^{page 1 note 1} All credit is due to him for and determinative for place-names: for ‘god’: in his suggested value, the god Tesup (I cannot agree with ‘Sandes’): s: his brilliant identification of the city-name spelt with Tyana (PSBA., xxv, 1903, 179), although I differ slightly from him in his ultimate values, reading Ṭ(a)- a-n(a): ‘king’, which Jensen held (ZDMG., 48, 1894, p. 302) (which I prefer to translate ‘lord’): ‘tree’: perhaps two numerals: ‘bowl’. In he first saw the idea of making a treaty (TSBA., vii, 1882, 276), which he altered later to ‘to love’ (PSBA., xxv, 1903,156); I believe that his former suggestion was nearer the truth, as I think it refers to the making of blood brotherhood (§ 87). In the two forms of he found the idea of ‘great’ or ‘lord’(I believe it to be used as the plural of ‘great’, ‘chief’); but unhappily his suggestion, made in 1882, that meant ‘killing’ or ‘conquering’ he changed to the incorrect one of ‘power’, a view which Rylands (PSBA., xxi, 1899, 210) also held. may be, as he suggests, the ideogram for ‘chief’. In translated from the earliest period of decipherment as ‘I (am)’ or ‘He says’, Professor Sayce, I believe wrongly, ultimately (like Jensen and others) inclined to the former. He is nearly right, I believe, ultimately in calling wa, ua (properly mi, wi), and correctly sees in it the mark of the first person singular of the verb, although his example (which I read kat-mi ‘I', the cuneiform kattimi, not a verb at all) is singularly unfortunate (PSBA., xxiii, 1901, 95): he is nearly correct in with is (I believe it is as); on what I believe are incorrect grounds he obtained correct values for nas and ar, and on unsatisfactory evidence ultimately called ar (I believe it is ir with a ‘tang’). (See his articles PSBA., xxv; 1903; xxvii, 1905.)

Jensen, although we need not much concern ourselves with his system, rightly 1 think, recognized that meant ‘lord’ (Hittiter und Armenier, sign-list), and he very nearly lighted on what I hold to be a most valuable clue in seeing that contained the name of Hamath, and even went so far as to explain the latter two characters as ‘king’, from a comparison with other texts, the whole reading according to him ‘King of Hamath’. But he failed entirely to give syllabic values to the name of Hamath, saying that its first character might in some cases be a plural ending, and in the translations in Hittiter und Armenier he relinquished the view that this group meant Hamath (ZDMG., 48, pp. 301 ff.: see also Messerschmidt, Bemerkungen zn den Hett. Inschr., p. 15, who quotes him). Sayce also came close to seeing this, but his incorrect division of the signs in the inscription prevented him from identifying it, and I cannot agree in the least with his latest translations of the Hamath inscriptions (PSBA., xxxiv, 1912, 217). Jensen was led astray, I think, entirely in seeing Syennesis in the name which I read A-r-ar-a-s (§ 12).

A word must be said for Menant (‘Eléments dn Syllabaire hétéen’, Acad. des. Inscr. xxxiv, 1892) who saw in ( which I believe to be e) a vowel a (p. 100); and Peiser saw in the division mark, and in the mark of an ideogram, according to Sayce, the plural (see, however, the sign-list at end of this article). Ball (PSBA., x, 1888, p. 447) recognized in the proper name, which I believe to be Benhadad (§ 33, note), a royal name of which the first character was the god Dadi. W. H. Rylands (to whose energy much of the collection of Hittite material in the early days of the study is due) noted that ‘on the shoulder of the [Mar'ash] lion at Constantinople is a human figure’, which, unfortunately, he says formed no part of the inscription (PSBA., ix, 1887,375): nevertheless, it has been omitted in the copy in Messerschmidt's Corpus. I do not think we need concern ourselves with the work of Conder or Gleye.

The greatest stumbling-blocks in the way of decipherment appear to have been the bilingual Boss of Tarkondemos, and two groups of hieroglyphs which occur several times in the Carchemish texts, in which has been recognized, I believe entirely erroneously, the name Carchemish, spelt variantly (a suggestion due to M. Six), and consequently to several characters were assigned values due to the supposed variants. Personally, I believe the latter part of these groups to contain the words Nineveh and Assyria, and that none of the phrases has anything to do with Carchemish at all (see § 51).

After the Society of Antiquaries had offered me the courtesy of hearing this paper read on November 21, 1912, Dr. Rusch, a German scholar, saw a brief notice of the meeting in the Orientalische Literaturzeitung of the January following. As he had been working on a system of his own during the same time as myself, he not unnaturally wished to draw attention to such claims as he might have to any priority of decipherment; and to this end wrote to the President of this Society giving references to notices of his system, and sending to him a manuscript copy of his labours. I think that I can satisfy Dr. Rusch that our systems are so fundamentally different that one of us is wrong. His work is referred to in Deutscher Reichsanzeiger, 1911, No. 269; 1912, Nos. 38, 114; and by von Scala in Internationaler Archacologen-Kongress, Okt., 1912. The following is the list of proper names which he has discovered in the hieroglyphs, according to the Deutsch. Reichsanz., in a reference to a meeting of the Vorderas. Gesellschaft:—Lapa, Lupastius, Teschubis, Teschub-Tarchu, Teschupiha, Teschuputias, Targurtisar, Argurstis, Motarvu, Hatti-Teschub, Arha, Arrapa, Kisch, Kuti, Kararkarti, Patesi, Sutech, Tarchus, Maarsi, Sigur, Huchu, Motar, Gurtis, Gurtius, Sepasuvu, Tarmispa, Teschup, Tarchi-Hattis, Ischtar-Gurtis, Teschupgurtisicha, Archa, Haartichamis, Motargurtis, Aryatarpa, Hapagurti, Luku, Teschuparpas, Pasaas, Tarchumispas, Teschupas, Tarchusapasus, Teschupucha, Teschuptis, Tarchuhattis, Argurmis, Gurtyas, Motaragurmis, Gurpas, Teteschup, Teschupgurtispas, Teschupicha, Teschupti-tarchurus, Arra, Hattisteschup, Teschuparra. From this list I think that it will be obvious that our two methods of decipherment have nothing in common, as only in one single word (the name of the god Tesup, long ago discovered) do we agree, and I hardly think it worth while to discuss his manuscript translations in which I cannot follow him. He goes so far as to adopt Professor Sayce's values for the signs for god, Tesup, ‘land’ (the double peak), s, the armed hand as Krieg, ‘throne’ (Ehrenplatz), and the ideographic value ‘water’, with the numeral nine, and the two (unused) numerals three and four, and in the second of the two numeral signs quoted in § 1, note, he sees ‘1000’: the sign of the two legs running was given the possible meaning of ‘to run’ by Menant (Eléments, 1892, p. 105); and he sees in a number of obvious ideograms their picture values, such as the sign of the head with the tongue protruding, the ideographic meaning ‘speak’, the foot (I deny the leg) ‘to go’, and the Tesupsign the lightning or serpent, in which I gladly concede to him any priority as far as I am concerned.

As far as I can see, the values for the remainder of his signs, which are liberally compared to both Egyptian hieroglyphs and Assyrian cuneiform signs, are different from mine. He has relinquished the view (I believe an erroneous one) that the larger figure at Ivriz is a god, but his attempt to read the name as ‘Hatti-Tešub ’is impossible, for he includes the first word ‘I am’ as part of the name.

I have gone thus fully into Dr. Rusch's system because it is unpublished for the most part, and I wish to make it quite clear that we differ entirely.

In the following pages it will be found that the footnotes frequently give sign-values and transliterations for convenience sake before the evidence of such equivalents in the main body of the article has been reached.

^{page 4 note 1} Abbreviations in this article: A i, A ii = the two Arzawa letters (see Knudtzon, , Die zwei Arzawa-Briefe: Die el-Amarna Ta/eln, 270 ff.Google Scholar). Al = Aleppo Tablet, Sayce, , PSBA., xxix, 1907, 91Google Scholar: B = Belck, , Anatolia, ii:Google ScholarC = Chantre, , Mission en Cappadoce:Google ScholarD = Sayce, , JRAS-, 1908, 985Google Scholar: E = Sayce, 1907, 913: F = ibid., the second tablet: G = Sayce, JRAS., 1909, 974: H, I, J,'K, L, M, N, O, P, tablets published ibid. 963 ff.: Id. = Ideogram: Liv. = tablets published by Pinches, Liverpool Annals, vol. iii: M i, M ii, &c. = Messerschmidt, Corpus lnscriptionuni Hettiticarum: Rams. = Ramsay, , PSBA., xxxi, 1909, 83Google Scholar (an inscription on Kara Dagh): TA, Tel Ahmar = Inscription from Tel Ahmar, Hogarth, Liverpool Annals, ii. 165 (in many cases I have added emendations to his text from my handcopy made from the actual stone while employed by the Trustees of the British Museum): Winckler = Mitteil. d. Deufsch. Orient. Gesellschaft, 1907, no. 35: Y = Sayce and Pinches, The Tablet from Yuzgat (Royal Asiatic Society Monographs, 1907): Z = some Hittite cuneiform tablets published by me in PSBA., xxxii, 1910, 191: ZDMG. — Zeitschrift der Deutschen Morgenländischen Gesellschaft.

^{page 4 note 2} Not ‘left there to be destroyed’, as Professor Sayce describes it (PSBA., xxvii, 1905, 210Google Scholar).

^{page 5 note 1} The Hittite hieroglyphs read boustrophedon, but for the convenience of the text I shall always write them beginning from the left.

^{page 5 note 2} I did not recognize for a long time that the obvious reading for the last word was San-gar-s, and not Gar-gam(?)-s, the characters being arranged so as to please the eye. This second Sangar must have been grandfather of this Sangar who wrote the inscription, according to the ancient habit of calling a son after his grandfather.

^{page 6 note 1} Or transpose these last two characters.

^{page 6 note 2} This must be restored m.

^{page 6 note 3} With regard to s as distinct from š, the Hittite cuneiform in seven or eight hundred words shows barely a dozen certain cases of 5: notably we find a word sa-an, A ii, 7. For the reason that so few cases occur I am for the present using only s in the hieroglyphs: for a discussion on this see § 90.

^{page 6 note 4} For a third indication of the value of this character cf. Gar-a-li, i.e. § 11.

^{page 7 note 1} M xxxiv (Nachtrag). My copy, made from the rock in 1909, is the same for mi-; on the reading of the proper name see the translation at end.

^{page 7 note 2} ‘In spite of the strangeness of the expression I am inclined to see in mê the Hittite first personal pronoun ‘(PSBA. xxi, 1899, 204Google Scholar). For the cuneiform text on the ‘Tarkondemos’ Boss see translation at end, ‘I am Targašša-?-wi’. Sayce was very nearly right in his final value wa or ua for . Halévy considered as the signs for a vowel; Hommel (in his list of signs settled ‘without any doubt’, PSBA. xxi. 233) considered this correct; and Jensen also couples as ‘à d.i. a und o (auch u?)’ (Hittiter und Armenier, sign-list). Sayce in PSBA. xxiii. 99 held that denoted the first singular of the verb, and consequently ‘ will be i orya’; this he altered to wa or ua in 1905 (PSBA. xxvii. 245), but curiously read as was..

^{page 8 note 1} An examination of the hieroglyphic texts will show that, in sense at least, the name may be written with or without the addition of

^{page 8 note 2} PSBA. xxvii, 1905, 192Google Scholar. His last rendering (PSBA. xxxv, 1913, 12Google Scholar) does not seem so good: ‘this bowl, in the temple of Sandes (the god) of Atuna I have made.’

^{page 8 note 3} The word following this group is distinct and well known, and is thus correctly separated.

^{page 9 note 1} postpositive occurs with and without suffixes many times in the Hittite hieroglyphs (see §33, note: for examples without suffixes see e.g. M ii, 2: xxi, 4, 5; lii, 3; TA 4, &c, in the translations at end.) would appear to have the value of both g and k; for it can take the place of g in Sangar and Gargamiš (which, however, the Hebrews wrote Karkemiš), and it is used in the following phrases: Ar-am ḫ K-as-k ‘Aram, chief of Kaškai’ (§§ 24, 35); K-r-a ḫ in M xi, 4, 5, the chief Kirrî (§§ 27, 35): K-a-u-a-u-t Kat-t-e ‘the Kauai of Katê’ (§§ 27, 60), and in a new Jerabis inscription on which are mentioned several kings of the ninth century we find K-a-k, i. e. Kâki (§ 24, see § 87). On TA 4 U-'m-k (the place-name'Amk) occurs: cf. M xxxii, 3, and Am-k on a new Jer. inscr., §52 (5). Moreover the sign gar is used to spell the first syllable of the chief's name (M Iii, 4), and hence we may consider that the distinction between g and k was not very great. In Egyptian, for ihstance, the word Carchemish could be spelt with ḳ or k. I shall therefore represent the Hittite symbol henceforth usually as k for the sake of simplicity.

^{page 10 note 1} There are even traces of the diagonal mark which indicates a proper name (see § 17).

^{page 11 note 1} E.g. Sayce (TSBA., vii, 1882, p. 278Google Scholar) thought that implied to speak or say, but later considered that it = ‘I’, though with much to be said in favour of ‘he says’ (PSBA., xxi, 1899, p. 213Google Scholar). Hommel held that Menant was correct in making ‘I (1st sing) (resp. ‘I am ’)’ (cf. his list, loc. cit. 233). Messerschmidt inclined to ‘I ‘and not ‘he says', as Peiser would have it (Mitteil. der Vorderas. Gesellsch., 1898, 6Google Scholar): but in The Hittites, 1903, 28Google ScholarPubMed, he admitted both possibilities. Jensen (Hittiter und Armenier, sign-list) also considered it ‘I’. For reasons stated later I-hope to show that the meaning ‘I ‘is impossible, and that ‘say’ is the probable one. In its usage a cursory examination shows that there is not much apparent difference in the sense of (a) (b) or (c) and we may regard the additions as auxiliaries. This is discussed in § 74.

^{page 12 note 1} The text runs: (32)... wa ^iluGul-aš-ša-an ^iluMaḫ ḫal-zi-iš-tin šum-ku-wa a-bi-e a-ki-i[r] (33) … e-ia im-ma *a-ki-tr mu-me a-bi-e-el urn BÂBU-aš ḫa-aḫ-ḫi-ma… (34) … ḫa-aḫ-ḫi-ma-aš ^iluIM-ni te-iz-zi ku-u-ši-wa bi-iš-ša-at-ti … (35) … -ši ḫu-u-ma-ante-eš a-ki-ir mu-me ki-i-ni GAL-ri ….

^{page 12 note 2} The text runs: LU-kan bi-e-te-ir …

^{page 12 note 3} Another instance of this final -ir in the hieroglyphs is =: a -ir (M xxxii, 5). occurs in M xxxii, 2, 4, and in xxxiii, 2, 6, and if be the root, then we can add this example to our list (on this question see translation to M ix, 3, note, at end).

^{page 12 note 4} For full proof see § 15. In order not to make matters too complicated, it is shown in § 17 that this little stroke or tang is frequently added to characters to show that a proper name is indicated.

^{page 13 note 1} In (M vi, 2), (M iv, A 2: B 2), (M xxxii, 1), the second character as it stands cannot be confused with ir, if the texts have been copied correctly: in the Restan text, which is practically a duplicate at this point of the Hamath texts, we find simply, so that the additional sign probably has no material value. On the other hand, I cannot find any satisfactory comparison for this final b(a) (as § 40 shows it be) in the cuneiform texts.

M xvi A shows a verb ending with an animal's head, but it is not quite clear whether this is -u or -ir (§ 50, 6). Incidentally, I should add that further proof that fi^ = ir will be found in the placename as-ir, which varies with as-r-a ( = Assyria, § 51).

^{page 13 note 2} Knudtzon, Die zwei Arzawa-Briefe, 1902, p. 19, gives the following list of names in which this god-name occurs: Tarkondemos, Tarkondimŏtos, Tarkodimantos, Tarkuaris, Tarkumbiou (gen), Trokoarbasios, Trokozarmas, Trokokombigrĕmis, (T)arkiônin, Tarkundberran, Tarkondarios, Dastarkon (a fortress in Kataonia), Lycian Trqqas, Trqqiz, Trqqnta, Trqqñti, Trqqñtasi, &c. As is well known, Tarḫu in Tarḫulara (the prefect of Gurgum or Markasas WAI. ii, 67, 45, 58: iii, 9, 52; B.M. tablet, K 1660) is the same god.

^{page 14 note 1} I question the correctness of my copy of this character (zi) from the stone; if I were copying it again I should look for or less probably . I believe the form Arhulini is also to be seen in M lii, 2, concealed in (reading ḫu for the second syllable).

^{page 14 note 2} M xxxiii, 1: cf. xxxi, c, 3: xxxii, 1, 2, 3: xxxiv, A, B, C.

^{page 15 note 1} Emend thus, instead of ir.

^{page 15 note 2} See note to § 1.

^{page 16 note 1} An examination of the photograph in Hogarth's article in Recueil de Travaux, xvii. 25, shows that these are possible readings.

^{page 16 note 2} The difference between this group and the others is so marked that it would be safer to collate it before laying too great stress on it. Compare, however, no. (k) further on.

^{page 17 note 1} This emendation of the form of this character in this passage (Restan i) is proper from the forms shown in other Hamath texts, e.g. M iv, A 2: B, 2: vi, 2: and the variations in the Mar'ash texts, M xxi, 4, lii, 4.

^{page 18 note 1} I see that I have accidentally omitted the small ‘tang’ to the character III here.

^{page 18 note 2} Professor Sayce thought that it indicated the adjectival termination, but, as will be seen from § 39, this is impossible.

^{page 19 note 1} The first figure is to be seen on the lion in the cast in the Museum (see note to § 1).

^{page 19 note 2} From this (slightly obscure) diagonal it is clear that the first part of this word is not to be compared (as Sayce took it to be) with the word ordinarily placed second after at the beginning of inscriptions.

^{page 20 note 1} It must, however, be noted that in certain cases characters marked by a tang do not denote a proper name, but in some respect call attention to it, and apparently the tang sometimes indicates a vowel sound.

^{page 20 note 2} I would suggest, in spite of the evidence afforded by M xvi, c (Menant) and the Malatia inscription published by Sayce (PSBA. xxvi, 1904, 23; see M xlvii), both of which read that we should read ḫu here, making the whole -ḫu-li-e-ni, and providing some value *r for so that the whole may represent Irḫulini. Two of the other kings at least in this inscription are known to be his contemporaries from the hieroglyphic texts. At the same time this is only a suggestion until we have a certain value for the first character: for another possibility see p. 112. On e see § 46.

^{page 21 note 1} I am indebted for much of this historical sketch to Maspero's Passing of the Empires, where an excellent and full account of the conditions prevailing in the Hittite lands in the ninth century is given.

^{page 22 note 1} For the latest evidence of the position of Kummuḫ, see L. W. King, PSBA., xxxv, 1913, 73.

^{page 22 note 2} Am-me-ba-'-la, the son of Zamani. The name appears to be Semitic. This Amme may well be the same as in Pan-ammu, the name of two kings of Sam'al (Sinjerli) some distance west of Bit-Zamani (cf. Heb. the father of Bath-Sheba, 1 Chron. iii. 5). Hence Amme-ba'al may mean ‘Amme is (my) lord’, just as Pan-Ammu would mean ‘Face of Ammi’ (cf. the Phoenician phrase and the Heb. name

^{page 22 note 3} Thanks to the kindness of the Trustees of the British Museum I was allowed to publish one of the results of an expedition on which they sent Mr. T. E. Lawrence and me from Carchemish to Tel Ahmar, a copy of the fallen lions inscribed in cuneiform which Mr. Hogarth had seen on his visit there. The inscription thereon shows beyond a doubt that Til Barsip was Tel Ahmar, and not Birejik (see PSBA., xxxiv, 1912, 66).

^{page 24 note 1} An interesting example of this is found on the Aramaic stele discovered somewhere in these regions by Pognon (where exactly he will not reveal) and published by him in Inscr. Sem., p. 158. It is a stele written by Zakir, king of Hamath and who describes his fight against ‘Bar-Hadad, the son of Hazael, king of Aram’ who had united against him the following coalition: ‘Bar-Hadad and his army and Bar-Ga's and his army, the king of Ḳaweh (Ḳauai) and his army, the king of ‘Amk (Assyrian Unḳi) and his army; the king of Gurgu[m] and his army, the king of Šam'al and his army, the king of Malaz (Malatia) and his army.’ Noticeable is it that Kummuh is not mentioned.

^{page 25 note 1} The texts say nothing of Mutallu joining the coalition of Aḫuni.

^{page 26 note 1} Are we to see on the slab from Carchemish M xii, 2 (perhaps a fragment of No. 1, which is sculptured with an Assyrian winged figure) the name Pitru (Witru) in M-t-r (= W-t-r)-1 country’ + ‘king’?

^{page 26 note 2} On the Bronze Gates of Balawat is represented the capture of another city ‘Aštamaku of Irhulini’.

^{page 27 note 1} Passing of the Empires, p, 83.

^{page 28 note 1} Pointed out by Sayce in JRAS. xiv. 44.

^{page 28 note 2} Read for my copy.

^{page 29 note 1} This group (the brother of Panammi) possibly occurs on M vii, i badly written.

^{page 29 note 2} See Cooke, North Semitic Inscriptions (the Sinjerli Inscriptions).

^{page 30 note 1} Restored from parallel at beginning of 1. 4 and M xxi, 2.

^{page 30 note 2} Also in M xxv, 3 and xxiv, B all from Mar'ash.

^{page 30 note 3} From these two variants a value 's is suggested for , which takes the place of This is also apparent in two other cases in the two quotations above. The ibex's head never, as far as I know, takes the place of 5 after i (in ni, mi), and hence we must read it as or us. The former becomes a certainty when we consider K-as-k (Kaškai, § 35), As-ir, As-r-a (Assyria, § 51). (See note to § 1.)

^{page 31 note 1} Tarḫulara is prefect of Gurgum (WAI. ii. 67, 45, 58: lii. 9, 52) or Marḳasa (Assyrian tablet in B.M. K 1660).

^{page 31 note 2} It must be remembered that both Assyrian and Aramaic forms of Gurgum would be transliterations of the native name.

^{page 31 note 3} u-'m-k = Unḳi = ‘Amḳ occurs on T.A. 4. See § 52 (5).

^{page 31 note 4} That the word ends here is clear from a comparison with M lii, 1 (quoted also § 29), where (Parallel with which follows our word immediately in M xxi) is clearly a distinct word.

^{page 33 note 1} If the cast of the Mar'ash lion in the B.M. (M xxi) be examined it will be found to have the second paragraph (i.e. the end of 1. i) thus: ‘Benhadad unto the son of his brother … li’ (i.e. Mu-tal-li?); and in the middle of the second line it is possible that four characters read *Mu-*tal-li-s (see translation at end).

^{page 33 note 2} We find similar ideas in Egyptian ideographs, = ‘to give’, = ‘to grasp’, ‘to fight’. The Hittite and may then be suspected to mean ‘to take’ and ‘to place’ respectively. This hand holding the graving-tool occurs in M iii, B, 3; iv, A, 3 (and is broken away in B), where presumably Irḫulina says ‘I have graven our covenant (?) with (So-and-so)’. Similarhy in a new Jerabis inscr. (see § 68 (10)) ‘So-and-so hath graven (?) covenants with me’. That it does not mean simply ‘to write’, as I first thought, is shown by M ii,4 ‘our allies have graven the leg(?)(= base?) of the memorial (?)’ (see § 48 (5)); moreover, the picture in the hieroglyphs points to a large tool held in the grip, unlike a pen. As is shown in § 48, we have the root t-e certainly meaning ‘to say’ (in M ii, 5 this is paralleled by the ideograph ‘engrave’); a third root 5 which occurs in similar passages must have a similar meaning, and 1 propose the value ‘write’ on the following grounds:—The actual root is certain from the word s-r-a following chiefs’ names (in exactly the same manner as t-e-r-a ‘they say’); e.g. ‘(NN in the land of?) s-r-a: kat-n:t-e: I; have written, We are of one speech (accord),’ &c. (TA 4: § 61) (cf. …; t_(a) s-r-a ‘enemy ‘: I-k-n-tn: g(k)-mu e-a-t(a) ‘have written, Against my (? = our?) common enemy I will go with thee ’, TA 3). Cf. also M xxxii, 3 …? an-nas ID (or U-) Am-k: ID; -n s-r-a: ‘ god’-‘friend’-k-ni-nis-k-n: Kar-a-tal(?)-k-n, ‘… annas, chief (?) of Amḳ a.. have written: Unto (or, By?) the god of (our?) friend, unto Karatal(?)’ (see M. xxi, 2, 5, 6, comparing there’ god’-‘brother’ -k-ni-n-nis and ‘god’-’ friend ’-k-ni-nis). The perfect of this verb s with the augment occurs in a newjer.inscr. in the form a-s-t. ‘ So-and-so: ID-n(a)a-s-t “brother”-e-k: san-t-n-s (or san(n)-st(a)) hath written a (our) … “Like a brother(s) thou makest us (or thou actest)” ’(§§69, 76). The word a-s-t occurs once elsewhere. A form s-u-n appears in ‘The pledges (?) of So-and-so s-u-n (I have written)’ (new Jerabis): n-m-n s-u-n ‘a covenant (?) I have written’ (new Jerabis). In the case of M xv, B are we to read 11. 2–3 kat-s n-u kat-u-n: “Pan-am-mi n-m-n s-u-n (§71)? In the next line n-m-n s-u-n ends the inscription following after a chief's name. A form s-u occurs M lii, 3.

I had at first thought that this root s meant ‘to send’, but I believe that the meaning ‘write ’ is the correct one, on account of the following noun s-e/which would seem to come from it. The most striking instances appear to me to be in (1) M xxxiv, A (Ivriz): ‘I am Tesup-mis.., I am Ariarathides; we have given our alliance (hands); (1. 3) s-e “ally ”-na t(a)-?, the writing of our alliance giv[ing].’ S-e here must mean some tangible proof of the alliance. (2) M xxxiii, 2: ‘This tablet of making alliance hath brought gifts (?): te (?) s-e-t(a): t-a: mi-t(a) ID; “T(a)-a-nas, thy letter did speak concerning (?) Tyanian wood.’ (On this quotation see translation at end.) The other instance which I know is:—(Mi) ”Gu-n-nas ntan(?)-mu s-e: n-m-n-e ‘god’ Targu (?)-r-r-s ‘god’ Sul (?)-e-s: mi-ni, ‘ Gunnas (?), my … (?), hath accepted (?) the writing of the covenant of Targu-ras (and) Sul(?)-es.’ (See translation at end.) Does s-e-n-tni in this text also belong here?

Are we to see the root in the verb after kat-tni (ibid.) and read s-un? (see § 70, note). In M*vii, 2 .gu(?);-e-u io(?) ID s-e n-m-n-an u-t(?).. san …. the phrase is the same (‘writing of our covenant-’) as in the previous example, and considering the limited possibilities of the verb s we shall probably not be far wrong in considering the meaning to be ‘write’ with a noun s-e (whatever number or case) ‘a writing’. Is this endorsed by the Hittite cuneiform?

The meaning ‘write’ fits the following case:—(Y r. 11) nu ma-a-an ^anUD-ns a-aš-šu ku-e ṭa- …. ‘Unto our lord (§ 44) the Sun-god (i.e. king) they have written, “Gifts giving … ”’(the next line ending ‘for a gift a poor man brings to thee a sheep’), where I take a-aš-šu to be the augmented tense of s with u termination, as in § 71.

Other possible occurrences of the root are (G 16) nam-ma-aš GIŠ-NI-it ša-ra-a ḫn-it-ti-ia-an-zi: (Y r. 1)… ta na iš AN.EN. ZU.NA ši-i-e-it.…: (Y r. 2 ) … a-ar BÂBU.GAL-aš bi-niš ši-i-e-it ameli MES: (Y r. 4) … ša li-it aḫ-ḫa-ti ša-at u-nl ṭa-aḫ-ḫu-un. Cf. ša-a-ak-ki (Y 20); [š]a-ia-at (A ii, 5).

^{page 35 note 1} A list of the suffixes will be found in § 58, with the reasons for their identification, and I have consequently not repeated them here. The examples for k-n with ist pers. sing, and pl. Are (i) TA 7: i.e.: ID; k-n-m ‘unto my table (?)’; (2) M x, 7 ID-k-n-m ‘against mine enemy’; (3) M lii, 5 ID--k-n-n-m ‘to my? ’; (4) M ii, 4 m-n-u-k-n-an a-ṭ(a)-t ‘for our memorial (?) he hath given.’

^{page 35 note 2} I am much inclined to suggest that these two signs are a-ḫu, and to recognize the word as one of those adopted, by the Hittites from Assyria, translating it ‘brother’ on the analogy of a-bi-e (Y 32, 33), a-bu-n-uš (Y 37, 38), a-bu-u-un-na (Y 27, 31), &c, ‘father’. There is the bare possibility of aḫm occurring once for ‘brother’ in the Hittite cuneiform (Al. r. 18, a letter) ili-lu GIŠ UD ‘”Za-ar-šc-AN-MAR-TU a-ḫa-ti-wa ḫat-ra-at (‘Zarše-Martu, thine other brother?’): the Sumerian ideograph is, however, often used. (See § 89.) Tesup-k occurs elsewhere in the hieroglyphic texts.

^{page 35 note 3} This is a name which occurs elsewhere: see § 49.

^{page 35 note 4} This name is so important and occurs so frequently that it is better to discuss it here. occurs on the long Jerabis inscription; on M ii, 1, 4, 6 (from Babylon): iii A, I (Aleppo): iii B, 3 (Hamath): xvi, A, I (Malatia): xxi, 1, 2 and lii, 1, 3 (Mar'ash). Ball, as far back as 1887 (PSBA., ix, 1887, 447), recognized that this was a royal name of which the first part was Dadi. This king is one of four who have written the Malatia inscription, the first being probably Irḫulîna: on the long Jerabis inscription he is mentioned again with one of these kings, and he is the actual writer of the two long Mar'ash inscriptions to Mutallu. In the Malatia as well as the Mar'ash inscriptions he is undoubtedly suggesting an alliance with the reigning king of those lands, and from the ubiquity of his name it is clear his power was widely recognized.

The name is made up of the sign for Tesup without the god-sign; then an unknown sign which I have not met outside this name; and finally the sign r. Thus we get Tesup-?-r, or, since Tesup is Hadad, Hadad-7-r, which looks very much as though we had the Assyrian form of the name of Benhadad II, Adad-id-ri. When we consider the frequence of his name, that he is mentioned on Irḫulîna's inscription at Hamath, that he writes to Mutallu of Gurgum and his adopted (?) son, and Arame, telling them of the alliances of several kings, among whom are Ḳaral, Katê and Nks, and probably Panammu and Irḫulîna, that he joins with Irḫulîna (?) and two other kings to ask alliance with the king of Malatia, and, as negative evidence, that he is not mentioned in the later published texts of Carchemish, in which occurs the name of Kirrî (who was not put on the throne of the Kauai until 834, § 27), we may well see Benhadad in this name, reading his name Tesup (Hadad)-id (or iz?)-r.

^{page 36 note 1} And also k-s, presumably by assimilation for k-n-s in M xxxii, i: xxxiii, i. At first sight k-n-nis would lead one to suspect ist pl. suffix rather than 3rd sing., especially when ‘make alliance with us’ frequently follows, and the difficult case ‘brother’ -k-ni nin-as ‘unto the son of.. brother’ (M xxi, 1) occurs. But we find k-n-s in the same line as k-n-ni-s, referring to the same subject (M xxi, 1), and in M xxxiii, 1 ‘make alliance with us’ follows the simplest form k-s. Moreover, the doubled n occurs in the ist pers. sing, k-n-n-m as well as k-n-m, and consequently the balance of evidence is in favour of our seeing the 3rd pers. in k-n-nis.

^{page 36 note 2} See the note to § 67.

^{page 37 note 1} The word kask undoubtedly ends at the -k, for the next word is well known. The Kaškai are a well-known tribe to the north-west of Assyria, and it is probable that this is a more definite description of one of the two kings called Arame by Shalmaneser, either of Bît-Agûsi or Urarṭa (see § 30). I have suggested ‘affirm’ or ‘swear’ for the ideograph of the head with protruding tongue (see end of note to § 1).

^{page 37 note 2} It occurs A i, 6, 7,10: Y [8], 15, 35, r. 33(?): cf. ḫu-iva-an-ti Y 9; ḫu-u-wa-an-da A i. 26: ḫu-iš Y 44, cf. Y 6: ḫu-u-i-ia-an-te-eš Z i, 7: ḫu-(u)-i-ba D 14: E 12: ḫu-u-wa-aš-ša C vi, 9, &c. Particularly compare the name in Assyrian letters ^mḪu-te-sup, B.M., K 1037, 1067 (period c. 700 B.C.),’ Great is Tesup’.

^{page 38 note 1} Bu. 88, 10–13, 43 (Bezold-Budge, Tell-el-Amarna Tablets, no. 1, 1. 6).

^{page 38 note 2} My copy made in 1911 from the stone gives but I have no hesitation in suggesting that it should be from a comparison with other texts.

^{page 39 note 1} In order not to stray too far from the subject of = zi, I append cases of the use of zi placed after names. An excellent example of this in the hieroglyphs is given in § 3: San-gars N-k-s “Gar-k-m-s+‘ place ’-zi. Now a postpositive zi in the case of nouns occurs in Hittite cuneiform, and Professor Sayce rightly hazarded that it meant ‘in ‘in the instances in cuneiform in which he met it:—

C 1, 15 Ma-a-an LUGAL-uš ^aluA-ri-in-na-az ^aluHa-at-iu-ši-pa-iz-zi (‘in the city Ḫattu-šipa’): ibid., r. 2 ^aluIa-ti-il ^aluḪa-iz-zi nu ^amelGIŠ.PA LUGAL-uš. Similarly we may see it perhaps in Y 30 … ak-ki-iš ḫar-aš-zi te-ri-ib-zi wa-a-tar na-a-i ḫal-di-in-na …: and possibly L 2 nam-ma-gan E(? or LAḪ?-iz-zi

Other words end thus in -zi which may be nouns, but it is difficult to be definite about them, as both -anzi and -izzi appear to be verbal terminations. But ^iluTenusizzi (K 7) is fairly certain with a slightly different meaning than ‘in’ for the preposition. ‘In’ is definitely the sense in our hieroglyphic passage ‘Sangar (and) Nks in Carchemish ’. That this is no quid pro quo is obvious from the Jerabis text M xv, B. For other examples of this postpositive -zi, cf. in a new Jerabis inscription : Mi-zi-ir-zi+ ‘place’,‘in Mizir’, i. e. Muzri, to the north-west of Assyria: (TA 4):?-‘place’-zi ‘In the land of?’ (unless it should mean ‘in the speech of the land’, which I think unlikely). The same use of -zi as in ^iluTenusizzi occurs in the hieroglyphs of M ix, 4, where -zi takes the place of -5 in the parallel passage in 1.2 in the same inscription after all three personal names; and M x, 2 ḫu IX-zi ‘against the chiefs of the Nine ’(quoting their names). Cf. also Ḫu-ni-zi ‘with (or against) Aḫuni’ (TA 1), Aḫuni being the chief of Bît-Adini, the neighbourhood of Tel Ahmar where this inscription was found; his name occurs twice in M lii (II. 1 and 2), : “Ḫu-n(u), with a tang, indicating a personal name (see § 24). For the cases of u-zi ‘ with them ‘see the list of pronouns, § 58.

Under this heading doubtless should come the cuneiform -za in such phrases as ma-a-an-za LUGAL-uš ‘unto our lord king’ (see § 44), nam-ma-za (A ii, 22), am-mu-uk-ku-za (W 19), &c. But as we find both i-ia-an-zi (Y 7; P 4), and i-ia-an-za (P n), which may be only careless variants owing to the final vowel of the Hittite zi being slurred, it does not seem unlikely that this -za is merely a variant of -zi.

^{page 41 note 1} These may be transliterated? -r-an-n-ḫ-s- ‘place’, Ar(?)-man(?)-a-ḫ-nas-e-a- ‘place’, B-s-ḫ-r-a-ḫ-nas- ‘place ’. See § 56.

^{page 43 note 1} The names which occur containing this character, Bauli the son of Mutallu,—banin, a chief of the ‘Nine’ (M lii, i, 2, &c), Bark, a chief of the ‘Nine’ (§ 73), give no further proof one way or the other, as I cannot identify them with known kings, although Ba'ali is of course possible as a name, and Ilu-ibni was king of Suhi. Indeed the opening Speech of M ii, ‘Saith Ṭ(a)-?-ar-s unto his lord (?) Mutallu, “Thy father (?) (and) Benhadad the great have given thee a memorial-stela (?) for the commemoration (?) (glory (?)) of Tesup (Hadad)” ’, holds out more prospect of confirmation of the value b(a), for we have seen that the Assyrian word abu ‘ father’ had been taken over into the language of the Hittite cuneiform, and here we have a-b(a)-u-t(a). (See § 89.) For a-b(a)-ir, see notes to trans, to M ix at end, which shows that b and p interchange in the hieroglyphs like k and g. For additional examples of the prep. a-b(a) see M i: ix, 2: x, 2, 5, 7, 8: xi, 2, &c.

^{page 43 note 2} In three similar texts (Hamath) the b(a)-a is omitted altogether.

^{page 44 note 1} The a-b(a)-t(a) which occurs on TA 1 immediately following the phrase ‘make alliance with us’ must belong to the succeeding sentence. In TA 4 it is interesting to see the phrase : ‘ally’: + ḫ-n-(ii(a)): ‘we have made alliance’ for the ultimate agreement (§ 37 (3)).

^{page 44 note 2} The seal M xxxix, 10 is a case in point, as it contains only the name Targu which should be a personal name, since it is on a seal.

^{page 44 note 3} -mu is a form of -mi, the 1st pers. sing, pron, suff., §§ 57, 58.

^{page 45 note 1} How are we to read occurring three times on a text M xxxv from near Tyriaeum (also perhaps Tyrasion, Tyganion, Totarion, or Tetradion)? Are we to see Ter-wu-um in it?

^{page 45 note 2} I cannot help thinking that the plural is correct here. it is true, is used as a determinative for a country, but it may be in its form ‘double ’, and consequently may well have the value of a pluralis excellentiae, as ‘chiefest god’ seems to have, for is used for the plural of ḫ ‘great’ (§38). The phrase on the ‘Tarkondemos’ Boss is translated šar mâti ali, literally, ‘king of country-city’.

^{page 45 note 3} On Arinna = the Egyptian Arnna of the Rameses-Khetasar treaty, see Sayce PSBA., xxi, 1899, 196: xxiii, 1901, 98.

^{page 45 note 4} Professor Sayce was the first to see the meaning ‘my lord’ for maš-ši-ia in this passage.

^{page 46 note 1} On the basis that the hieroglyphic ma-ḫ-n-s, &c, meaning ‘our great lord’ is found in cuneiform as maḫ-an, &c, it might be profitable to see if the hieroglyphic groups ‘god’ -ḫ-m-n ‘by my great god’ (M xxi, 4), mi-t-m: ‘god’ -ḫ-m ‘(as) my great god (is) with me’ (or similar oath, new Jerabis inscription, § 81) can be identified in cuneiform also, so that we might learn the Hittite word for ‘god ’.

A word which might possibly solve this difficulty occurs as araḫzanṭa in A ii, 19; arḫa, G 4, 12, 14, 16, 19 (cf. Liv. ii, 15): S i, 2: ar-ḫa-a-an, S i, r. 9.

A-ra-aḫ-za-an-ṭa can be divided up into a noun araḫ with za + anṭa a compound preposition (cf. §§ 37 note, 79) similar to ma-za-gan (W 19); ara + ḫ can then be compared to arḫa quoted above. Examples of its occurrence are: —

(A ii, 19) nn-ut-ta ḳat-zun-nš a-ra-aḫ-za-an-ṭa [aš-š]u-u-li ḫar-kan-[d)u ‘ to thee his ḳat for (?) the great god(?) ….’ (G 4, 12, 19) ma-aḫ-ḫa-an ma-aš ar-ḫa la-a-an-zi. (G 14) / ub-na u-zu-nḫ-ri-in-UD-DU-a ar-ḫa ia-ṭa-an-zi (‘I ubna of … to the great god (?) they have given’). (G 16) na-aš ar-ḫa a-ṭa-an-zi ‘this to the great god (?) they have given’. The simple word a-ra-an (?) occurs C iii, 6.

That ara = ‘god’ is therefore only a suggestion: at the same time it is interesting to see the number of personal and place-names beginning with this: Ariamnes and Ariarathes (both names of kings of Cappadocia), Arame (of Bît-Agûsi), Aranda, Ardys, Arnuanta, Arinna (place-name = Boghaz Keui?), Arantu (Orontes), Arpad, Araziki, Argana, Arzašku. It is hardly necessary to compare the Assyrian and Babylonian Bâb-ili, Irba-il, Dur-Aššur, Dur-ili, Dur-Šamaš, Kar-Aššur, &c, for placenames compounded either simply with ‘god ’or a god's name. But, on the other hand, ar is a possible value for the ‘house’-sign (§ 18; note 2), and ar-ḫa may mean simply ‘palace’.

^{page 46 note 2} In M xxiii A, 2 there is a group which might conceivably be read the same way.

^{page 47 note 1} Note, however, that in the proper name (Kat-t-e) it is written thus, to make the distinction (§ 60).

^{page 47 note 2} But besides this initial a we find iand e: e.g. i-ia-an-zi (Y 7, P 4), ia-ṭa-an-zi (G 14), i-ga-it (A i, 27), im-mi-ia-an-zi (G 15), e-ša-at (W 19). This seems to indicate that the sound of was not given a definite equivalent in cuneiform, but that the scribe wrote down the sound as he thought he heard it This is supported by the cuneiform value me-e for mi-.

^{page 48 note 1} Brugmann, , Comp. Grammar of the Indo-G. Languages (tr. Conway and Rouse), § 477.Google Scholar

^{page 48 note 2} See Sayce, Y p. 64.

^{page 49 note 1} Menant arrived at the conclusion that it was a vowel (‘Eléments du Syllabaire hétéen ’, Acad. Des Inscr., xxxiv, 2nd part, 1892, p. 100Google Scholar). He considered that it = a.

^{page 49 note 2} The preponderance of Assyrian words in this line makes it comparatively easy to translate.

^{page 50 note 1} It is possible that this is a sculptor's error for n (‘our’ and not ‘my’), but my copy from the stone and Hogarth's from the cast both read m.

^{page 50 note 2} Proved in § 60.

^{page 50 note 3} Proved in § 68, note.

^{page 50 note 4} Ši-i-e-it Y r. 1, 2 is probably an instance: perhaps e-ša-at W 19.

^{page 51 note 1} Torp, , loc. cit., compares the warat-mu of A i, 18 with bhrātō ‘brother’.Google Scholar

^{page 51 note 2} The ki-i here looks rather like the Assyrian ki-i, a conjunction, ‘that’.

^{page 52 note 1} (or its abbreviation) = u, from the following:—In Hittite cuneiform -u is the termination of the 3rd pers. pl. suffix (see §57), kat-tu, a-ba-u, nu-u-ZUN, n-ul: this is found in hieroglyphs in kat-u, a-b(a)-u, t-a-u, mi-r-a-u, u-zi: the particle a-u which appears to be the of the hieroglyphs (§ 83): the verbal termination -un (§ 71) appears in III. s-u-n: the verbal u- (in u-ṭa-an-zt, &c.) is found in u-b(a)-r-a-t(a) (see translation of M ix at end). It would be clear that since Nis-t and Nis-t-e are found, any addition would probably only be the mark of a case-ending, although this need not be so necessarily.

That is an abbreviation for is obvious from a-b(a)- (M xxi, 4, &c), a- (xxi, 4), and § 61.)

^{page 52 note 2} My hand-copy from the stone has zt, which seems obviously wrong: the inscription was very often difficult to copy with certainty.

^{page 52 note 3} I think Professor Sayce is practically right in translating ‘king’ from the ‘Boss of Tarkondemos’: I have preferred the word ‘lord’, as it appears to me to be a term of respectful address to an equal. The value is apparent from a comparison of two groups: (1) the group (e.g. Rams.; Boghaz Keui, M xxviii: Kolitolu-Yaila, M xxxv, 2: Fraktin, M xxx, A). This group seems to be used chiefly if not entirely in the Western States. (2) used, as an epithet of the king at Fraktin, M xxx, B, and of the king on the Carchemish inscriptions M ix, M xi, and one new one.

Professor Sayce's suggestion that is the royal headdress seems a good one. At any rate, it is not improbable that is closely allied to in meaning: and since we have seen that =‘chiefs’, the plural of (§ 38), we should have the value for these groups ‘lord of chiefs’ or similar meaning, which is exactly what we should expect, the equivalent of the ‘king of kings’ of the Oriental. It is clear that (§ 44) ‘lord of lands’ is not far different from on the ‘Boss of Tarkondemos’

^{page 53 note 4} To avoid a repetition of the character in type, I am using my value nin, always with the reservation that its proof rests on what follows in this section.

^{page 53 note 5} Sayce reads from the stone in Constantinople, but the B.M. cast suggests a parallel to M xvi.

^{page 53 note 6} Emended from a comparison of M lii, 4, with the B. M. cast.

^{page 53 note 7} Read thus for M xxi nas. [The value =nas is shown on M iii, B, 2 where n-s takes the place of this character on M iv, A 2, and iv, B 2: and also on M xxi, 4, where after the ‘chair’ hieroglyph occurs , which is replaced by n-as in a similar passage in M lii, 5.]

^{page 53 note 8} Probably a place-name but not easy to read.

^{page 53 note 9} u or ir: see § 69.

^{page 53 note 10} For Tabal see translation of M xvi at end.

^{page 53 note 11} See § 73 (a).

^{page 54 note 1} These characters are uncommon, and form a group. It seems to me that this must mean ‘the dead’ or some similar phrase, and certainly the whole phrase ‘ally of our dead fathers’ is a most probable one. The, other occurrences of which I know are M xxi, 2, bis, and M lii, 4, where the same meaning is suggested (see translation at end and §87); if this be right, the ideographs might be explained as a burial-shaft and a coffin.

^{page 54 note 2} See § 14 (c).

^{page 54 note 3} I would suggest the name Bar(?)-ḫu which occurs on TA 3, but it is a doubtful reading. See also § 73.

^{page 54 note 4} Character doubtful, but it may be ma, or perhaps the title discussed on p. 77.

^{page 54 note 5} For syntax and nominative of this word, see §§ 66, 84.

^{page 54 note 6} Allowing, of course, for the emendation of my hand-copy mentioned in § 49, note.

^{page 55 note 1} This reading is due to Professor Sayce.

^{page 55 note 2} This character is obvious on the stone in the B.M.

^{page 55 note 3} Rams. 4 appears to contain this name As-ir-zi- ‘country’, ‘in Assyria’.

^{page 56 note 1} Read:? -s(?)-an Atn-k ‘?-san of Amṭ, &c.’ s-an (§ 68, note) appears to be equivalent to : cf. (?) also s-n-zi, TA 6 (see translation at end). Comparable to this name is the name M xxxi. The god-name occurs on M xliii, 8, with ‘god’- Targu, under the winged sun-figure, and hence I have assumed it to be the sign for the sun-god. With regard to Amṭ, I cannot help thinking that U-'m-k is the proper reading on M xxxii, 3.

^{page 57 note 1} Abbreviations in Assyrian are not uncommon: Šuzubu is short for Nergal-ušezib or Mušezib-Marduk. Compare also Pul. Indeed, on M x, it looks very much as though the Hittite king's name also was abbreviated (see translation at end).

^{page 57 note 2} This character occurs or is omitted apparently arbitrarily after the god Tesup's name; it would appear to be the winged disc (see Ramsay's inscription, PSBA. xxxi, 1909, 83). It occurs alone syllabically in M iii, B, 2: vii, 1 (?): xii, 3: TA 6.

^{page 58 note 1} I can only offer a very poor suggestion here, that this name occurs in M viii, 4, M-n-n(a)-m.

^{page 59 note 1} Sayce gives the following forms (Y p. 49): miš or mêš, pl. mâš, ‘mine’, gen. dat. mi (mu), acc. min: taš, liš ‘thine’, gen. dat. ta, ti, tu, acc. tan, tin, pl. tâš; šaš ‘ his’. See also Torp and Bugge in Knudtzon, Die zwei Arzawa-Briefe.

^{page 60 note 1} Knudtzon, , Die El-Amarna Tafeln, pp. 270 ff.: Die zwei Arzawa-Briefe (with additions by Torp and Bugge)Google Scholar.

^{page 61 note 1} A division-mark has been omitted in the hieroglyphs here.

^{page 63 note 1} I copied these two characters an obvious error.

^{page 63 note 2} See § 68, note i.

^{page 64 note 1} The nominative and accusative have long been known; Sayce considers that the genitive-dative case was expressed by a vowel, and that -ša denoted a case of dependency and probably the vocative. ‘The nom. and acc. pl. terminated in -aš and -uš, as well as in -t or -d, but the relation between the two terminations is not yet clear, -an appears to have been the suffix of the gen. pi.’ (Y p. 48). See Torp and Bugge in Knudtzon, Die zwei Arzawa-Briefe.

I believe the dative is best exemplified by (1) AN.IM-aš AN.UD-i bi-i-e-it (Y 21) where the god Hadad is nom. and ‘sun-god’ (i.e. the king?) is in an oblique case, not the accus., after bi-i-e-it, a verb. In Y 9, 17, 26, 31, 34, 37 occurs a word or name ḫa-aḫ-hi-ma-aš of which ḫa-aḫ-ḫi-im-mi is found in Y 38. Certain names occur in the nominative without -š.

^{page 64 note 2} AN.IM makes AN.IM-ni, Y 34.

^{page 65 note 1} This is made certain from the hieroglyphs.

^{page 65 note 2} GAL-GAL-as occurs in the opening phrase of A i, ‘it is well unto the chiefs of my people’.

^{page 65 note 3} Sayce and Jensen both recognized this.

^{page 66 note 1} Are we to regard this n-as as an oblique case of the plural of nis (§ 65)?

^{page 66 note 2} From the comparison of texts it would seem as though were possibly a variant for : cf. M ix, 5 a chief's name, ‘a chief of ’: M lii,5 Gar-[a?]-li(?) ni where the alternative in 1. 4 is IX-a-e: cf. xxi, 4: xxiii A, 2.

^{page 66 note 3} Are we to read thus instead of mi-am on account of the ‘tang’ which calls attention to mi?

^{page 67 note 2} Other possible instances of n assimilated before a sibilant are: cuneiform i-ia-zi (Y r. 39) by the side of i-ia-an-zi (Y 7, P 4), pa-iz-zi (for pa-in-zi?) (S 1, 12: F 1, 3, &c), te-iz-zi (for te-in-zi?) (Y 17, 23, 34, r. 10).

^{page 68 note 1} This is the ‘hand’ sign upright, without distinction of the fingers (as in § 34), the four fingers being merely indicated like a glove and not spread apart. It will be observed that these two forms rarely occur on the same inscription (cf. M. ix, 2,3), and a comparison of texts will show that they are the same character with the same meaning ‘friend’; also on M ii, the similarity of the horizontal hand (‘ally’) with this upright hand will be at once remarked. The form which can almost be called transitional is to be seen in M xi, 2,-‘ally’ -zi a-an-zi ‘they have counted as an ally’ (or, ‘in alliance’). In M ix, 3 we meet this hand three times: ‘I will make sonship (§ 73) with thee “ally ”-r-s u-b(a)-r-a-t(a) nis-zi mi-t: mi: “ally”-n-ui ir-r-a-t(a) “ally ”-zi-t(a):? -e-s-k mi-ni…,? will take thee as a son (in sonship) with me, our allies (alliance) will join thee,, in thy alliance … eskwe (?) accept (?).’ In M ii, 3, 4, what I at first thought was a personal name, reading it as must be explained as ‘ally’ -e-ni ‘ our allies’, with an accus.’ ally’ -ni-u in 1 6. Clearly in 1. 4 we have a plural verb after it, and ‘ally’ -e-ni here is certainly not preceded by a nominative, but rather an accusative from the preceding sentence, whatever its meaning may be: and I should translate this phrase ‘our allies have graven the [leg?] of the memorial stela’. Similarly we must read ‘our allies’ in 1. 3.

^{page 69 note 1} (which is the linear form of as was pointed out by Sayce, PSBA., xxi, 1899, 205) = an is proved (1) from M iii, 2, ḫ-AN-n-s and iv, A, 2 ḫ-AN-nas, varying with iv, B, 2, ḫ-nas: (2) M ii, 4 m-n-u-k-n-AN a-ṭ(a)-t ‘for our memorial (?) he hath given’, and ibid.6 mi-r-a-AN: t-e ‘ before us he saith’: (3) the common n-m-n-AN ‘ our covenant’ compared with n-m-n-a-n once (TA 7); (4) n-m-n-t-an ‘thy covenant’ (M xv, A, 3). The oath in a new Jerabis inscription with a singular subject”takes the form mi-t-m ‘god’ -ḫ-m ‘ (As) my great god is with me’, with a plural subject becomes ‘as for me and So-and-so Tesup-s mi-t- (if my reading is right, which seems in every way probable), i.e. mi-t-an ‘ (as) Tesup is with us ’.

^{page 70 note 1} Cf. ar-mi, translation to M xxi at end, note.

^{page 70 note 2} That this form san(n)-s-t(a) might have an intransitive force is possible, san having all the active meaning of'to make’, ‘to do’. We are too much hampered by lack of examples at present to say that this -s- in san(n)-s-t(a) forms a middle voice, or that the form -s-t(a) for the 2nd person singular termination is comparable to vid-isti in Latin (an 5 Aorist).

^{page 71 note 1} I have not enough examples in the hieroglyphs to say definitely whether this -r sound was used in the singular in unaugmented tenses. Yet the cases te(?)-r-r M viii, 2: te(?)-r-nis (M xxxi): te(?)-r (ibid.): the possibility of a badly written b — ir in M iv, A, B, 2: M vi, 2: M xxxii, 1: and of the word r-k-r (M ii, 6: xi, 5) must not be lost sight of.

^{page 71 note 2} A possible form with a suffix is r-nin-n-t(a), TA 5, &c.

^{page 71 note 3} Are we to see in this san-n-zi t(a)-k-n?si (or san-n-zt-t(a)?-k-n-zi) either ‘they have made (it) for thee … “(or they have made it, they have …), or, still more probably, comparing it with TA 4, ‘they have made [a feast? (takna)]?’

^{page 72 note 1} In M ii, 1 ff. we might see in a-ṭ(a)-t the 3rd pers. plural with the 2nd pers. suffix t(a), ‘Saith Ṭ(a)-?-ar-s unto his son (?) Mutallu “Thy father (and) Benhadad the great have given thee (a-ṭ(a)-t(a)) a memorial-stela (?) for the glory (?) (commemoration (?)) of Tesup ”’.

^{page 73 note 1} On this variant see § 64, note 2.

^{page 74 note 1} This has long been accepted in the Arzawa letters. It rather suggests that might have the value of un in M 1 ‘a bowl for Tesup (as) our covenant he (?) hath given : a-n kat-mi: s-: a-b(a) Bar-k-u which I (have) written (?) with Bark ‘: but the suggestion is without support. (See § 32, note.)

^{page 74 note 2} At the same time in two of the passages in which it occurs it appears to be preceded by a numeral, once ‘ten’and once a ‘hundred’, if Professor Sayce's very plausible explanation of as ‘440’ be correct.

^{page 74 note 3} Ṭa-a-i (Y r. 26, 27, &c, D II, &c, E 7, 12, &c.) has all the appearance of an imperative

^{page 76 note 1} In the long text from Tel Ahmar the name occurs in 1. 8 spelt ordinarily (without Panammi); in 1. 3 we find a group which, although the lower half of the line is broken away beneath, gives at least at first sight some colour to the belief that this group is a variant (‘head’-ḫu) for the name of the brother of Panammi (‘head’-ḫi (?)). Against this, however, I must mention that M xv, B I (§ 3) contains a possibility that this name ‘head’ -ḫu is the father of Sangar, and son of another Sangar: the first character is very difficult to read, but if the line on M xv, B I runs as I have suggested in $ 3, it would be difficult to reconcile this name ‘head’ -ḫu as the brother of Panammi. However this may be, it has, of course, no effect whatever against our reading the name of the brother of Panammi as ‘head’ -ḫu.

^{page 78 note 1} Cf. the inscription of Kalammu referred to in (c).

^{page 79 note 1} A. = occurs in Assyrian inscriptions, H. = Hittite, S. = Sinjerli.

^{page 79 note 2} Father of Panammu I, occurring (if this system of decipherment be correct) in a Hittite inscription (M lii) sent by Benhadad to Mutallu of Gurgum (§§ 28, 87).

^{page 79 note 3} Occurring in Hittite on inscriptions contemporary with Sangar (M xv, B) Irḫulîna (M vi), Muttallu (M xxi?), Katê (M ix?), Kirri (M xi), Aram of Kask (M xi), and Benhadad (M xxi?). As Panammu II died probably not long after 745 B.C. (see the Sinjerli inscription of Bar-Rekub), Panammu I may quite well have been on the throne as early as 845. He must have been contemporary with the last years of Benhadad, who was murdered some time between 846–842.

^{page 79 note 4} Bar-Rekub describes his ancestors as living in the palace of Kalammu.

^{page 80 note 1} The Hittite inscription from Kirtschoglu (M vii), not very far from Kalammu's kingdom, seems to contain the name Bar-ẖi badly written in the first line. (See translations at end.)

^{page 81 note 1} It is a fitting place here to discuss the meaning of gar(kar) (§ 3 ff.); besides occurring in Sangar, Gargamis, Garali, Garsans (?), &c, and with san, as above, we find it in (1) M ii, 2 ‘Thy father (?) (and) Benhadad, the great, have given thee a memorial-stela for the kar of Tesup’; (2) M ix, 5 ‘A great tablet (?) of our kar ir-r-a k … (i.e. they have joined)’. Now this can either be referred to the kar ‘commemoration’ mentioned above, or we can consider it as equivalent to the Assyrian karu, supposed to be the Kar or Gar in Gargamis, ‘a fortress’, ‘wall’. I incline to the former, ‘the commemoration (praise) of Tesup’ being indicated by his figure beside the inscription, ‘the great tablet (?) of our commemoration’ being the actual inscription on which the covenant was written. (On the meaning of the sign ‘tablet’ see § 73, a.)

^{page 85 note 1} A perplexing group which occurs three times is t(a)-a mi-t in the following passages: M vii, I ‘Saith … Make alliance with us (the b(a) of b(a)-a is omitted) t(a)-a mi-t Bar-ḫi (§ 73) “brother”-nas’; M xxxiii, 2 (see translation to M ix at end, notes) te(?) s-e-t(a): t(a)-a: mi-t: ‘wood ‘; ″Ṭ(a)-a-nas; and in a new Jerabis inscr. t(a)-a mi-t-k-n. Are we to regard it as a compound ‘regarding, for, on behalf of’? In the last case it is possible to see in mi-t the reading mi-t(a) ‘with thee’, and as is shown above (M ii, 6) mi-t.. k-n is a proper compound like nu-mu-kan; but at the same time the existence of the word t-k-n must not be forgotten (see translation of M ii, notes at end).

^{page 86 note 1} I have suggested ir instead of the text-reading of as (the ibex head) which is close to a-break in the stone: as-r-a is unintelligible to me, while Ir-r-a is a known name. But this is very doubtful.

^{page 86 note 2} Text has

^{page 86 note 3} Transpose and read ar-k-n-as.

^{page 86 note 4} On this word ar-k see §88.

^{page 86 note 5} Are we to add here the terminations in -r, i. e. Ir-ḫu-(?)li-ni-r M xxiii, 3: ‘enemy’ -a-ar TA 5?

^{page 92 note 1} It is curious to see this same reminiscence of previous alliance appearing in Asa's message to Benhadad (1 Kings xv. 19), ‘There is a league between me and thee, and between my father and thy father’.

^{page 92 note 2} See § 33 for this identification: § 25 for his history.

^{page 92 note 3} See § 24 for his history: his name occurs on the following published monuments—M ii, 1: xix, 2 (?): xxi, i, 2, 5: [xxii]: [xxiii, 2, 3?]: lii, 1, 5 (?): Seal xlii, 5: TA 1.

^{page 92 note 4} Unidentified at present (§ 41, note 1): name occurs M xxi, 2: lii, 1, 4.

^{page 92 note 5} Unidentified at present (§ 49): M v, 4 (?): xxi, 2: lii, 3: TA 4.

^{page 93 note 1} See § 24 for his history: name occurs M lii, 1, 2: TA 1. For the lost a, cf. Gusi (Shalm. Mo., II, 12) with Agūsi (ib., 27). I have used the phrase Bit-Agūsi for his district for convenience. (Cf. Maspero, Les Empires, p. 34.

^{page 93 note 2} See § 73: name occurs M i: vi, 4(?): x, 2: xi, 4: xxi, 3: xxiii, c, 2(?): lii, 1: TA 3, 5.

^{page 93 note 3} Name occurs M xxxii, 1, 2, 4 (?): xxxiii, 3, 12: xxxiv, A, 1: lii, 2.

^{page 93 note 4} Name occurs M xv, B, 2: lii, 2, 4.

^{page 93 note 5} See § 25 for his history: name occurs M iii, B, 1: iv, A, B, 1: vi, T: Restan 1: xvi, A, 1 (?), c, 1 (?): xlvii, 1 (?): xxiii, 3 (?): lii, 2 (?).

^{page 93 note 6} See Sinjerli inscription of Panammu I (Von Luschan, Ausgrabungen), § 11: name occurs M lii, 4, 5(?).

^{page 93 note 7} See § 24 for his history: name occurs M vi, 2 (?): ix, 4: xix, 3 (?), 8: lii, 3 (?), 5.

^{page 93 note 8} See § 56.

^{page 93 note 9} Name occurs M ii, 3 (?): vi, 3: xi, 2: xv, B, 3: TA 2 (?).

^{page 93 note 10} See § 24 ft. for his history, and § 30: name occurs M xi, 4: (? xvi, B): xvi, c, 2: xxi, 1, 3.

^{page 94 note 1} See translation of M xvi, A at end.

^{page 94 note 2} On the difficulty which follows see § 80.

^{page 94 note 3} His name occurs [M iii, B, 2]: iv, A, B, 2: vi, 2: Restan 2: xix, A, 5.

^{page 94 note 4} Unidentified: name occurs M i: v, 1 (?): xxiii, 2.

^{page 94 note 5} I cannot help thinking that the ‘Nine’who are so often mentioned in the hieroglyphs are connected with these ‘Twelve’ in some way.

^{page 94 note 6} Note that Shalmaneser calls himselt mušamḳit ^mátuMuzru u ^mátuUrarṭu on his Til-Barsip inscription (see my article, PSBA., xxxiv, 1912, 72, 10Google Scholar).

^{page 96 note 1} Name occurs M viii, 4: xi, 4, 5.

^{page 96 note 2} Occurs also in TA 5.

^{page 96 note 3} Name occurs M i (as Targu-r-r-s): ix, 2, 4: xi, 1: cf. xi, 2: xii, 4: xxxii, 2.

^{page 96 note 4} See § 73.

^{page 96 note 5} See translation of M x, at end.

^{page 97 note 1} See § 25 for the history: name occurs M xvi, c, 1: Kellekli, § 70.

^{page 97 note 2} The position of the land of Ṣalla appears to be to the north-east of Bît-Adini, and hence is perhaps too remote for us to compare its king Adad-'me (Adadimmi, Adadmi) who paid tribute to Aššurnaṣirpal, with Tesup mis. The names, however, are worth comparison. His name occurs as Tesup-mis M xxxiii, 12: xxxiv, A, I: Tesup-mi-n(a), [M xxxii?]: xxxiii, 3: lii, 2.

^{page 97 note 3} Name occurs M xxxi, c, 3: xxxii, 1, 2, 3: xxxiii, A, 1: xxxiv, A, 2, B, I.

^{page 97 note 4} Occurs M xxxii, 1: xxxiii, 1.

^{page 98 note 1} Can A-r-ar-a-nin-s be the original of the Greek form Ἀριάμνης (the name of two kings of Cappadocia, one the father of Ariarathes I); or should we see the -μνης in the -mina of Tesup-mina, in which case Tesup- would take the place Aria- (= Ara ‘god?’, § 44, note 1)?

^{page 98 note 2} I have taken the grammatical forms from Brugmann's Comp. Gram, of the Indogermanic Languages. I am much indebted to Professor Conway of Manchester University and Mr. Lionel D. Barnett of the British Museum for advice on this matter, and particularly for their timely caution against the danger of making comparisons: they are, of course, not responsible for anything in this section, which is, after all, only a collection of suggestions.

^{page 99 note 1} We have to add to these the following words from Hittite cuneiform: at-ti-iš-ši ‘his, her father (?)’, like a-t(u)(?) above (Y 17) (cf. a-ta-mu, A ii, 1); an-ni-iš-ši ‘his, her mother (?)’ (Y 17) ἀννίς, anus; e-eš-mi, e-eš-tu, Indog. as- ‘to be’ (A i, 7, 10: Y r. 3): ḫat-ra-a, ἓτερος? (or Þa-ra-a A i, 20: ii, 10: B 2: P 16: Y r. 38 (but see § 48)).

^{page 100 note 1} ŠI (=Þa-an) is used thus apparently, in D 18, E 15, O 2.

^{page 102 note 1} Unless this has a syllabic value: see sign-list, No. 8.

^{page 108 note 1} Probably to be read thus.

^{page 108 note 2} Doubtful: I have read the s as ḫu.

^{page 112 note 1} li is fairly plain on the photograph (Recueil, xvii, 25).

^{page 122 note 1} ‘Ana-Aššur-utir-aṣbat … which the Ḫatti call Pitru’ (Ob. 38).

^{page 124 note 1} (bar), certain, but the group may not be Barḫu.

^{page 125 note 1} It cannot be the ⁿ?-ni-li of M xxi 1?

^{page 131 note 1} In this latter case, however, the variants (the name is written five times) show that the name Tesup has no share in the name.

^{page 131 note 2} The seal M xlii, 5 is proof that Mutallu of Gurgum spelt his own name Mu-tal, and that it is not merely the error of a foreigner such as Benhadad.