^{page 83 note 2} Strabo, xiii, 4, 12, P. 629.

^{page 83 note 3} Strabo, ib.

^{page 84 note 1} Studio Pontica, iii, p. 80 f.

^{page 84 note 2} Colonia Niniva was long supposed to be an exceptional colonia situated on alien soil; but the name is now proved to be an error for Ninica in Cilicia Tracheia. Kubitschek, however, would attribute the foundation of Colonia Ninica Claudiopolis to Claudius, long before the province arose; but, though he explicitly dissents in this from my article on Colonia Ninica (Rev. Numism. 1894, p. 164 ff), yet I still adhere to the view stated there that Claudiopolis was founded as a city by Antiochus in honour of Claudius, and Col. Iul. Aug. Claud. Ninica as a colonia by Domitian, named after Iulia Augusta: Kubitschek, , Rundschau über ein Quinquennium d. a. Num., Wien, 1896Google Scholar (see footnote on p. 86).

^{page 84 note 3} In Pauly-Wissowa s.v. Colonia, p. 532.

^{page 85 note 1} This function of the client kings is clearly stated by Strabo, pp. 671, 840; and has been discussed by the present writer, Was Christ Born at Bethlehem ? pp. 120-123

^{page 85 note 2} The use of the name Kaisareia by foreign kings must not be quoted as exemplifying Roman usage. Archelaos of Cappadocia renamed Eusebeia Mazaka between 12 and 9 B.C. calling it Kaisareia. Again Kaisareia Germanica in Bithynia was not a ‘colonia,’ and the time of its foundation is unknown (probably it is named after Julius Caesar) : it was a Greek polis.

^{page 86 note 1} Ninica Claudiopolis offers an excellent parallel. It was renamed Claudiopolis by Antiochus in honour of the emperor; but it continued to be a Greek polls. Kornemann, P.W. s.v. Colonia, 551, follows Kubitschek (see footnote on p. 84). Another example is found in Iconium, which was granted the name Claudiconium by the same emperor, but remained a polis till Hadrian made it Col. Ael. Hadr. Aug. Iconiensium. Kornemann is wrong in this also; and he quotes Eckhel as authority for the false name Col. Claudia Iconium, which has no ancient authority: C.I.G. iii, 3991, 3993.

^{page 86 note 2} Mommsen gives the same list, Res Gestae D. Aug. 1883, p. 119.

^{page 87 note 1} Nothing of this is stated in the article on Antioch in Pauly-Wissowa; but the geographical side of that valuable encyclopaedia is the weakest, so far at least as the eastern provinces are concerned. In these geographical articles the facts are rarely stated in a complete or correct form even from the older authorities, and little attempt is made to co-ordinate and value the evidence. Antioch is reckoned to Pisidia, and Acts is quoted tor this assignment; but Acts rightly understood is of the contrary opinion, and does not speak about ‘Antioch of Pisidia,’ but about ‘Pisidian Antioch.

^{page 88 note 1} An alternative explanation as to which Drusus is meant, although admitted as possible by Mommsen in C.I.L. iii, 6843, has now been disproved by Professor Calder in J.R.S. 1913, p. 100 f.

^{page 88 note 2} Nat. Hist. v, 94.

^{page 88 note 3} Kornemann, p. 532, states the facts exactly and correctly.

^{page 89 note 1} Oa Caesianus see below, p. 95. It is, however, possible that Caesianus is taken from his mother's name according to a common imperial fashion.

^{page 89 note 2} The use of nomina in the place of the cognomen was beginning at this period: examples J.R.S. 1913, P. 254.

^{page 89 note 3} C.I.L. iii, 6824 (Campusius), 6825 (Cissonius), 6828 (Tiberius). I have copied all more than once, Mr. Calder has, I think, seen them all, and Mr. Anderson most or all of them.

^{page 90 note 1} Mr. Cheesman mentioned to me that the legion which lost an eagle in the Jewish war (Suet. Vesp. 4 f) must have been XII Fulm., a conclusive proof that the loss of an eagle did not always terminate the existence of a legion. Leg. V Alaudae existed as late at least as Nero (Dessau, 974, 991); but the older identification of Alaudae and Gallica is not accepted in Dessau's Index (published 1914). The Fifth Legion was in Germany till 69; then in Italy and Moesia, and was annihilated in Moesia in 86–87.

^{page 90 note 2} The Pomponian family at Iconium originated C.A.D. 100, when a Greek citizen of the πόλις took on enfranchisement the name M. Ulpius Pomponius after the emperor and the provincial governor.

^{page 90 note 3} Claudius, among his affected archaisms, re-introduced -ai.

^{page 91 note 1} A form due to Greek would, of course, be not a survival of Italian archaic Latin, but a corruption under foreign influence. I have made the suggestion in the manuscript of a book now preparing that the use of -a instead of -ae in the first decl. dat. C.I.L. iii, 6842, 6856, 6861, at Antioch was due to Greek influence. Mommsen rebels against the use in no. 6856, but it is there: he allows it in the other two places. In no. 6856, NONIAE PAVLLA was actually written and was altered by the stone-cutter, who deleted E: the name is dative.

^{page 91 note 2} Even the existence of a form -ηι in the Greek, dative at Antioch would not produce a Latin dative -ai.

^{page 91 note 3} A son would perhaps be absent serving in a legion, as service in the army was customary for the sons of veterans of the colonia; but even though absent his co-operation would be presumed naturally, as being the heir.

^{page 92 note 1} Dessau, Inscr. Lat. Sel. 1033 (where the office is held by a man of senatorial rank), 8859, etc.

^{page 92 note 2} The interpretation, scribai, q(uaestori) urban(o) v(oto) s(oluto) l(ibenter), though epigraphically posinsible (as no stops are written in the last line) must be rejected. The quaestor in Antioch can hardly bear the title ‘quaestor urban(us)’: Urbs means Rome, and not a provincial town, which was oppidum or civitas. Moreover, the form of the sentence, with a dative but no nominative, could not in that case be defended; the meaning would have to be taken as ‘to Pomponius Niger (he himself made the tomb) in accordance with a vow,’ but this way of expressing the meaning seems too awkward.

^{page 92 note 3} Mr. Hill tells me also of two worn and in decipherable coins purchased at Antioch which seem to correspond perfectly to the Berlin coin: they are countermarked with a figure of Mên holding Victory 1. and leaning r. on long sceptre (Num. Cbron. 1914, p. 305, pl. xix. 10). These prove that CC on the Berlin coin must mean Colonia Caesarea, as in the inscription of Claudius's time: J.R.S. 1914, p. 258.

^{page 93 note 1} Bell. Jud. vii, 5, 3. For the date cf. Chambalu, Philol. xliv. (1885), p. 507ffGoogle Scholar. and Filow, Die Legionen d. Provina Moesia (Klio vi Beiheft), 1906, P. 35.

^{page 93 note 2} This is discussed at length in the introduction to my Hist. Comm. on Galatians, pp. 113–126.

^{page 94 note 1} C.I.L. iii, 6826 and 6827. The veterans are called Cissonius and Coelius, of tribes Sergia and Anlensis. [Col. Caes. was classed to tribe Sergia. Two inscriptions, at least, mention citizens of Collina and Aniensis. These were coloni of the original foundation, who came possessing the ‘civitas.’ One inscription is obviously quite early; the other is at the sanctuary of Mên, where the character of the stone generally used makes the lettering almost useless as evidence of date.]

^{page 94 note 2} The whole type is four standards, the two in the middle supporting eagles (Num. Chron. 1914, p. 303, pl. xix. 6, 7).

^{page 95 note 1} J.R.S. iii (1913), p. 253Google Scholar.

^{page 95 note 2} But the size in my copies, made at different times, cannot be trusted as exact. They vary in scale. Moreover, if the daughter were the devotee who placed the offering and added her father's name, she might write her own in larger letters: the son puts his name larger than the parents in the inscription J.R.S. iii (1913), p. 262Google Scholar.

^{page 95 note 3} An exception in the appeal of Nikias to the soldiers in the siege of Syracuse γυναῖκας καὶ παῖδας καὶ θεοὺς πατρῴους (Thuc. vii, 69), where the climax makes the apparent exception really an example of the general custom.

^{page 96 note 1} In J.R.S. iii (1913), pp. 253ffGoogle Scholar. I add a word of mourning for the loss of two scholars whose work on problems of Asia Minor was just beginning—Mr. Cheesman and Mr. Hunter. The investigation of the history of Asia Minor suffers severely from the death of these two excellent young scholars. Mr. Cheesman had arranged with me, if possible, to go out to co-operate in the excavations at Antioch during the summer of 1915, where his help would have been invaluable.

^{page 96 note 2} The remainder of section iii was added in 1916. Sections iv. ff. belong to the original MS. written in 1914.

^{page 97 note 1} The too early death of J. R. S. Sterrett has been a great sorrow to me and a great loss to Anatolian studies.

^{page 97 note 2} It was probably a stone in the wall of a stoa covered with honorary inscriptions. On such porticos, painted or inscribed, see the account of Apamea Phr. in Ramsay, C.B. ii, pp. 431–434Google Scholar.

^{page 98 note 1} Many cases might be mentioned in which the attempt to combine two independent copies in one text has led to error. The most remarkable and extreme instance was in the treatment by Pauli of the oldest Phrygian inscriptions, where quite astounding results were attained by selection of readings from different copies.

^{page 98 note 2} Hence I took no notice of it in my copy.

^{page 98 note 3} This restoration only suggested itself in 1916 when I was revising the notebook containing the inscriptions copied at Antioch in 1882, as in the intermediate time I had used only the published copy in C.I.L. iii, 6834. It is never safe to prefer a text printed in type, and neglect the original copy.

^{page 99 note 1} Arrius Calpurnius Frontinus Honoratus, consul (6810–2), was patronus of this colony, but not governor of the province. Two of his names, Calpurnius and Frontinus, were borne by earlier governors; and his connexion with the colony may have been hereditary.

^{page 100 note 1} He could hardly be the son of Cornutus Arruntius Aquila, as the latter was of senatorial rank and his son would enter the senatorial cursus honorum: the procurator of Pamphylia was presumably equestrian; but he might perhaps belong to a collateral branch of the family of Arruntius Aquila the senator.

^{page 100 note 2} See Cheesman in J.R.S. iv (1914), P. 258Google Scholar.

^{page 100 note 3} On coins of Augustus COL.CAES. is the name, and C.C. occurs alone; but already in the first century the name is C.C.AN.; under Titus ANT. COL. (Num. Cbron. 1914, pp. 303 f.).

^{page 100 note 4} The idea of secondary capitals lasted long in Roman affairs. In 372 A.D. Basil calls Iconium in relation to Antioch the secondary capital of the newly instituted province Pisidia (371), post maximam prima, Ep. 138. This real fact furnished some real basis to the disputes about the honorary title ‘First,’ for which the cities contended in many provinces, Asia, Bithynia, Cilicia, etc In Macedonia the same quarrel existed, as we know from Acts xvi, 12, and no other authority; but this admirable touch of real life has roused much discussion among the theologians who refuse to accept the ready explanation, and alter the text.

^{page 101 note 1} See Kubitschek, Rückgang des Lat. im Orient, Wiener Studien, xxiv, 2Google Scholar.

^{page 101 note 2} Mommsen suggests a Co[manis], but Comana was 200 miles distant, and the ‘caput viarum’ for Cappadocia was Caesarea-Mazaka, which lay on the road between Colonia Archelais and Comana. The distance on the stone is stated from the colonia. and not from Caesarea, because it stood in the territory of Colonia Archelais.

^{page 101 note 3} Hermes, 1884, pp. 33 ff. A summary of this statement is given in my article in Studia Biblica, iv, p. 36 ff. The province, again, was composed of cities, and a soldier's ‘domus’ was his city, except where the municipal Greek or Roman organisation had not spread, in which case his ‘domus’ was his tribe; but ‘Lycaonian’ did not designate a Roman unit in the province. ‘Lycaonian’ designated a ‘classianus,’ because those troops were originally servile. Mommsen recognises on p. 33 the difficulty caused by the fact that the same term, e.g. Sardus or Thrax or Galata, often was used both for the province and for the country, Roman and non-Roman.

^{page 102 note 1} It is, however, not used in the indexes to Dessau's Inscript. Lat. of which the first vol. has just appeared (1914).

^{page 103 note 1} Ath. Mitt. 1887, p. 181.

^{page 103 note 2} e.g. Staatsrecht, ii, p. 1024, ed. 1, p. 991, ed. 2; ibid. p. 237, ed. 1: in the latter place one kind of procuratorial province, but the idea applies mutatis mutandis to all procuratorial administration.

^{page 103 note 3} Yet this mistake has been made by inexact writers through confused thinking. Pamphyliarch and Lykiarch were each head of a separate κοινόν. The κοινὸν Λυκαόων contained the cities of Lycaonia as one of the Tres Eparchiae (second and third century).

^{page 103 note 4} C.I.L. iii, 471, 7116, group together Galatia and Lycia-Pamphylia. When Tacitus says, Hist. ii, 9, ‘Galatiam ac Pamphyliam provincias Calpurnio Asprenati regendas Galba permisit,’ it has been customary to understand that Asprenas was Galba's legatus. The question may be asked whether he was not merely procurator put in charge of the provinces, during the interregnum that resulted after the death of Corbulo. Lycia was not yet connected with Pamphylia, which was at this time under a procurator (C.I.L. iii, 6737). The connexion with Pamphylia was perhaps not arranged until A.D. 74. X. 7583 f. show that such lists are hardly ever complete, and the omission of one item does not prove that the part omitted was not included in the single administration.

^{page 104 note 1} Tacitus, , Hist., 2. 9;Google ScholarAnn. 15, 6; C.I.G. 3991, refer to the time when Galatia province included. Pisidia and the procuratorial was then probably identical with the imperial province,

^{page 104 note 2} Strabo, p. 569, xii, 6, 3, uses a similar metaphor about another Roman arrangement ἡ Δέρβη μάλιστα τῇ Καππαδοκίą ἐπιπεϕυκός (τυραννεῖον).

^{page 105 note 1} As is pointed out by Mommsen (quoted above, p. 101, n. 3), there are many cases in which the provincial title was the same as the national name, and the two senses have to be carefully distinguished: so in Galatia, Thracia, Sardinia, Corsica, etc. The boundaries of the province sometimes were the same as those of the nation (e.g. Corsica, Sardinia), but more frequently differed.

^{page 105 note 2} Above, p. 87; Church in Rom. Emp. p. 25; St. Paul the Traveller, pp. 103–105; Cities of St. Paul, p. 273ff.

^{page 105 note 3} The remarkable road-system alluded to in the passages quoted in the preceding note implies a definite controlling plan, the idea of which is that Antioch should be the centre of a garrison system. The roads eastward and westward alike bear the same name, Via Sebaste, a remarkable latinisation of the Greek translation of the Latin word Augusta.

^{page 106 note 1} Hissar-ardi, ‘behind Hissar,’ is a village close behind and above the ancient site, Hissar, the fortress. The modern town Yalowadj lies below and in front of Hissar, on the plain.

^{page 106 note 2} Parts of the surface are broken even in these letters; but we considered the text here certain, and in 1914 I revised it.

^{page 106 note 3} [ạ], [ṃ], indicates that a small part of the letter only is preserved consistent with, but not necessitating, the restoration a or m.

^{page 106 note 4} This ala prima miliaria is known otherwise only from Notitia Dign. Or. xxxiv, 36.

^{page 106 note 5} Flamens were part of the colonial constitution. The gymnasiarch is an institution that marks the growing hellenisation of the colonia: its occurrence may perhaps indicate that the inscription should be assigned to the second century (which suits the lettering). The existence of a gymnasiarch at Antioch is attested also by an unpublished inscription.

^{page 106 note 6} J.R.S. iii (1913), p. 254Google Scholar.

^{page 107 note 1} J.R.S. ii (1912), p. 102Google Scholar.

^{page 107 note 2} The orientation is far from exact, being conditioned by the rock background. For east strictness would require ESE.

^{page 107 note 3} At first I was disposed to connect them with the temple, but they are far from it, and the temple exterior was of marble (now almost all lost), whereas the inscription was on limestone.

^{page 108 note 1} Perhaps the explanation of the erection was contained in an appendix (see notes). But we should expect that some more formal decree and statement would be placed at the beginning.

^{page 109 note 1} Colonia Caesarea was intended to be a purely Roman city (see above, § ii), and a Greek copy, therefore, was probably considered out of place, although the great majority of the population were incolae speaking Greek.

^{page 110 note 1} Annual of British School at Athens (1912), p. 65.

^{page 110 note 2} Too late the excellent edition of Cagnat-Lafaye comes before me.

^{page 110 note 3} See note to par. 24.

^{page 111 note 1} Written in July, 1914.

^{page 122 note 1} Sir J. E. Sandys notes that this is a very rare construction, lege being much more common; but that the evidence compels us to accept per legem and to assume that this construction was directly due to Augustus himself (in which opinion Professor J. S. Reid agrees).

^{page 123 note 1} So here; but quingenties 16.

^{page 123 note 2} It is doubtful whether m was omitted, or left over to the next line.

^{page 123 note 3} The space seems broad for xv, but narrow for quindecim in full.

^{page 124 note 1} This hope, written in July, 1914,is left standing as a protest against the absurd and ruinous war which was forced on Europe.

^{page 125 note 1} In the facsimile the line above PIAM should be marked strongly and clearly, indicating the top edge of the stone. Above XIM the edge is not certain, as the fragment is small and broken.

^{page 126 note 1} There still remained some days of the time specified in our permit; but perhaps the Vali might justify his action by the plea that the years should be counted as lunar, not as solar, in which case the irade had lapsed.

^{page 126 note 2} Sir J. E. Sandyson 22nd Nov. 1917, read before the Cambridge Philological Society a paper on the restoration of the lacunae in this paragraph, showing that the Supplement (1890) to Sir John Evans's British coins (1864), supported by other new evidence, confirmed his preference for Tincommius as the true restoration of the text. This evidence (naturally unknown to Mommsen in his edition of 1883) has: not been taken into consideration by more recent editors or commentators on the text of Anc. This paper has been published in the Num. Chron., 1918.

^{page 127 note 1} Perhaps a better place for this fragment may be found. In that case the reason for supposing a new pagina here would disappear; and a new pagina would begin at the line to which ART is transferred.

^{page 127 note 2} Ant. has quadrigis where Anc. has quadrigeis.

^{page 128 note 1} Perhaps I may have mentioned to him my conjecture, but I cannot remember.

^{page 129 note 1} This was wrongly adopted in Hist. Geogr. A.M. p. 302. Capersana with bridge over the Euphrates (Amm. 18, 8, 1) and Capharda (south of Antioch, identified by Muralt, Chron. Byzant. A.D. 1098 with Cappareas) contain the name Kaper or Kaphar (village). [Itin. Anton. has Caperturi between Antioch and Apamea Syr.—J.S.R.]

^{page 130 note 1} Since the text was printed, I have been tracing the old hieratic families at various Anatolian religious sanctuaries, and feel growing confidence that the suggestion in the text is right. Augustus did not destroy or impoverish the great priestly dynastic family at Antioch, but merely took from it the chief priesthood, leaving it wealthy and influential in the colony: J.H.S. 1918, pp. 146–8.

^{page 131 note 1} See Röm. Mitt. 1891, p. 338.

^{page 131 note 2} See Prosop. where Borghesi's identification of Aemilianus is rejected.

^{page 131 note 3} The same is the case in a filth inscription of this class, Calder, in J.R.S. ii (1912), no. 40Google Scholar (see below app. II, no. 40).

^{page 132 note 1} We verified it in 1908.

^{page 132 note 2} Ptolemy was confused by employing two authorities and not comprehending rightly their relation to each other. He actually gives Antioch twice, in Galatia and in Pamphylia, though C. Müller cuts out the latter.

^{page 133 note 1} C.I.L. iii, 6815, 6816.

^{page 134 note 1} C.I.L. iii, 6831.