^{page 1 note 1} In the notes to this article, the following abbreviations are used. I.G.R.R.=Inscriptiones graecae ad res Romanas pertinentes (Paris, 1906—); I.L.S.=Dessau, H., Inscriptiones Latinae Selectae (Berlin, 1892–1916)Google Scholar; I.L.Al.=Gsell, S., Inscriptions latines de l'Algérie I (Paris, 1922)Google Scholar; Dio. Exc. Ug.=Excerpta Ursiniana de legatis exterarum gentium ad Romanos; Dio. Exc. Ur.=Excerpta Ursiniana de legatis Rom. ad exteras gentes.

^{page 1 note 2} ὲνιαύσιος ἦν ἡμέρα, έν ᾗ Τραιανὸς έπὶ διαδοχῆ τῆς Ῥωμαίων άρχῆς ὑρχῆς ὑπο τοῦ πατρὸς Νερούα εὶσεποιήθη if Roos (Studia Arrianea, p. 32), followed by Jacoby (F. Gr. Hist. BD. 575), are right in referring this to Arrian's Parthica.

^{page 1 note 3} Dio, Exc. Ug., 51.

^{page 2 note 1} Dio, l.c. He must mean Seleuceia in Cilicia. Probably Trajan sailed from here to the Pierian Seleuceia, the port of Antioch. The diploma published in A.J.A. for 1926 concerns some seamen serving under Q. Marcius Turbo in the first half of 114, and it is there suggested that they may have formed part of a fleet which conveyed or escorted Trajan to Syria (o.c. pp. 418–421).

^{page 2 note 2} Mommsen, , Provinces (Eng. Trans., 1909), vol. ii, p. 66.Google Scholar

^{page 2 note 3} It has found champions, e.g. G. A. Harrer in Studies in the Roman Province of Syria, p. 23: but the reasons which make it impossible are too well established to need repetition.

^{page 2 note 4} Lightfoot, , Apostolic Fathers, Part II, vol. II, pp. 413 ffGoogle Scholar.

^{page 3 note 1} Trajan's Armenian and Parthian Wars (Cambridge, 1897), pp. 77–104Google Scholar.

^{page 3 note 2} Five Roman Emperors (Cambridge, 1927), pp. 318–330Google Scholar.

^{page 3 note 3} Optimus Princeps (2 vols. Messina, 1926), ii, pp. 293–300Google Scholar.

^{page 4 note 1} Malalas' evidence on the general history of the war is by most regarded as worthless (e.g. Mommsen o.c., p. 69). But, as an Antiochene, his remarks about Antioch deserve closer attention: Sills says his date is ‘probably derived from contemporary Antiochene annals.’ He does not, however, seem to be any more trustworthy about Antiochene than about external history (cf. Lightfoot o.c., pp. 437 ff. and see below pp. 29 ff.).

^{page 4 note 2} In Dierauer's, Beiträge zu einer kritischen Geschichte Trajans, p. 154, n. 4Google Scholar.

^{page 4 note 3} This suited Mommsen, who dated his arrival in the East to 115: but it does not suit the tale of later historians who have rightly accepted the winter 113–4 for this event.

^{page 4 note 4} See below, p. 10.

^{page 4 note 5} Rev. arch. 1911, p. 486Google Scholar. This man is L. Catilius Cn. f. Clau. Severus Iulianus Claudius Reginus, who was subsequently governor of Armenia with Cappadocia and was promoted to Syria by Hadrian.

^{page 4 note 6} This date is accepted without more ado by Groag in P.-W. xiii, col. 1879.

^{page 5 note 1} A different objection has often been stated: e.g. by Sills (o.c. pp. 83–89), de la Berge (Essai sur le règne de Trajan. Paris, 1877), etc. ‘Puisque Pedon etait à Antioche, c'était qu'il n'était plus en fonction: pendant l'exercice de sa charge il ne pouvait quitter Rome’(de la Berge, o.c. p. 175 n.). This objection cannot stand. Setting aside other examples, we have evidence that all the generals during the Parthian wars whose names we know, except Trajan himself, Hadrian (who was consul in 118 without returning to Rome), and Appius Maximus Santra—who was not a success—held the consulship during this period: and there is little to recommend the view that they were sent home to hold it. The modern practice by which wars are largely directed from home affords no parallel. In no sense was Rome the headquarters for the prosecution of the Parthian war. Rather was the headquarters of the Roman government where Trajan was. The consulship was bestowed by him as a reward for distinguished service.

^{page 5 note 2} C.I.L. xiii, 6798; I.L.AI. 3978.

^{page 5 note 3} C.I.L. ix, 5894; x, 6887.

^{page 5 note 4} C.I.L. xi, 6622; ii, 5543.

^{page 5 note 5} I.G.R.R. iv, 172.

^{page 5 note 6} C.I.L. iii, p. 870. Cf. also iii, p. 2328⁶⁷ to Pannonian auxiliaries. C.I.L. iii, 7537, an official inscription of Tomi, reads TR.P. XXI, IMP. XII, which must be an error. It looks as if the provincial headquarters staff here were under a misapprehension and had supplied false information to the people of Tropaeum Traiani, if the new reading of C.I.L. iii, 12470 as TR.P. XXI, IMP. XII be correct.

^{page 6 note 1} C.I.L. ii, 2097=I.L.S. 297. Cohen no. 177=Mattingly and Sydenham no. 322. Others with Imp. VIII and VIIII lack the title: Mattingly and Sydenham nos. 655, 657=Cohen 176, 178.

^{page 6 note 2} Of these C.I.L. ii, 5543 gives special emphasis by assigning it a line to itself.

^{page 6 note 3} According to Vogt (Alexandrinische Münzen, i, 66) it also occurs on coins of 114–5. This is also the case with coins of Laodicea (B.M.C. Galatia, etc, P. 253).

^{page 6 note 4} E.g. Beroea, where there are eight successive issues (according to Wroth's interpretation of the coin-marks (B.M.C. Galatia, etc. li and 130–131), all of which show the title.

^{page 6 note 5} The remoter parts of Egypt were often behind the times: cf. I.G.R.R. i, 1371, Wilcken, Griech.- Ostr., pp. 799 ff. There are two papyri which are also objectionable. But in P. Ryl. 191 the reading Δακικοῦ άθῦρ 4—i.e. omission of the title Parthicus on November 4, 115, is only conjectural: and in any case the papyrus was itself written in 117. In P. Oxy. 74, however, a registration of sheep and goats, the title Parthicus is lacking on January 27, 116. Papyri are, however, untrustworthy evidence, e.g. the title Optimus is missing in a declaration of municipal bakers of October 28, 116 (P. Oxy. 1454); and many such instances could be given. Inscriptions, of course, accidentally omit well-established titles from time to time: such omissions are Dacicus in I.L.Al. 236, C.I.L. ii, 4797 (where other milestones of the same series all have it); Optimus in I.L.S. 303 (which is full of errors), I.L.Al. 2829, 2989 (if the tribunician number is right).

^{page 7 note 1} Turner, C. H. in Journ. Theol. Studies, i, pp. 187–192.Google Scholar

^{page 7 note 2} When, however, Eusebius gives an exact date in the Ecclesiastical History it is usually accurate by ordinary reckoning. Turner says ‘Eusebius will be found to repeat in the History the exact dating of the Chronicle when, and only when, it was more than guess-work,’ l.c. p. 199.

^{page 7 note 3} Dio, Exc. Ug. 53.

^{page 7 note 4} Cohen no. 39=Mattingly and Sydenham The Roman Imperial Coinage II, Trajan no. 642, where the legend-reference needs correcting.

^{page 7 note 5} The news was not common property in Rome on January 28. C.I.L. vi, 43.

^{page 8 note 1} See below, p. 13, n. 1.

^{page 8 note 2} Cf. Gutschmid, von, Mén. de l'acad. imp. de St. Petersbourg, 1887, p. 18.Google Scholar One might also mention the storm on the Tigris (Xiph. 239), which if the revolt belongs to 115 must be put in the summer of that year. Such a phenomenon is not impossible in Mesopotamia in summer, witness the accident to the Chesney expedition on May 21, 1836; but, in combination with a high tide, it seems to consort better with the autumn, when strong westerly winds blow and the river is low and therefore dangerous.

^{page 8 note 3} Before August, at any rate. Cf. C.I.L. iii, p. 870.

^{page 8 note 4} In Cappadocia: Suet. Vesp. 8. Probably at Melitene until 89 and thereafter perhaps at Satala.

^{page 9 note 1} Had Trajan been at Samosata a meeting would have been almost inevitable. But they do not meet until the end of 114. Mommsen supports his thesis by Fronto (ed. Naber pp. 206 f.), but this refers to L. Verus' campaigns. Samosata does not lie on any convenient route from Antioch to Melitene, which is the obvious headquarters for an invasion of Armenia and where Trajan did actually go.

^{page 9 note 2} There is a hint that it may be right in the MSS., see Dio, ed. Boissevain, iii, 207.

^{page 9 note 3} N.H. vi, 26.

^{page 9 note 4} Anderson, in J.H.S. xvii, 25Google Scholar.

^{page 9 note 5} I.L.S. 232.

^{page 9 note 6} Procopius de Aedif. iii, 4.

^{page 9 note 7} Traces of their passage are found in I.G.R.R. iii, 173, from Ancyra, where Julius Severus is honoured for having received them.

^{page 9 note 8} Dio—or Xiphilinus—singles out Anchialus, king of the Heniochi and Machelones, who was rewarded with gifts, what for we do not know—unless an answer can be supplied out of Suidas' entry about the Lazi, under the heading Δομετιανός or Jordanes, Romana, 267. The Lazi were neighbours of Anchialus. We learn besides of Julianus, king of the Apsilae, and no doubt the contemporary rulers of all the other tribes that Arrian mentions in Perip. II also came: and Eutropius mentions the kings of the Albani, Iberi, Colchi, Bosporani and Sarmatae. The last two peoples although living on the north of the Black Sea may have been summoned to give pledges for their good behaviour during Trajan's absence on campaign. On Ardaches, see below, p. 17, n. 4.

^{page 9 note 9} Parthamasiris had been summoned to Satala. For a discussion of his real position (see below, p. 25).

^{page 10 note 1} Optimus Princeps ii, 293.

^{page 10 note 2} Suidas s. v. γνῶσις Άρριανὸς ὲν Παρθικοῖς περὶ Παρθαμασίρου δὲ οὐχι Άξιδάρου εἶναι, άλλά έαυτοῦ τῆν γνῶσιν, ὅτι πρῶτος παραβαίνων τὰ ξυγκείμενα ἔτυχε τῆς δίκης Fronto, ed. Naber, p. 209, cf. Suidas s.v. Παραβαλών

^{page 10 note 3} Xiphilinus 235.

^{page 10 note 4} This is generally assumed to be proved by a combination of C.I.L. iii, p. 870, with Xiphilinus, 236. The first joins the titles Optimus and I mp. vii, on a diploma of Sept. 1, 114, while the second says that after Trajan had conquered the whole of Armenia the senate bestowed on him the title of Optimus. The epithet is first certainly applied to him by Pliny in the Panegyric, for its restoration in I.G.R.R. iii, 914, must be an error; and it is said not to appear on coin reverses before 104. Thereafter the word, generally in conjunction with princeps, is often applied to him, but not officially included in his titles until at least 114. Mr. Mattingly maintained in a lecture to the Roman Society that it was not official until 115: but this is not an essential deduction from the coin evidence, so far as I can see, and it is contradicted by that of all the inscriptions of 114. In fact, while Trajan did not have the title when he left Rome (M. & S. nos. 253, 4, 263, 633), if M. & S. no. 263a can be referred to the Athenian embassy (unless it is an error), the coin evidence would not prevent us following that of inscriptions in thinking it was officially assumed by Trajan in the New Year of 114. However, since, in addition, all coins of 113–4 and even a few of 114–5 of Alexandria lack the title, it is best to follow Xiphilinus. This leaves A.J.A. 1926, p. 418, a diploma with Optimus and Imp. vi, unexplained. Two even earlier inscriptions, C.I.L. iii, 15021 and I.L.S. 293, give the title wrongly; in the latter case Mommsen has an explanation which Dessau gives. Gsell's restoration in I.L.Al. 1230 does not seem to be a necessary one. Inscriptions are more prone to erroneous anticipation than coins; but when all the inscriptions of a year give a certain title the inference cannot be resisted. The coin evidence, as it is available at present, is a baffling ally in settling the chronology of the Parthian wars; much may be hoped from an advertised work by P. L. Strack which will deal with the subject. Meanwhile, J. Vogt's Alexandrinische Münzen is valuable. He notes (p. 92) that the subjugation of Armenia seems to be indicated by coin types of the year 113–4, i.e. before August 114—but types of this character may date from any time after the beginning of the invasion.

^{page 10 note 5} See above, p. 5.

^{page 10 note 6} Dio. Exc. Ur. 16.

^{page 10 note 7} I.L.S. 1338.

^{page 10 note 8} C.I.L. iii, 39.

^{page 11 note 1} Themistius, Or. XVI. Cf. Suidas s.v. άμϕίβολοι

^{page 11 note 2} Joh. Lydus, De Mag. iii, 53.Google Scholar Jacoby 〈F. Gr. Hist. BD. p. 575〉 follows Rocs in tracing activity of Trajan in the Caucasus in this passage. If so, we might have an explanation of his boast to have gone further than Alexander.

^{page 11 note 3} Pliny reckons it a part of Adiabene; N.H. vi, 42. Moreover, both Singara and Adenystrae are said by Dio to be in Adiabene; which may even at this time have marched with Osrhoene. The territory of Mannus, who was the vassal of Mebarsapes and the neighbour of Abgarus, probably lay on the border. Nine kings of Osrhoene were apparently called ‘Mannus,’ including him who succeeded Parthamaspates there in 123. Both this man and Abgarus were sons of an Izates, and this name also occurs in the royal family of Adiabene at a slightly earlier period.

^{page 11 note 4} See Syria, 1927, p. 53, for a milestone of Trajan from the Diebel Sinjar, showing that the normal Roman development of the new provinces was begun immediately.

^{page 11 note 5} Dio, Exc. Ug. 54. Exc. Ur. 15 also belongs to this year. Sentius had been sent to Adenystrae as an envoy to Mebarsapes, who imprisoned him. He had been sent from Armenia probably to announce Trajan's approach, and in due course contrived to deliver the town over to the Romans. Adenystrae has been identified by Hoffmann 〈Z.D.M.G. xxxii, 741〉 with the medieval Dunaisir near Mardin, some 50 miles W. of Nisibis.

^{page 11 note 6} Exc. Ug. 53. Details are given here and in the extracts of Suidas s.vv. ἄκρα, έλλόβια, Έδεσσα, ϕυλάρχης It is possible that Trajan himself went from Armenia to Osrhoene and thence to Nisibis, if the order of the extracts represents the right chronology.

^{page 11 note 7} Suidas s.vv. στόλος, ὑϕηγήσονται Eutropius, viii, 3.

^{page 12 note 1} Manisarus, who, when Trajan entered Mesopotamia, sent him encouraging messages from a locality where he was making trouble, is hard to account for. He has been supposed (by Gutschmid, von, Geschichte Irans und seiner Nachbarländer (Tübingen, 1888), p. 143Google Scholar) to be the ruler of Carduene (cf. Eutropius l.c.), but if Osroes was making a campaign against him this is unlikely.

^{page 12 note 2} Steph. Byz. s. Χωχή

^{page 12 note 3} The only absolutely fixed points are Elegeia in 114, 8th book, and Hatra in 116, 17th book. But the chronological arrangement here adopted enables the books to be divided among the campaigns very plausibly. Thus, 8, 9, 10, and possibly 11, belong to 114. 12, which perhaps began with the earthquake, 13, which certainly described the conquest of Adiabene beyond Tigris, 14 and 15 belong to 115. In 16 came an account of Trajan's journey to the Persian Gulf, and probably the story of the revolt was begun here and concluded in the 17th book. Roos, however, o.c. p. 56, thinks that the mention of a town in Mesene in Book 16 must refer to some unknown event during the revolt; and that Trajan's known visit to the Persian Gulf must have been described in an earlier book.

^{page 12 note 4} B.M.C. Parthia, pp. 209–214. Vologesias is suggested, but in that case we should expect him to be more prominent.

^{page 12 note 5} B.M.C. Arabia, etc., p. cxcii, and references there.

^{page 12 note 6} ibid. p. clxxix ff. They must not be confused with Manisarus.

^{page 12 note 7} e.g. ibid. p. cxc.

^{page 13 note 1} The Acts of Sharbil in Cureton, Ancient Syriac Documents, p. 41, synchronise Sept. 4, 112 with the third year of Abgar and the fifteenth of Trajan. Cf. von Gutschmid, , Mém. de l'Acad. des Sciences de St. Petersbourg, VIIe serie, xxxv, p. 17, cf. pp. 25–28. p. 17Google Scholar. Suidas s.v. ὠνητή · καὶ τὴν χώραν ἐπιτρέπειν Τραιανᾦ Αὔγαρον, καίπερ ἄρτι ὠνητὴν ἐκ Πακόρου ἔχει λαβὼν πολλῶν χρημάτων· καὶ τοῦτο ἀσμένῳ τῷ βασιλεῖ γίνεται.

^{page 13 note 2} B.M.C. Parthia, p. 205.

^{page 13 note 3} See below, p. 25, where the question is more fully treated.

^{page 13 note 4} Cf. Suidas s.v. εὐθεῖαν

^{page 13 note 5} The actual place is not to be identified. Perhaps Za-feran (Bell, Amurath to Amurath, p. 296). The later important trade route crossed lower down at Wanna. He may simply have followed Alexander's route as far as that was known.

^{page 14 note 1} Amm. Marc. xxiv, 2. Cohen no. 178 (=Mattingly and Sydenham no. 657) may celebrate this, as the latter editors affirm. Zosimus iii, 15 has Zaragardia.

^{page 14 note 2} He also visited the Ήϕαίστον νῆσοι the bitumen sources for the walls of Babylon, and the modern Kirkuk (Arrian, Parthica, 13, from Steph. Byz. s.v.).

^{page 14 note 3} E.g. by Henderson, o.c. pp. 327–9.

^{page 14 note 4} Xiphilinus, 240, (following Tillemont's correction to Βαβυλῶνι adopted by Boissevain) the sense of which passage is certain.

^{page 14 note 5} Steph. Byz. s.vv. Χαζήνη, ϕάλγα, Νάαρδα de la Berge, Trajan, pp. 172–3. I.L.S. 9471, έπιμελητῇ εὐθηνίας έν τῷ πολέμω τῷ Παρθικῷ τῆς ὂχθης τυῦ Εὐϕράτον The ‘praef. ripae fluminis Euphratis’ of I.L.S. 2709 belongs perhaps to this period; but it is impossible to tell what his office was. Cf. Jacoby, F. Gr. Hist. BD. pp. 576–7 he quotes other extracts from Arrian which doubtfully refer to the Euphrates force.

^{page 14 note 6} Suidas s.v. ναῦς

^{page 15 note 1} But it must be admitted that Malalas xi, 274 is against this.

^{page 15 note 2} For a reference to the season of the storm see above, p. 8, n. 2.

^{page 15 note 3} B.M.C. Arabia, etc., p. cciii.

^{page 15 note 4} Jordanes, Romana 268.

^{page 15 note 5} Warmington, E. H., Commerce between the Roman Empire and India, p. 96Google Scholar.

^{page 15 note 6} Henderson, o.c. p. 330.

^{page 15 note 7} See below, p. 26, n. 2.

^{page 15 note 8} Art. ‘Seleukeia’ in p.-W. cf. his list of authorities on either side.

^{page 15 note 9} o.c. p. 331.

^{page 16 note 1} Henderson's map is at fault here.

^{page 16 note 2} P.-W. VI, col. 553.

^{page 16 note 3} Dio describes them as ὑποστράτηγοι that is to say legionary legates. What the rank of Lusius was at this time remains not very clear. At some time before the end of 116 he was adlectus inter praetorios and perhaps consul before the end of that year also. Groag (P.-W. XIII, col. 1882) thinks that the words of Dio (Exc. Val. 290) imply an adlectio only after his successes in this year, though he admits that the analogy of Maximus suggests that Lusius was already a senator. We do not know in the least how close the parallel was; still the most natural thing is to suppose that for his performances in 114, which cannot have gone unrewarded, Lusius received the honour of adlectio and that his victories in 116 won him the consulship, as those of Erucius and Julius did for them.

^{page 16 note 4} Wiener Studien xxxviii, 167.

^{page 16 note 5} If this man was legate in Lower Germany in 88, he might seem a trifle elderly to cope with the emergencies of 116. Pichlmayr (Hermes xxxiii, 667) declared, on the evidence of Aur. Vict. Ep. 11, 10, that Norbanus' name was A. Lappius. The difficulties involved in this led Groag (P.-W. Suppl. I, col. 112) to christen him L. Appius Norbanus Lappius Maximus. The final addition of Santra by Hauler apparently caused him (ap. Hauler, p. 170) to suggest that the Parthian victim might have been a son of the loyalist of 88.

^{page 17 note 1} Malalas XI, p. 273–4.

^{page 17 note 2} Exc. Ur. 16, correctly placed by Boissevain.

^{page 17 note 3} See above, p. 10.

^{page 17 note 4} Hauler (o.c. p. 173) has an alternative identification with Ardaches, whom the history of Armenia ascribed to Moses of Chorene described as the King of Armenia in this period. The identification cannot be supported. if for no other reason, because the history says that, having rebelled against Domitian and beaten his soldiers (cf. above, p. 9, n. 8), Ardaches came to Trajan, paid his dues, and remained tributary and undisturbed under both him and his successor. We have here no more than the tale of some petty monarch among those summoned to Satala, the record of whose affairs the historian has magnified into a history of the whole of Armenia.

^{page 17 note 5} Boissevain, vol. iii, p. 219.

^{page 18 note 1} A cousin of Parthamaspates, Mal. xi, 273.

^{page 18 note 2} P.-W. xiii, col. 1880. It may, however, be observed that, if Malalas copied from Arrian, his authority is of a different character from those quoted in col. 1884. Von Gutschmid was the first to make Lusius Sanatruces' conqueror.

^{page 18 note 3} S.H.A. Hadrian 5. 3. But cp. Dio, lxviii, 33, 1.

^{page 18 note 4} Provinces (Eng. Trans.) II, p. 69.

^{page 18 note 5} Malalas XI, 270–273.

^{page 18 note 6} E.g. by Longpérier, Mémoire sur la chronographie des rois Parthes, p. 141Google Scholar; Gardner, , The Parthian Coinage, p. 55Google Scholar.

^{page 18 note 7} Hill, however, (B.M.C. Arabia, etc. p. clxxix) seems to prefer von Gutschmid's authority to that of Wroth and to identify Mithradates with an otherwise unknown king who appears on a coin of Persis of about this date and whose name is read as MTRDT. How a king of Persis is found invading Syria is not explained.

^{page 19 note 1} There are genealogical discrepancies between Malalas' two stories.

^{page 19 note 2} Dio, lxviii, 33, 1.

^{page 19 note 3} Hermes lxi (1926), pp. 192–202.

^{page 20 note 1} Pliny, however, did not comply, as appears from Ep. lxvii.

^{page 20 note 2} Ep. lxvii.

^{page 21 note 1} The means by which he proves that the other four cities in question were only civitates liberae when the Elder Pliny wrote may be accepted here, since our argument does not require their rejection.

In the case of the Aphrodisians, he admits they had a foedus in 35 B.C., but as Pliny calls them liberi they had lost it.

^{page 22 note 1} E.g. in Ep. xxxii.

^{page 22 note 2} Or after that of Varenus in 106.

^{page 22 note 3} According to Vogt (Alexandrinische Münzen i, 77), there was a rumour afoot in Alexandria in 108–9 that Trajan was intending to visit the city. Such an intention he might well have had, in view of his recent annexation (see below, p. 26) of Arabia, but Vogt's arguments, though ingenious, do not seem to me to prove it.

^{page 23 note 1} What those intentions of Nero were which were cut short by his death cannot be determined; and since he did not carry them out, discussion of them is unprofitable.

^{page 23 note 2} Jos. Bell. Jud. vii, 7. Cf. Hegesippus v, 51, 2 who follows Josephus, of whom his work is indeed a rough translation, but ascribes the invasion to A.D. 71.

^{page 23 note 3} Otho's nova iura of Tac. Hist. i, 78 were probably in the nature of a financial concession to the provincials: but in any case they were ‘ostentata magis quam mansura.’

^{page 23 note 4} XII Fulminata may have been there since the autumn of 70: Jos. B.J. vii, i.

^{page 24 note 1} I.L.S. 8795 (minor correction by A. Aminraschwili reported in Berl. Phil. Woch. xlviii, 838). The inscription does not say that a Roman force was permanently stationed here, only that it helped King Mithridates of Iberia to build some walls. To connect these operations with the trade route across the Caucasus mentioned by Strabo and Pliny is, I think, to allow undue importance to a route which there is no evidence to show was freely used at this time. (Cf. Arrian Peripl. 9, 5.)

^{page 24 note 2} Anatolian Studies presented to Sir W. Ramsay, p. 114.

^{page 24 note 3} The details of trouble on the Syrian frontier under the elder Trajan (Plin. Pan. 14, Dessau, I.L.S. 8970. Suidas s.v. ἐπίκλημα) in 76 or 77 and its causes are equally conjectural. It is worth bearing in mind that in 77–8 the procurator in Bithynia, a man of much military experience, seems to have enjoyed exceptional powers (I.L.S. 253, 9199. B.M.C. Bithynia, p. 104). Cumont, Études Syriennes, pp. 329–330, relies on two inscriptions from near Cyrrhus, together with C.I.L. iii, 192, 195, to show that legion VII Claudia came to Commagene as a temporary garrison in 72 and stayed to take part in Trajan's war. But it fought in the Dacian wars, and I think none of the inscriptions need be earlier than 114.

^{page 24 note 4} Without better authority we cannot use the statement in Moses Chorenensis that Ardaches of Armenia had defeated Domitian's troops (see p. 9, n. 8; 17, n. 4), or that of Suidas under the heading Δομετιανός

^{page 25 note 1} Exc. Ug. 51, 52.

^{page 25 note 2} Trajan's pronouncement about Parthamasiris (Suidas from Arrian, Partbica s.v. γνῶσις) shows that after the invasion of Armenia, Axidares was still regarded as the responsible king of the country, but indicates also what was to become of him. (Cf. Suidas s.v. ἀμϕίλογον). Parthamasiris did not keep his first appointment with Trajan—probably at Satala where his humiliation was to have encouraged the other vassals there assembled—(Suidas s.v. χρῆναι). His excuse was as follows: τῆς δὲ τριβῆς τὰ αἴτια οὐ δυνατὸς γενέσθαι εὐθεῖαν παρὰ βασιλέα ὲλάσαι τῷ δεῖσαι τὰς φυλακὰς τὰς Ἀξιδάρου καὶ περιελθεῖν ὲν κύκλῳ καὶ οὕτω διὰ μακροῦ ἀφικὲσθαι. Clearly then Axidares was still holding some part of Armenia, and considering where Parthamasiris came from and where he could not get to, this was probably the north-western district. If Trajan took measures to secure Arsamosata, it may be that Parthamasiris' supporters were known to be active in that part. It seems likely that he had dislodged Axidares from the capital, Artaxata, and from the Araxes valley.

^{page 26 note 1} Exc. Ug. 52.

^{page 26 note 2} Other motives may have influenced this decision, such as a desire to improve trade relations with India by more efficient control of the Red Sea. Trajan put a fleet there (see above, p. 15), and built a first-class road to it through the Naba-taean territory. He also repaired and enlarged the canal between the Nile and the Red Sea from Babylon on the Nile to Clysma. One result of this may have been the Indian embassy in the following year.

^{page 28 note 1} In fact, when the Romans withdrew, the trade returned to its old channels, though a change took place in the third century. (See note below.)

^{page 28 note 2} Warmington, , Commerce between the Roman Empire and India, pp. 91–100Google Scholar, panegyrises Trajan as a ‘second Augustus in Roman commerce’ and either explicitly or implicitly ascribes a commercial motive for all he did and some things which he may have done in the East. The arguments do not seem to me convincing, but it is impossible here to deal with them piecemeal. Vespasian's reorganisation of Cappadocia is quoted as the first of a series of steps which were ‘begun already in A.D. 100 or thereabouts by Trajan’ in order to link up the eastern boundaries of the Empire commercially and otherwise (p. 92); this seems to illustrate not unfairly the nature of the evidence for such a view. Trajan is said to have formed a scheme (‘which had been suggested at the imperial court for some time’) for the invasion of India. As evidence of this the remark of Plutarch (Pomp. 70) that ‘given leaders such as Pompey and Caesar acting conjointly, India could not have resisted 70,000 Roman troops’ is adduced, but, to say the least, does not seem to be worth very much. As for Martial's epigram quoted (xii, 8, 8–10), it refers to Trajan's accession. and contains no idea of foreign conquest but is a rhetorical way of saying what a good man Trajan was, while Statius' lines (Silvae, iv, 1, 40–42), if they are evidence of anything, are evidence for court flattery of the irresponsible ambitions of Domitian. That Trajan was forced to return from the Persian Gulf by the news of the revolt is not the case. That ‘the general results of Trajan's visit to the East are reflected in the rise of Palmyra, in the occurrence of three gold coins (of Domitian, Trajan and Sabina) together with coins of Kushan kings at Jellalabad, and above all in the detailed information of North-West India contained in Ptolemy’ (p. 95) seems to me to be unproven. The position of Palmyra had indeed made it important since early imperial days: but its wealth surely belongs to a later time; and the second century inscriptions are singularly weak in evidence that any traffic of value from the Far East was passing through the town. If Trajan's wars had, as the author in the previous sentence declares, for a time closed to Romans the silk route through Parthia (thus assisting a rapprochement between Romans and Kushans), Palmyra must have suffered. The coins by themselves are pale reflectors: and it is hard to see what the war that brought Trajan to the East, which served rather to close the avenues of intercourse and information, can have contributed to the work of Ptolemy. Ptolemy (and the coins, but cp. p. 300) owed more to the improved relations between India and Alexandria, for which Trajan was in part responsible (cf. p. 299).

^{page 29 note 1} Cf. some remarks of de la Berge Essai, pp. 151–5, which are worth reading. His account of the Parthian wars is still the best available. The prominence in the later stages of the war of Lusius Quietus, the first of those barbarian adventurers who served the empire well upon occasion, to do it greater harm in the long run, should not be overlooked. Themistius even says that Trajan had designed him as his successor, and we can easily see how in 116 he may have appeared the only bulwark in a sea of incompetence.

^{page 31 note 1} Bourier, , Über die Quellen der ersten vierzehn Bücher des Jo. Malalas (Munich 1899–1900)Google Scholar: Patzig in Byz. Zeitsch. 1901 sums up with the sentence ‘Nach Bourier hat Malalas in den ersten 14 Büchern im ganzen nur Timotheos, Domninos und Nestorianos benutzt; ich dagegen halte Domninos und Nestorianos für identisch und glaube, dass von Malalas alle diejenigen Quellen, die er im Prooemium aufzählt, wirklich benutzt worden sind’ (p. 610). Cf. also the analysis of Malalas' account by von Gutschmid in Dierauer's, Beiträge zu einer kritischen Geschichte Trajans pp. 154–158Google Scholar.

^{page 31 note 2} S.H.A., Hadrian, 5.

^{page 32 note 1} The older date for this king is preferred and the arguments of Wroth (B.M.C. Parthia, pp. lix, lx) are ignored. These are not, indeed, absolutely conclusive; but so much mystery surrounds these coins that they cannot be admitted as evidence to corroborate anything.

^{page 32 note 2} The locality is not further specified.

^{page 32 note 3} Ap. Dierauer, o.c. p. 157.

^{page 33 note 1} Von Stauffenberg has a very complicated theory of the Armenian antecedents of the war. According to this, Osroes first broke his contract with Rome by appointing Axidares king of Armenia, but he deceived himself in thinking that his nephew would be the ‘ergebener Vollstrecker seiner Wünsche,’ the ‘in allen Dingen gehorsame Kreatur’ that he desired (p. 263). In fact, in the diplomatic intrigues which followed his action (?), Axidares declared for Trajan and received a provisional recognition. Thereupon Osroes deposed him in favour of his brother Parthamasiris, who obediently —but without first effectively ousting Axidares— joined the Euphratesian expedition, but did not therefore, it seems, give up hope that Trajan would ultimately recognise his claims in Armenia. On Trajan's approach Osroes sent an embassy to Athens, but after such preliminaries it is hard to imagine that his envoys had much confidence in the plea which Dio says they advanced. Axidares also sent envoys to Athens. Trajan concluded a provisional alliance with him (which he afterwards repudiated); and he had already made some headway in Armenia before Trajan arrived there. The reason why Dio does not relate these facts is the same as that already given. I can here only refer to the alternative version which I have adopted in the above article: it must suffice to say that this very elaborate reconstruction seems to turn on a less probable interpretation of Dio Exc. Ug. 51 and a wish to find a place for the invasion of Euphratesia.

^{page 33 note 2} Cf. p. 18.

^{page 34 note 1} Denkschr. d. Wien. Akad. lvii, 3 Abt.

^{page 35 note 1} ix, 216.

^{page 35 note 2} The victory, however, continued to be celebrated on August 9. Cf. Fasti Amit. Antiat. Maff. Allif. in C.I.L. I².

^{page 35 note 3} The theory was advanced by von Domaszewski in Abhandlungen zur römischen Religion, p. 206 ff.