on the names of the months may be hoped for from the Asia Minor calendar. The ordinary Semitic evidence is given by Lagrange, , Études sur les Rel. Sém. 2nd ed. (1905), pp. 275Google Scholarsqq.

¹⁰ Sayee conjectures that Báκηλος is the Hellenised form of the Hittite, Aba-kali (P.S.B.A. xxiii. 106, n. 2)Google Scholar; if so, it is to be compared with the Assyrian abkallu, wise man, etc. As an official or priestly title the word seems to be found in Nabataean and Palmyrene (Cooke, pp. 223, 296, with references).

¹¹ Denkschr. Wiener Akad. liv. (1911), ii. No. 182.

¹² To add to these conjectures one may note in 7₇, the combination brafṛsis brafṛlu (? -ṣlu) which suggests both the above brvãć and the islũ of the bilingual. That -iũ is merely an ending is probable on other grounds; see below, n. 18 (3c).

¹³ He includes Śfarvad (= Śfard, Sardis) in the same inscription; but the form, recurs in 16₁₉, which does not appear to be metrical.

¹⁴ The fragment 23 mentions taaqdać. For the ending, cf. ñeũmdać (29₄), qidaũmdać (4₆), śfeñdać (11₉, but śfendavmũin, 1. 12).

¹⁵ In la, Śanñaś, Kuoadk, Marivdak, are presumably three gods; the second is Kuoad, but it remains uncertain whether the last is Marivda (pp. 43 seq.) or, as Ehelolf suggests, Merodak = Marduk (p. 85).

¹⁶ Above, p. 83, 1. 7, of the Lydian, read kulmsis with -m- instead of the doubtful -ṡ- which seemed preferable at the time of writng. It is interesting to recall that the place Koloë, near Sardis, with a famous sanctuary of Artemis, probably gave its name to Colossae, whence the ‘Colossians’ of the New Testament (Woodhouse, Ency. Biblica, col. 859 n.).

¹⁷ P. 71. The same inscription contains buk (‘or’) repeated five times, the last with the ending -in. For an exception to this use of -in, see 13₂ (n. 30 below).

¹⁸ Littmann's remarks on affixes and endings (pp. 70 seq., 73 sqq.) may be extended by the following note on typical variations:—

(1) As regards the oblique case in -ṡ, it may be observed that -aṡ (or ad) -eṡ, -oṡ (or -od) beeome -aũ, -eũ, -oũ; but vratoś (19₉, ₁₁) becomes vratuũ in 12₃, an inscription with several peculiarities, and dummś (4₂) becomes dummũit, dumũit (27₂, ₉), where -it is presumably an affixed particle.

(2) For the relation between -s and -d, cf. his, hid whoever (p. 67), iskoś (12₂), -od (1. 10, 29₆), aũaś (7₁₆), -ad (30₆), cf. also also aũẽć (11₇).

(3) Other endings:—

(a) -t, in astrkoś, -koũ (4₁₄, ₂), -kot (19₄), -kotak (29₂); bukmũad (30₅), -at (1. 9), -ũis (1. 8); bitad (7₉), -aad (30₇), -at (19₆); (b) -is, in akmũad (4₁₅), -ũis (30₇); ẽmć (7₁₉), ẽmũ (13₄), ẽmis (19₂), ẽminać (13₄); dumis (29₁), -mćit (l. 14), cf. 1 above.

(c) -lũ, in mũimnś (4₃), -mlũ (1. 14), -mnaś (1. 1), -mnać (1. 8); niviślũ (30₁₀), -śqć (6₆); Artimulũ (Falanga, ); Hũdãnl Artimuũk (7₁)Google Scholar. Cf. also tarblaś (19₆, 34₂), -latil (34₅); and alarmś (3, 26₂, 27₆), -maś (19₁), -mũ (27₂), -mn (16₁₇); and … larmlć (29₅).

(d) -i in sellis srmlis compared with serlik (-i + k) srmli (30₃, ₁₆).

(e) -idć, in qẽnsidć (19₁), inãnidć (16₁₉), faqasidć (l. 16), kotiśfamrasidć (29₉); bidć (30₆, but bidẽć, l. 11); hid trodé … historidć (16₅); cf. also hisredć (14₁, but hisred 26₃).

(f) śfarvad (12₁, 6₁₉); śfarũ (4₅, ₁₀, from śfard?), śfardak (12₉), śfardẽnũ (4₁), -ẽtũ, -ẽtak, -ẽtik, -ẽtać (ll. 2–13).

¹⁹ In 4₁₂ it follows qẽhraũ, but the context does not appear to contain any threat.

²⁰ hiri is the only parallel I have observed among the many Caucasian words collected by Kluge, (Mitt. d. Vorderasiat. Gesell. v. 1907, p. 46)Google Scholar.

²¹ So Dr. Gray independently suggests this possibility as regards the Aramaic, and relating and and asks whether may not be some very general term corresponding to ‘his possessions, in soil and water.’

²² ‘May (Artemis) break up (his) house, destroy (his) goods, spoil (?) his land—may they drive him away’—three verbs in the singular (ending in -k) and the last verb in the plural.

²³ buuk bilis in the inscription noted above (n. 11) is too uncertain. In hũk bũu (7₁₆, cf. hũ l. 6), the oblique case of his, hid (he who, that which, p. 67), can scarcely be in combination with a possessive. It will be seen that the ending -is in the oblique case is --ũ not --iũ.

²⁴ The in in § III. may be an error for the definite affix (Cowley), or (with Gray) an anticipatory suffix, ‘(above Sepharad is) his parbar (viz.) the property of,’ etc.

²⁵ The form kudkit seems to resemble that of dqtdid in 11₈ (Littmann takes q to be an error for -e, pp. 18, 50) and of fadofid in 16₁₁. But the clue may be illusory. Bũtar-vod (b-t-r-v-d), too, resembles the form p-r-b-d which Andreas everywhere reads in place of parbar (p. 26), seeing that -t- may be merely a sign of a derived stem (so as regards varb-tokid, p. 45).

²⁶ Karolaś follows immediately in 11₉, but Sabũalid comes in 11₄ after akad Karolid and before Iśtubeũmlid. Cp. also Sivamlis, -lid and -lũ with -ś;, -d (see above), and -ũ in 15₁, 13₂, 27₁.

²⁷ Cf. also Littmann's discussion p. 16. For -l, and -lũ, cf. above, n. 18 (3c).

²⁸ In the Aramaic papyri from Elephantine, , Sachau, (Aram. Pap., p. 268)Google Scholar cites (with Iranian ending -kan), and (with the further addition of the Aramaic ). It may be added that from the same source comes the Persian the name of an official class, wherein -k is an affix (see Andreas, in Lidzbarski, , Ephemeris, ii. 215)Google Scholar.

²⁹ Cf. the inflection above, n. 26.

³⁰ The inscription begins: (1) eśś vãnas Atalis Tivdalis Tarvnallis, (2) akin kudkañaũres. ak Teśaśtid Sivãmlid, (3) mũola Śrfaśtid Mẽũalid mũola ak nãhis, etc. -in here, however, is not a concluding particle, contrast above, n. 17.

³¹ See above, n. 25. qitalad is the verb in the protasis of 13; and strangely enough Littmann has not recorded the parallels in the (as yet unpublished) inscriptions 7 and 30.

³² Unfortunately not all these and other necessary details are given in this fascicule, and judgment must therefore be suspended. So, for example, nihaaslad (29₆) compared with ninin nid haaslis (27₅) suggests the possibility of the use, in the latter, of a double negative, nin and nid.

³³ Viz. the old Hadad inscription (Cooke No. 61), e.g. f-m-z, l. 3, ‘whatever’; and before verbs in the perfect and imperfect, ll. 14, 31.

³⁴ See Amer. Journ. of Arch. xvi. (1912), 28 sqq. Among the names are Ἀρτϵμᾶς, , a man's name (A.J.A. xviii. 61seq.)Google Scholar. Ἀτταλος, (A.J.A. xviii. 35)Google Scholar, cf. Atalis in L. 13, Dionysias as the name of a tribe (ib. p. 57); Menelaus (ib. p. 66), cf. Manelis 1, etc.; Μῆτρας (p. 58) ?. Mẽtrid, 7, 30; Myrsilus, , etc. (A.J.A. xvi. 45)Google Scholar? cf. Mrslaś—if a proper name, 29₇: Nannas, , Ninis, (A.J.A. xviii. 35, 38)Google Scholar, cf. Nannas, 25; Σαδυάττης, (A.J.A. xvi. 41)Google Scholar? cf. the first syllable of Sadkorfũ 29₃.

³⁵ As further Semitist opinion is necessary, it may be as well to mention that in (l. 2, end), the final is assured by plain traces upon the negatives (as Mr. Buckler kindly informs me), and by Aramaic usage. The traces do not come out however, on the photograph, p. 78 above.

³⁶ Dr. Cowley remarks that ‘the “curious usage of ” (Littmann, pp. 24,29) is common in the Elephantine papyrus of Ahikar and of the Behistun inscription, where it corresponds to the Old Persian pasāva. There is no need to compare the Pehlevi. It is simply due to Persian influence.’ As regards foreign influence the archaeological facts are of interest, and Mr.Butler, H. C. has drawn attention to the resemblance between the jewellery found in Lydia and the Etruscan. The expedition also found seals, gems, etc., of Persian design, perhaps cut for Persian nobles; these may have been of local manufacture (A.J.A. xv. 157, xvi. 479)Google Scholar.

³⁷ To the non-classical student kudkit and vãnaś suggest quidquid and fanum, but a conscientious study of Semitic and Persian lexicons would produce equally curious resemblances elsewhere.

³⁸ It is at least a very curious coincidence that at Sardis there was evidently a cult of ‘Artemis of k-l-w (Koloë) and Ephesus,’ and that the coordination of this Colossian and Ephesian Artemis recalls the close relationship between the Colossians and the Ephesians, and between the Pauline Epistles addressed to each. But it is taken for granted that the Phrygian Colossae is meant, even though the name of the city itself is actually of Sardian origin (see n. 16).