¹ The name of the village is wrongly given on the new G.S. map as Asar Köy. Asar (literally ‘works, monuments’) is a common place-name in Turkey, being applied, like hisar and kale, to many ancient sites; and the acropolis hill at Arsa is, in fact, called Asar Tepe. But the name of the village is Arsa.

² This ‘coursed polygonal’ style seems characteristic of the early Hellenistic period; see Scranton, , Greek Walls, 52, 69, 165Google Scholar.

³ Reliefs representing the dead man on horseback, generally as a warrior, are of course common; one such, very similar in general style to the present relief, is that on the tomb (of Alcetas?) at Termessus, illustrated in Lanckoronski II, p. 67, fig. 19.

⁴ See Robert, L.Hellenica III (1946), ch. 3Google Scholar. M. Rober has cleared up much confusion regarding Kakasbos and his relation to the other equestrian deities, and has shown that Kakasbos never carries a sword and always carries a club At Tlos, a few miles from Arsada, on the north side of the acropolis hill a little below the summit, is another equestrian relief, much damaged, which I have not seen mentioned, The horse is stationary, with long tail falling vertically nearly to the ground; his head, and the rider's body above the waist, have been hacked away. Here again there is no tomb, and no visible inscription.

⁵ I saw nothing of the Lycian inscriptions reported by Spratt and Forbes. Nothing, in fact, seems ever to have been seen of them since.

⁶ Supposedly a corruption of Bolsulu, ‘the place of abundant water.’ The ground is, in fact, abundantly irrigated from the river. The whole of the Xanthus valley is, or could be, exceedingly fertile, but in the absence of communication with Fethiye, the villagers grow no more than enough for their own needs, and the land is much neglected.

⁷ I was told by the villagers that some thirty years ago the statue was seen by two foreigners, who wished to remove it, but were unable owing to its weight. No mention of it appears to have been published. The figure was transferred in 1947 to Fethiye, where it now stands in the garden of the school.

⁸ The stone, when I found it, was in use as a washing-board outside the house of Salih Öztürk. [Compare the case of the Aspendus decree published by Paribeni and Romanelli, in Mon. Ant. XXIII (1914)Google Scholar. A second version of this decree has recently come to light, and will be published shortly.] We had it moved into the house for safer keeping. When I revisited Ören Köyü in 1947, the house was locked up and the owner away at the yayla; the stone was presumed to be still inside. In the yard of the same house I saw also a broken table of liquid measures, with the usual hemispherical depressions, each with a hole in the bottom.

⁹ τὸ α′ might indicate the first year of a priesthood held for a term of years; but the article is surely indispensable. In some parts of Asia Minor ὁ δεῑνα α′ occurs apparently in the sense of the familiar ὁ δεῑνα β′, e.g., at Iasus (REG VI 195, Βάλακρος α′); but this again is unsatisfactory here. I mention the slight uncertainty of reading because the use of the abbreviation μ(ηνός) or μη(νός) might be thought to have some bearing on the date of the inscription (see below).

¹⁰ Holleaux, REG XXXV, 200Google Scholar = Ét. d'Ép. et d'Hist. Gr. III, 3Google Scholar.

^10a Army officers' pay at Athens is discussed by Larsen in Cl. Phil. XLI (1946), 91 sqq.Google Scholar For army pay in general see the references collected there (n. 4). What rank, if any, ἔφιππος may denote, I do not know; but it is evident from the use of δωρεάν here that Orthagoras was not in receipt of a regular salary, but would normally have been paid for the period of actual service. This agrees with Larsen's conclusions as to pay at Athens.

¹¹ Of the nameless ancient sites known in the neighbourhood, the most attractive is perhaps that at Koristan, briefly described in TAM II, 263. I visited this site in September 1947. It lies back from the Xanthus valley, 1200 m. above sea-level, on a small hill (Fig. 11) overlooking a lateral valley, some five or six hours from Araxa. The top of the hill is enclosed by a wall of tooled-face polygonal masonry of rather striking appearance (Fig. 12); many of the blocks are large, one measuring as much as 1·85 × 1·82 m. in maximum dimensions. (For this rather unusual style of masonry see Scranton, Greek Walls, 47, 165Google Scholar). The north side of the hill falls in an almost vertical precipice for several hundred metres to the side valley; along its edge are a number of rock-cut houses. On the south slope are several ‘Gothic’ sarcophagi (Fig. 11), one of which has an illegible inscription in letters 33 mm. high, apparently of Roman date; I saw no other inscription on the site. In the face of the hill are several ‘pigeon-hole tombs and the more ornamental tomb described in TAM loc. cit. (Fig. 14). The space enclosed by the wall, some 150 m. in diameter, comprises two slight eminences, of which that on the north-west carries the ruins of a small temple; the style of its decoration may be judged from Fig. 13. Figs. 12–15 are given here in response to L. Robert's complaint (Ét. Ép. et Phil., 166) of the inadequate photographic documentation of Asia Minor; in the total absence of any positive evidence, an identification with Orloanda would be in the highest degree hazardous. At the two sites to the north-west of Araxa, Gürmedere Köyü, which I also visited, and Hisar (Dereköy and Mesenis respectively on the map attached to TAM I), there seems to be nothing earlier than mediaeval, with the exception of a built tomb at the former site well constructed of good squared blocks, now rapidly disappearing for building purposes; no inscription visible. I have not seen the site at Cokek Asari (TAM 11, 263).

¹² Treuber, Gesch. der Lykier, 223, n. 4Google Scholar. Cf. especially IGR III, 563 = Ditt. OGI 556, , dated by Dittenberger ca. 40 B.C., by Cagnat ca. A.D. 40. I do not know what the games at Xanthus ‘in honour of Apollo Patrous and the ancestral gods,’ mentioned in Cagnat's note ad loc., have to do with the cult of Rome.

¹³ From IGR I, 61 = Ditt. OGI 551 it appears that the Lycians dedicated a statue of Dea Roma at Rome, with an inscription expressing their gratitude for the recovery of their democracy. This is presumably the same Dea Roma Epiphanes mentioned here. The date of this dedication is disputed; Dittenberger, loc. cit., accepts Mommsen's dating to the time of Murena; Treuber, p. 168, questions this, but without declaring in favour of any other; the recent tendency has been to place the date earlier, preferably about 167 B.C., when the Senate declared Lycia free after the war with Rhodes: so Magie, , Anal. Studies Buckler, p. 176Google Scholar, n. 1; Larsen, , Cl. Phil. XL, 2, 88Google Scholar; Jones, on the other hand (C.E.R.P., 404, n. 13), accepts the Sullan dating. But there is no necessity that the dedication in Rome should be contemporary with the foundation of the cult in Lycia, though it is not likely to be earlier. See further n. 28 and n. 40 below.

¹⁴ In the following discussion I have had the benefit of the advice, generously given, of Prof. A. H. M. Jones, M. Louis Robert, Dr. J. A. O. Larsen, and Mr. Peter Fraser. Prof. Jones, in particular, has been good enough to send me a reasoned and detailed exposition of his view of the problem. To all these scholars I am sincerely grateful. In view of the abundance of incident and proper names in the inscription, it is really remarkable that the question of the date should present such difficulty. I am conscious of the inadequacy of the commentary I am able to offer here, but have thought it right to publish the inscription without further delay, and so make it available for general discussion.

¹⁵ Strabo XIII, 631. I assume that the incidents in the inscription are narrated, at least approximately, in chronological order, and that the whole public career of Orthagoras is recorded (, 1. 7). That this should begin with his election as ambassador I take to be evidence of his outstanding qualities rather than of maturity in years.

^15a It is enough to refer to the lettering of the Aspendus decree (third century B.C.) mentioned above (n. 8). See the photograph there given on Pl. II.

¹⁶ Wilhelm, A. in Sitzungsb. Wien Akad. CCXXIV, 1, (1946) 71–2Google Scholar, says merely, ‘eine in griechischen Inschriften hellenistischer Zeit vielfach begegnende Form des Pei … dessen Strich ein ausgesprochene Bogen, ein Halbkreis ist.’ He refers to Breccia, E., Iscrizioni greche e latine, tav, I, 3Google Scholar (Egypt, time of Ptolemy Philadelphus) and XLIV, 104. Another example in L. Robert, Coll. Froehner, no. 73, Pl. XXXVII (Egypt, 172–169 B.C.).

¹⁷ Statistics for the present inscription are: (a) Dative iota, added four times, omitted seventeen times, (b) Assimilation: (i) between words, only ἐμ πολοῑς (27), never with the article, eight cases of non-assimilation (13 bis, 17, 21, 34, 37, 53, 71); (ii) in the same word, three cases of assimilation (19, 53, 67: 57 is doubtful), four cases of non-assimilation (14, 60 before gutturals, 14, 62 before labials, and cf. παραλήνψεως in 44). (c) Movable nu: added before consonants thirteen times (three times before a pause), never omitted, (d) Itacism: ἡμεῑν (8, 49: in 29 the termination is missing), ὑπομίνας (11), ἀστυΥιτόνων (54); on the other hand, πολίτης is correctly spelt (4, 17, cf. 57, 62). The stems νικ- and τιμ- do not occur.

¹⁸ The forms with πο- are naturally suggestive of an early date; but a similar variation occurs e.g., in GDI 3087 = Ditt. Syll. ³, 709 in the first century B.C. (time of first Mithridatic War, two cases of πο-, one of ποι-).

¹⁹ Remembering that a quite uncertain amount is missing at the end.

²⁰ The series of drachrns illustrated in BMC. Phrygia, nos. 6–15, bears on the reverse ΚΙΒΥΡΑΤΩΝ with various names, mostly abbreviated or in the form of monograms; one of these is ΜΟΑΓΕΤΗ[Σ]. Imhoof-Blumer, , Kleinas. Münzen I, 250–1Google Scholar, Head, B. V. in Hist. Mum.¹, 560–1Google Scholar and in BMC Phrygia, p. xlvi, where he observes that the names are too numerous to have been all of reigning dynasts.

²¹ Dindorf Teubner text 1868; Müller, FHG II, p. xviiGoogle Scholar; Feder, Excerpta e Polyb. et Diod.; Boor, de, Exc. Const. 205, no. 40Google Scholar; TAM III, 1, p. 4, no. 23.

²² On the reliability of the order of the extracts for purposes of dating, see Holleaux, REA XXIX 91Google ScholarÉt. d'Ép. et d'Hist. Gr. III, 269Google Scholar.

²³ So, recently, by Heberdey in TAM loc. cit., with note of variants, and by Larsen, in Cl. Phil. XL, 2, p. 80, n. 80Google Scholar. The name Moagetes is also corrupted to ΜοαΥέστης or ΜοΥέτης in certain MSS. of Polybl. XXI, 34.

²⁴ Araxa, being a border city, would naturally be particularly concerned in any territorial adjustments; the above consideration would therefore apply equally in the case, e.g., of the formation of the province Asia.

²⁵ Such a state of affairs was already suggested by Larsen, in Cl. Phil. XL, 2, p. 80Google Scholar; the present inscription most happily confirms, as it seems to me, the views there expressed.

²⁶ Strabo's words (xiii, 631 ) suggest a reasonably long period; but it is uncertain whether the subject is the tetrapolis or Cibyra.

²⁷ In a private letter, the substance of which he has kindly permitted me to set forth here.

²⁸ This is, if I am not mistaken, the principal reason for dating the inscription IGR I, 61 (see above n. 13) shortly after 167 B.C., in preference to Mommsen's dating to the first Mithridatic War. Now it is evidently unlikely that the Lycians would erect a statue of Dea Roma in Rome without having a cult of Dea Roma in Lycia. Assuming, therefore, the identity of Dea Roma there with Dea Roma Epiphanes here, IGR I, 61 can be very little, if at all, earlier than the present inscription. If, then, the date 167 is right for IGR I, 61, the Araxan decree should be either contemporary or, as Professor Jones proposes, earlier. The difficulties in either of these suppositions are mentioned below; and if they are thought valid, it will probably be necessary to revert to Mommsen's dating of IGR I, 61. See also n. 40 below.

²⁹ Evidently only incidental mentions of Romans are to be taken into account; in honorary decrees and similar official documents, the simple praenomen is obviously out of the question. The following examples seem relevant: (1) Ditt. Syll. ³, 591 (Lampsacus 196–5 B.C.), l. 68, : so in l. 17 the simple praenomen Λευκίωι is restored; the reference is to T. and L. Quinctius Flamininus; (2) Inscr. de De'los, 442 B (Demares inventory ca. 180 B.C.), l. 34, , ll. 85–6, , (again T. and L. Quinctius Flamininus), (A. Atilius Serranus): in the actual dedication the gentilicia were used in the ordinary way, but the summary inventory is content with the praenomina: see Holleaux, M., , pp. 150Google Scholar sq. Later in the century the gentilicia appear regularly: (3) Inscr. Cret. iii (1942), 91 sq.Google Scholar, nos. 9–10 = Ditt. Syll. ³, 685 (ca. 140–130 B.C.), l. 49, , , a strict parallel to the Araxa text; so in l. 74, where the variant reading does not affect the present point; (4) Ditt. Syll. ³, 683 (similar date), ll. 43, 54, 64, casual mention of and ; (5) Ditt. Syll. ³, 748 (Gythium first century B.C.), numerous incidental mentions of Romans, all with gentilicia.

³⁰ I am deeply sensible of his kindness in allowing me to quote and, with apparent ingratitude, to criticise his views in this article.

³¹ Polyb. XXI, 35, Livy XXXVIII, 15.

³² XXXVII, 55.

³³ Polyb. XXII, 5, .

³⁴ They had at least begun by 181 B.C. (Polyb. XXIV, fin.)

³⁵ So Professor Jones himself in C.E.R.P., p. 101: ‘the Lycian League, which had long been in abeyance, now sprang into life once more.

³⁶ Ditt. OGI 99. See also the evidence from victor-lists collected by Larsen, in Cl. Phil. XL, 2, p. 72Google Scholar, which has equally little significance for the present purpose.

³⁷ It is perhaps worth recalling that Polybius V, 90 mentions a Lysanias as one of three dynasts who sent aid to Rhodes after the great earthquake. Date not later than 226 B.C. (in spite of the chronological sequence in Polybius: see Holleaux, REG 1923, 488Google Scholar sqq. = Ét. d'Ép. et d'Hist. Gr. I, 452 sqq.Google Scholar).

³⁸ That the tyrant's name is written Εὔδημος, not Εὔδημος, is naturally without significance. The name Eudemus occurs at Xanthus (TAM II, 288) and at Patara (id., 429), Lysanias at Tlos (id., 592, 602).

³⁹ Polyb. XXV, 4, (177 B.C.); GDI 3788, , if indeed this inscription dates from 189–167 B.C.: see Holleaux, , Ét. d'Ép. et d'Hist. Gr. I, 394, n. 1Google Scholar.

⁴⁰ That at this period a certain Pancrates was tyrant of Cibyra (Polyb. XXX, 9) is not in itself a difficulty; the events at the beginning of the inscription may, and probably do, precede those at the end by many years. Reverting once more to IGR I, 61 (above, nn. 13, 28), it seems difficult to suppose that a dedication by the Lycian League as should be contemporary with our decree, in which there is no suggestion whatever that the Lycians have recently lost their democracy. If a date near 167 B.C. is to be given to both, he difficulty is indeed formidable. If both are dated after the Mithridatic War, the question arises, in what sense the Lycians had lost their democracy at that time. There is some suggestion that in 102 B.C. Lycia was placed under the authority of the governor of the new province of Cilicia (see below), though this authority, if it extended at all beyond the eastern coast, was hardly more than nominal, the League continuing to function as before; the dedication IGR I, 61 would then be in gratitude for the formal recovery of a freedom that de facto had never been lost. This is the view taken by Jones in C.E.R.P., p. 105 and 404, n. 13, and there is nothing in it inconsistent with our present inscription. See the following paragraph. The exact situation in Lycia during the Mithridatic War is not very clear. Appian begins by saying (Mith. 20) that Mithridates ὑπηΥάΥετο Λυκίαν, and a little later (id., 27) the king lands at Patara and cuts the timber at the Letoum; but the Lycians were resisting all the time (id., 21, 24, 27), and the country was evidently not in his hands; it hardly seems that he could be said to have deprived Lycia of its democracy.

⁴¹ I hardly know whether any inference can fairly be drawn from the expression ‘Moagetes and the Bubonians’ in ll. 9, 12 as to his relation to the city. The words might seem to suggest a less close association than that of a tyrant and the people under him; compare the remarks of Holleaux, M. in BCH 1907, 97–8Google Scholar = Ét. d'Ép. et d'Hist. Gr. III, 57–8, with reference to the expression as opposed to .

⁴² Verr. II, 1, 95.

⁴³ Anat. Studies Buckler, p. 300, n. 2.

⁴⁴ In Fouilles de Delphes III, iv, 37Google Scholar are associated in some way with the newly formed province of Cilicia; but the text is quite fragmentary.

⁴⁵ So Jones C.E.R.P., p. 404, n. 13.

⁴⁶ The name is alternatively given as Gürdef or Girdev.

⁴⁷ In the inscription no. 187 in Reisen loc. cit. the fine is made payable to Oenoanda.

⁴⁸ G on the plan in TAM II, p. 142.

⁴⁹ See TAM II, p. 141.

⁵⁰ Rather surprisingly, no other traces either of the oracle or of the temple of Apollo have come to light, and Beaufort's view has accordingly met with more approval than it deserves. It appears from Hdt. 1, 182 that the oracle was in, or at least attached to, the temple.

⁵¹ We went from dawn till late afternoon before obtaining a drink at a spring outside the north gate; and this was by far the worst water I have tasted in Anatolia. Remains of two aqueducts are visible in the neighbourhood, one on the road from the Letoum to Gâvuraǧlı (? Pydna), the other on the hills above Kalkan (Kalamaki), illustrated by Texier. One or both of these presumably fed Patara, though no remains survive in the immediate vicinity of the city.

⁵² The side-walls may of course have risen higher than they do now. At present they are flush with the ground.

Additional note to no. 11, l. 71. While the present article was in the press, I received Vol. VI of L. Robert's Hellenica. On p. 50 the author publishes a dedication to , discovered by him at Sazoba in Lydia in the summer of 1946, and illustrated on p. 51 by a reference to the present decree from Araxa. M. Robert notes that the epithet Epiphanes was previously unknown or the Senate or for Dea Roma, and suggests (p. 52) that this dedication (from Hierocaesarea) should possibly be dated to 23 A.D., in gratitude for the confirmation of the right of asylum. A further reference to the Araxa inscription on p. 91, n. 4.