¹ I am very grateful to Mr. G. T. Griffith for his kindness in reading and commenting on this paper. I have employed the following abbreviations: Badian, Studies = E. Badian, Studies in Greek and Roman History (1964), Ch. I (a reprint of his article ‘Notes on Roman Policy in Illyria’, in PBSR xx, 1952).

Hammond, Epirus = N. G. L. Hammond, Epirus (1967).

Holleaux, CAH = Holleaux, M. in Cambridge Ancient History VII (1928)Google Scholar.

Holleaux, Études = M. Holleaux, Études d'épigraphie et d'histoire grecques IV (1952).

Holleaux, Rome = M. Holleaux, Rome, la Grèce et les monarchies hellénistiques (1920).

Walbank, Comm. = Walbank, F. W., Historical Commentary on Polybius I (1957)Google Scholar.

Walbank, Philip = F. W. Walbank, Philip V of Macedon (1940). The new evidence is in Studia Albanica, a periodical published at Tiranë.

² As Caesar, , BC 3, 42Google Scholar, 5 remarked, ‘plerumque frumento utuntur importato’.

³ Holleaux, Études 74, n. 1, mentions the flooding of the plain in October but not in the spring and early summer. His belief, that when the plain is flooded ‘toute opération de guerre devient impossible; la saison militaire prend donc fin sur cette côte avec le mois de septembre’, would certainly lead to an extremely short military season. But his belief is unfounded; for armies are not confined to the plains but may move on higher ground. The two arms of the Via Egnatia, which started from Epidamnus and Apollonia, followed the higher ground; a minor road linking the two cities direct must have been built on artificial embankments. See Miller, K., Itineraria Romana (1916), fig. 161 and p. 519Google Scholar.

⁴ Papagos, A., The Battle of Greece 1940–1941 (Athens, 1949), 285 f.Google Scholar and 289.

⁵ This is easily seen from the plans of operations in A. Papagos, op. cit. facing p. 296 and p. 300.

⁶ The Mediterranean Pilot, 8th edition (1957), 169. Delatte, A., Les Portulans grecs (Liège, 1947)Google Scholar, Portulan 2, p. 202 f., gives a detailed description of this coast which was then ὅλη δάσος.

⁷ Ibidem, Portulan I, p. 24: Καὶ ἀποκεῖ περνᾶς εἰς τὸ Ντουράτζο καὶ ἀπὸ μακρέα δείχνει ὡσὰν νησὶν μὲ πέντε κεφάλια. See also Portulan 2, p. 202.

⁸ The Drilon, as the Drin was called, entered the Buenë in antiquity, as a part of it still does; see May, J. M. F. in JRS 36 (1946), 55 fGoogle Scholar. In Portulan 2, p. 202, ἀπὸ τὸ Σὰν Τζουᾶν δὲ Μέδευβα ὡς τὴν σάκα τοῦ Λέσου ἔναι μίλλια β´ ἀπὸ τὴν μποῦκα τοῦ Λέσου κτλ., the word μποũκα (implies that the Lesos was a river, that is the Drin, then entering the sea by Lesh.

⁹ This was shown by the conflict over Epidamnus between Corcyra and Corinth with her allies in the years before the Peloponnesian War and by the difficulties of Julius Caesar and Antony when Pompey held Dyrrachium and Oricum; the former landed on the Acroceraunian coast ‘saxa inter et alia loca periculosa’ (Caesar, , BC 3, 6Google Scholar, 3) and the latter ran the gauntlet past Dyrrachium and reached the Nymphaeum (Caesar, , BC 3, 26Google Scholar).

¹⁰ Strabo C 283 gives the additional reason that Epidamnus affords better access to the interior. The Franks also preferred this crossing.

¹¹ For the lembus see Torr, C., Ancient Ships (1895), 115 f.Google Scholar At the extremes of the time scale the Taphian pirates probably and the Venetians certainly used similar craft.

¹² See Hammond, Epirus 591 and 595 f. I disagree with the views of Badian, Studies 26, n. 6 and Walbank, Comm. 156 that ‘the collaboration with Scerdilaïdas shows that the expedition of 240 was part of a policy of planned expansion into Epirus’. This is not how Polybius describes the action of the Illyrians on coming to Phoenice, and I put more faith in his understanding of Illyrian methods and Illyrian opportunism than I do in a rationalistic reorientation of the facts supplied by Polybius.

¹³ For kingdoms within Illyria see my article in BSA 61 (1966), 239 fGoogle Scholar. Agron's power was both military and naval and exceeded that of any previous dynast in Illyria (Plb. 2, 2, 4).

¹⁴ The threat to the independence of the states holding this coast is stressed by Polybius (2, 6, 8) as a new development, arising from and after the operations of 230 at Phoenice: οὐκέτι περὶ τῶν ἀπὸ τῆς χώρας ἠγωνίων, καθάπερ ἐν τοῖς ἔμπροσθεν χρόνοις, ἀλλὰ περὶ σφῶν αὐτῶν καὶ τῶν πόλεων. Dell, H. J., ‘The origin and nature of Illyrian piracy’ in Historia 16 (1967), 356 f.Google Scholar, sees the significance of the change, but he seems to me to oversimplify in saying that before 230 ‘the Illyrians had raided primarily for food’ and ‘the region was so poor that food gathering in the form of raiding was required in order for the Illyrians to subsist’. A lembus was not built on the lines of a transporter to bring back pilfered corn or cattle (indeed Torr, loc. cit., reckons two horses a load for a lembus), but on the lines of a light galley or racer; the Illyrians were after valuables, including slaves, which could be converted, of course, into foodstuffs later, if it was necessary.

¹⁵ Badian, Studies 3, has advanced the hypothesis, that ‘if Italian blockade-runners carried cargoes to the enemy the Illyrian forces were then fighting, the Illyrians can hardly be blamed for interfering’. This hypothesis does not seem to be feasible. In the first place the Illyrians seized Phoenice so suddenly, thanks to the action of the Gauls, that there was never any blockade of ‘the enemy the Illyrian forces were then fighting’, namely the Greeks of Phoenice, and subsequently it was the Illyrians who were blockaded in Phoenice and not the Epirotes who held the countryside and had supplies from Epirus (see Hammond, Epirus 595 f. and fig. 29). In the second place the hypothesis is not compatible with the account of Polybius, that some Illyrians went off on their own from the fleet when it was lying off Santi Quaranta (πλείους χωριζόμενοι ἀπὸ τοῦ στόλου); for his words show that they were engaging in a bit of private enterprise and not instituting a preconcerted blockade for Italian ships to run the gauntlet. As for the statement of Holleaux, CAH 831, that the Italians were ‘in Epirus’, it lacks any support in the evidence and any probability.

¹⁶ These seem to be the facts of the case. The different interpretations set upon the facts by Polybius (2,8) and by Appian (Illyr. 7) go back to at least two Roman annalistic sources, of which Fabius was almost certainly one (see Gelzer, M. in Hermes 68 (1933), 142 f.Google Scholar). Appian has Issa, a Greek city, appealing to Rome for help against Agron, Rome sending an embassy with the Issaeans to Agron, and Illyrian lembi attacking the Roman embassy en route to Agron and killing the leading Issaean envoy and a Roman envoy, Coruncanius, while the others got away. A fine case of Rome protecting a little Greek state and Rome wronged by the wicked Illyrians ! Polybius gives an account which cannot be reconciled with that of Appian and which is less blatantly pro-Roman. He says that Roman envoys were sent out to investigate (2, 8, 3 ἐπίσκεψιν ποιησομένους), but only when the Senate had been pressed by many individuals (incidentally one wonders if the Senate of 229 was much moved by individual complaints); the envoys arrived to find Teuta besieging Issa and had an audience with her. The ensuing conversations are given at some length. It is most unlikely that minutes of the actual conversations were kept or that reports of them were transmitted; Polybius (or his source in part) is expressing not τὰ ὂντα but τὰ δέοντα, and it is uncritical to press any details of the alleged conversations, as do Holleaux, CAH 832, and Dell, H. J. in Historia 16, 357Google Scholar. The circumstances under which the Roman envoy was killed are different in Appian and in Polybius. The statement of Polybius that Teuta instigated the assassination should not therefore be accepted as necessarily true. Indeed there is an obvious motive for invention in the desire to show Teuta breaking the ius gentium (ὀλιγωρήσασα Τὦν παρ᾿ ἀνθρώποις ὡρισμένων δικαίων, Plb. 2, 8, 12). Dio, frag. 49, which is amplified in Zonaras 8, 19, is a conflation of Appian's account and Polybius' attitude to feminine psychology in regard to Teuta. It is clear that Roman writers enjoyed whitewashing Rome and blackening Teuta; but even when we allow for that, the actions of Teuta show a folly and a confidence which find an echo in the Corfu Channel incident of 1945.

¹⁷ See Hammond, Epirus 590.

¹⁸ Many conjectures have been made as to what these reasons may have been. I incline to the view of Badian, Studies 4 f., but I do not think, as he does, that Rome's actions were ‘based on a misapprehension’, since Teuta's successes and ambitions were plain to see. The view of Holleaux, Rome 99 and CAH VIII 136, that the war ‘was imposed on Rome by the enemy’ is altogether too simple, and it is clear that even Teuta herself was unaware that she was at war with Rome.

¹⁹ Several scholars have regarded the force as unduly large; as Badian, Studies 4 remarks, ‘this strong deployment has baffled scholars and cannot be explained by Holleaux who thinks it due to Roman fear of Macedon’. The size of the force must be considered in relation to its purpose. The naval contingent was large enough to tackle the full naval strength of the Illyrians, whose lembi were far inferior in line of battle to Rome's quinqueremes, and the naval force would have been enough in itself if Rome's purpose had been to clear the seas of Illyrian ships. This, however, was not her purpose at all. She was out to control a part of the Illyrian coast, and therefore had to land an army. The Illyrians had fielded at short notice 10,000 infantrymen, highly regarded as fighters by the Greeks, in the campaign at Phoenice; they had cavalry forces too; and they had allies close at hand in Epirus and Macedon. I regard the Roman army of 20,000 infantry and 2,000 cavalry as no more than adequate for the purpose.

²⁰ There were two groups of Ardiaei, the main tribe in the region of Dubrovnik and a lesser tribe fairly close to Epidamnus. It is this lesser tribe which the consul who was stationed at Epidamnus in winter 229–8 was to watch (Plb. 2, 12, 2) and which was in dispute later between Rome and Macedon (Livy 27, 30, 13). See Hammond, Epirus 598 f., with map 15 on p. 614; and BSA 61 (1966), 249Google Scholar.

²¹ The ancient name of the Drin is Drilon. Holleaux, Rome 102, takes the Roman advance on land beyond the Drilon. G. Zippel, Die römische Herrschaft in Illyrien 51, and Walbank, Comm. 163, think the Roman forces proceeded to ‘the farther recesses of the Adriatic’; I see neither evidence nor probability in support of their view, as the farther recesses of the Adriatic were far beyond Teuta's kingdom.

²² Annalistic sources were at pains to show Teuta totally at fault, perhaps because Rome's conduct was not above criticism, and it is certainly true that Rome's attack on Teuta was hardly less sudden than that delivered by Italy on Albania on Good Friday morning in 1939.

²³ Plb. 2, 11, 17 τῷ Δημητρίῳ τοὺς πλείστους ὑποτάξαντες τῶν Ἰλλυριῶν καὶ μεγάλην αὐτῷ περιθέντες δυναστείαν means, I take it, the majority of the (subjugated) Illyrians and a kingdom for Demetrius which was intended to be a lasting one, as Walbank, Comm. 164 believes, and not just for autumn 229 or so, as Badian, Studies 8 holds (the Roman annalistic source in App., Illyr. 8 wrote with a knowledge of later developments). These Illyrians are described rather euphemistically by Polybius as τοῖς δεδωκόσιν ἑαυτοὺς εἰς τὴν πίστιν.

²⁴ I accept Polybius' account as a summary of the treaty (2, 12, 3). It is drawn probably from a Roman annalistic source, because the clause limiting the range of the lembi is said to have been concerned mostly with the (interests of the) Greek states, a piece of Roman ‘philhellenic’ propaganda which was anachronistic in Z28. Appian gives a different account; this is based on another Roman annalistic source. It is to the effect that Pinnes, the stepson of Teuta, was to keep his kingdom, apart from Corcyra, Pharos, Issa and Epidamnus and apart from the Atintani among the Illyrians, and was to be a friend of Rome on condition that he respected Rome's division of spheres of authority and that no more than two lembi sailed past Lissus, even unarmed. The account of Dio, frag. 49, ends at the vital point τὴν ἀρχὴν ἀφῆκεν without giving a definition of its extent (cf. Zonaras 8, 19); it may have included a change which took place only after the Roman settlement. On the positive side there is no doubt that Rome kept under her direct control only Apollonia, Epidamnus, the Parthini and the Atintani on the mainland. When Teuta evacuated all ‘Illyris’ except a few districts, it seems that ‘Illyris’ here, as elsewhere, is that part of the territories where the Illyrian tribes lived which was familiar to the Greeks, that is the mainland from Ulcinj or Scodra southwards to the vicinity of the Gulf of Valona and not northernmost Albania and the Dalmatian coast. Her own kingdom was still, as in the past, based on Dubrovnik, and it was this kingdom of which Rome left the control to Pinnes (App., Illyr. 7). The opinion of Holleaux, Rome 105, n. 4, citing Täubler, , Imp. Romanum I, 77Google Scholar, and Zippel, op. cit. 53, is that the frontier for Teuta's army was drawn also at Lissus; this is unlikely to be correct, because the tribes north of Lissus were probably included in Demetrius' kingdom and were intended to form an independent group as a buffer between Teuta and Rome.

²⁵ See Hammond, Epirus 232 f., and map 18 on p. 700.

²⁶ Caesar, , BC 3, 13Google Scholar, 5 ‘ad Apsum flumen in finibus Apolloniatium’. On fig. 1 Apolloniatis is confined to the plain and its edge of rising ground, because there are a number of fortified sites in the hill-country east and southeast of Apollonia, a district called ‘Malakastra’, which Praschniker (see n. 46 below) described as the most fertile and most thickly populated area in Albania. Part of it was evidently held by the Bylliones whose chief town was Byllis (for their position see Hammond, Epirus 674).

²⁷ Caeser, , BC 3, 30Google Scholar, 7. The outlets of the Shkumbi, Semeni and Vijosë may have been in ancient times not where they are today (see Hammond, Epirus 133).

²⁸ Livy 43, 9, 7 ‘ad Lychnidum (MS. lycidinum) Dassaretiorum’, whereas Uscana nearby was on the Macedonian frontier when Perseus was king (Livy 43, 10, 1) and also later (Plb. 34, 12, 6). I do not think that Dassaretis and ‘Lychnidia’ in Plb. 5, 108, 8 are exclusive of one another. We should expect the Parthini to have held the whole valley, since a division of the upper valley and the middle valley between two tribes would be difficult to maintain.

²⁹ See Hammond, Epirus 600. The territory of the Atintani was at some time said to be Macedonian (Steph. Byz., ᾿Ατιντανία μοῖρα Μακεδονίας, with Atintas a son of Macedon). In 229 Atintanis seems to have been at some distance from the Macedonian frontier; so too the area round Lake Lychnitis was not Macedonian but Dassaretian.

³⁰ See CAH 7, 748 and BSA 61 (1966), 249Google Scholar.

³¹ The zone of direct dependence which I have defined is completely different from that shown by Holleaux, , CAH 7, 768Google Scholar, map 10 and 825, map 14. See my Fig. 1 inset. His area is twice as large and twice as vulnerable. His error in this matter affected his theory of Rome's policy and Macedon's reactions to some extent. Fine, J. V. A., ‘Macedon, Illyria and Rome 220–219 B.C.’, JRS 26 (1936), 27Google Scholar follows Holleaux. Badian, Studies 7, realized that ‘the usual map of the protectorate is entirely misleading’, but he did not distinguish the Illyrian Atintani from the Epirote Atintanes, nor did Fine, J. V. A., ‘The problem of Macedonian holdings in Epirus and Thessaly in 221 B.C.’, TAPA 63 (1932), 130Google Scholar.

³² The orthodox view tha t the coinage of Scodra started in 211 after Philip's capture of Lissus has been challenged by H. Ceka, who would date it from Rome's intervention in 229, and by S. Islami, who moves the date back to about 250 (see Studia Albanica 1966, 1, 225 f.Google Scholar). May, J. M. F. in JRS 36 (1946), 52 fGoogle Scholar. attributes the coinage to 186–168. If 229 is preferred, it illustrates the economic interest of Scodra.

³³ Their status is discussed by Walbank, Comm. 161–162 and Badian, Studies 9. The Roman annalists regarded them as ‘free’ (App., Illyr. 8 init.) and Pinnes as a ‘friend’, but Polybius is more realistic in describing them as subject to Rome (3, 16, 3 τὰς κατὰ τὴν Ἰλλυρίδα πόλεις τὰς ὑπὸ Ῥωμαίους ταττομένας); the treaty between Philip and Hannibal speaks of Rome's control of these states de facto rather than de iure (Plb. 7, 9, 13 μηδὲ εἶναι Ῥωμαίους κυρίους Κερκυραίων κτλ.).

³⁴ This is a common enough trick of twentieth-century diplomacy, as we have seen most recently in Russia's relations with Israel and the Arab States; a threat may be indirect, it is none the less a threat. Rome played the same diplomatic trick in 200 (see Hammond, Epirus 614 f.). Various views have been taken of Rome's intentions. Holleaux, CAH 839, argued that Rome intervened in fear of Macedon and that by her settlement Rome took advantage of Macedon's weakness ‘to infringe grievously her historic rights and thwart her traditional ambitions’. This view has been challenged, most recently by Dell, H. J. in Classical Philology 62 (1967), 99Google Scholar. On the other hand Dell's argument that ‘Antigonus could only have viewed the intervention of the Republic as a piece of good fortune for himself and for Macedonia’ seems to confuse the intermediate effect, the weakening of the Illyrians, with the final effect, which was the juxtaposition of a new and much more powerful near-neighbour in the shape of Rome, a point appreciated by Oost, S. I., Roman Policy in Epirus and Acarnania (Dallas, 1954), 16Google Scholar. Badian, Studies 29, n. 44 contends that at no stage of the war did Rome show any fear of, hostility towards or indeed interest in Macedon. This is true if the emphasis is on ‘show’; but that is not to say that fear, hostility and interest were absent. In my opinion Rome's settlement in Central Albania and her diplomatic disregard of Macedon reveal that they were in fact present. Holleaux, Rome 115, reduces Rome's diplomatic moves ‘à des démonstrations de courtoisie, à un échange de politesses diplomatiques’; but this is to impute to Rome a degree of diplomatic refinement and political innocence which is not typical of Rome's relations, for instance, with Carthage in the third century.

³⁵ The point is well made by Holleaux, CAH 837.

³⁶ Appian (Illyr. 8) associates the moves of Demetrius with the three-year war against the Celts of the Po valley, and this synchronisation is given also by Polybius (3, 16, 2). Appian, but not Polybius, mentions the revolt of the Atintani from Rome and piracy by the Istrians, also being ‘Illyrians’. The latter event is recorded by Livy Ep. 20, Oros. 4, 13, 1, Zonaras 8, 20 (under 221 B.C.), and Eutropius 3, 7, 1 (piracy against Roman supply-ships, presumably going to the Po valley theatre of war). Appian says that Demetrius had an understanding with the Istrians, a statement which is not found elsewhere but is highly probable, since both were anti-Roman and Illyrian in outlook. The record and the timing of these events are due to the Roman annalists, and Polybius accepted their chronology at 3, 16, 2. I see no reason for the attitude of Holleaux, Rome 132 f., who claims that Appian is wrong in placing the revolt of Demetrius from Rome during the first three years of the Celtic War and that the Roman defeat of the Istrians in 221 ‘est presque sûrement apocryphe’.

³⁷ So also Fine, J. V. A., op. cit., JRS 26 (1936), 30Google Scholar. However, Dell, H. J. has argued in Classical Philology 62 (1967), 101Google Scholar, that this passage shows rather that Demetrius and his Illyrians were not allies of Antigonus, and that they were serving as mercenaries. I am not convinced by his argument. Polybius begins by saying that Antigonus προῆγε μετὰ τῶν συμμάχων, and he then enumerates the forces as follows: (1) the Macedonian army including Gauls, (2) mercenaries, (3) Peloponnesians (in fact allies), (4) allies (from outside the Peloponnese) Boeotians, Epirotes, Acarnanians and Illyrians. Now if the Illyrians had been mercenaries serving with the Macedonian army, they would have been enumerated in (2) μισθοφόρους δέ and not in (4) τῶν δὲ συμμάχων. A few years later the Illyrian Scerdilaïdas became an ally in the same way (Plb. 4, 29, 7).

³⁸ The chronology of these years is difficult (see Walbank, Comm. 324 f.). (1) Appian (Illyr. 8) places together but within a three-year period the war against the Gauls in the Po valley, a piratical expedition by Demetrius, a piratical venture by the Istrians and the revolt of the Atintani. The three-year period ἐπὶ τριετές is probably 225/4, 224/3 and 223/2. Then, when ‘the Gallic affairs were settled’, perhaps finally in 221, (Appian continues), a Roman attack was made on the pirates, presumably the Istrian pirates in the same theatre of war, i.e. in 220 (two wars against the Istrians are attested in other sources, probably in 221 and 220); then ἐς νέωτα, in the next year i.e. in 219, Rome attacked Demetrius. (2) Polybius (3, 16, 2–3) gives a sequence of events as a prelude to Rome's decision (in winter 220/219) to enter Illyris and attack Demetrius. The time base to which the sequence is attached is thus 220/219. Polybius gives us (a) Rome's benefits to Demetrius (229/8), (b) Demetrius' contemptuous treatment of Rome due first to the menace of the Gauls (225/2 probably) and ‘at that time’ (220/219) to the menace of Carthage, and (c) the attack of Demetrius on the dependents of Rome (220/219). Polybius says also that Demetrius had every hope (in 220/219) in Macedon's royal family, because he had fought on the Macedonian side and had shared in the hazards against Cleomenes (223–222), and that Demetrius had sailed past Lissus and had ravaged the Cyclades (summer 220). (3) On my interpretation the chronologies given by Appian and Polybius are in accord, which is not surprising if the annalist tradition is the source for most of the events with which we are concerned.

³⁹ Badian, Studies 13, finds Polybius' account of Demetrius' motives untenable and holds that Demetrius had not so far (i.e. by 220/219) done anything anti-Roman. As regards the latter point he has either missed or tacitly excluded the statements of Appian that Demetrius practised piracy and that Demetrius instigated the revolt of the Atintani from Rome (within 225/2; see last note), although Badian mentions the revolt on p. 30 in n. 66 apparently without sharing the extreme view of Holleaux, Rome 135, n. 1, ‘il n'y a nul compte à tenir des indications d'Appien Illyr. 8’ etc. As regards Demetrius' motives it is important to note the perfect tenses in Polybius' text and the changing circumstances as they affected Demetrius: thus in 225–2 Demetrius acted as he did because Rome was involved in war with the Gauls of the Po valley (Demetrius' actions, as we learn from Appian, being a piratical expedition, the instigation of Istrian piracy, and the instigation of the revolt of the Atintani), and then in 220–219 winter (τότε) thinking Rome would be involved in war with Carthage he began to attack the Roman sector in Illyris. His motives for action are intelligible, though not necessarily intelligent (but in making our judgement we must shake off our after-knowledge of Rome's remorseless and successful expansion). Demetrius' raid into the Cyclades in summer 220 is tacked on at the end by Polybius; this action was unrelated to actual or future involvements with Rome, and we see perhaps here an irrational step by Demetrius.

⁴⁰ Badian, Studies has made the latter point well. As he observes on p. 30 in n. 72, we cannot tell when Demetrius acquired Dimallum, but Holleaux, Rome 112, has a less open mind—the Romans took it, he says, in 229.

⁴¹ Plb. 3, 16, 4 and 7; 3, 18–19. App., Illyr. 8 fin. adds nothing of value. The story that Demetrius was killed by the Romans—a story common to Zonaras 8, 20—is a fictitious piece of Roman justice, as Plb. 3, 19, 11 has a more trustworthy account of Demetrius' death in Messenia.

⁴² Polybius (3, 16, 4) adds the point that in winter of 220–219 Rome observed the flourishing state of the Macedonian royal house, and this observation (he implies) played some part in her decision to attack Demetrius and secure her eastern flank. This seems to me a sensible suggestion by Polybius; but it is no more than a suggestion, because Polybius cannot have known the precise motives which actuated the members of the Senate in their debates and in their decision. And if Philip's alliance with Scerdilaïdas (Plb. 4, 29) was concluded in this winter, and if the alliance became known to the Senate, Polybius' suggestion is even more sensible. The basic fact, however, is that Demetrius would have ousted Rome altogether from Illyris, if Rome had taken no action.

⁴³ See Badian, Studies 13 f., for a very different account of this period. Holleaux, Rome 143 f. enlarges upon Philip's acceptance of Demetrius.

⁴⁴ On the coast: A. Philippson s.v. in P–W v, col. 646. Not on the coast: Walbank, Comm. 330. Near Epidamnus: Badian, Studies 16 and Walbank 330. In Parthinian territory: Holleaux, Rome 135, n. 1 ‘dans la contrée des Parthiniens, arrière d'Epidamnos’. Not in Parthinian territory: Badian ibid., Walbank ibid.

⁴⁵ Walbank, Comm. 330, makes Bargullum and Eugenium ‘townships of the Parthini’ but not Dimallum; the Latin text suggests that all three are one or the other.

⁴⁶ Reported first by Praschniker, C., ‘Muzakhia und Malakastra’ in Jahreshefte d. Öst. Arch. Instituts in Wien 21–22 (1922), 103Google Scholar with fig. 40.

⁴⁷ Plu., GQ 17.

⁴⁸ The names occur in Epirus with ῎Αρμενος as a variant of Άρμήν and Nestor with a double sigma, which is typical there (see Hammond, Epirus, Index I, 795 f.). Dautaj reads [Π]APMHNOC and AMYNTA[C]; but the latter is probably in the genitive (unless he was a potter like Nestor), and his Plate IV shows little space for the extra letter C.

⁴⁹ See the illustrations in Islami's, S. article in Studia Albanica 1966, 1, 241 fGoogle Scholar.

⁵⁰ The MS of Polybius give Διμαλλη, Διμαλη, Διμαλον and Διμαλλον at 3, 18 and 7, 9, 13. See δίμαλλος in Glossaria, and Hdt. 9, 92 for the famous sheep of Apolloniatis feeding ‘by the river’.

⁵¹ Geoponica II, 3, 4 εὐγένιον.

⁵² Or it might have been an offshoot of Apollonia, like the colonies of Syracuse which became independent. As a general rule the Illyrian tribes did not found cities of the Hellenistic type.

⁵³ The text of Plb. 5, 108, 2 is τῆς δὲ Δασσαρητίδος (Δασαρητίδος AB, δαρητίδος R) προσηγμένον πόλεις τὰς μὲν Θιβώτιδας (Φοιβάτιδας C) ἐπαγγελίαις Ἀντιπάτρειαν χρυσονδύωνα Γερτοῦντα, πολλὴν δὲ καὶ τῆς συνορούσης τούτοις Μακεδονίας ἐπιδεδραμηκότα. The emendation of τὰς μὲν Φιβώτιδας ἐπαγγελίαις to τὰς μὲν φόβῳ τὰς δὲ ἐπαγγελίαις has the merit of neatness but is not therefore correct; an alternative possibility is that Φιβώτιδας or Φοιβάτιδας is a cantonal name and that a lacuna is to be postulated before Άντιπάτρειαν or after Гερτοũντα, the explanation of the lacuna being that a phrase beginning τὰς δέ had dropped out because of its visual similarity to τὰς μέν. In either case τούτοις should refer to Pelagonia and Dassaretis and not to the three named towns of which Antipatrea is far from the Macedonian frontier.

⁵⁴ See my article in JRS 66 (1966), 43 fGoogle Scholar. The suggestion of Holleaux, Rome 165 f., Oost, op. cit., 113, n. 96 and Fine, op. cit., JRS 26 (1936), 39Google Scholar, that Rome prompted Scerdilaïdas, is such an obvious one that Polybius must have considered and rejected it. Rome had no desire in 217 to add to her troubles by stinging Macedon into action, and her experience with Demetrius had taught her not to put her trust in Illyrian princes. Holleaux, in making Rome prompt Scerdilaïdas, attributes rather gratuitously to Rome a folly, which he then points out somewhat complacently was a folly: ‘tel est le résultat, exactement contraire au calcul des Romains, de l'inopportune agression de Scerdilaïdas’. Badian, Studies 18, rejects Holleaux's suggestion. Scerdilaïdas was evidently a short-sighted opportunist, rather like Teuta and Demetrius. The proximity of Hannibal may have excited his ambition (for Scerdilaïdas had the lembi to cross the Adriatic Sea), just as the proximity of Napoleon excited the ambition of Ali Pasha of Ioannina.

⁵⁵ This passage of Polybius contains a number of obscurities. He represents Philip as anxious to ἀνακτήσασθαι τὰς ἀφεστηκυίας πόλεις and he makes a distinction between ἀνεκτήσατο and κατελάβετο. This seems to indicate something more than that Scerdilaïdas had just taken these three cities and to point to a Macedonian claim of previous ownership, which was probably true only in the latter part of the fourth century, unless Philip had encroached on Dassaretis in 220 when Demetrius was active. Of the places here mentioned Gertous and Gerous are not the same (despite Fine's view, op. cit., 26). Gerous is evidently the Gerunium of Livy 31, 27, 2, a castellum in the vicinity of Berat and ‘situated probably on the limestone range of Mt. Shpiragrit’ (as I said in JRS 66 (1966), 43Google Scholar). Orgyssus of the Pisantini is likely to be the Orgessus of Livy 31, 27, 2, another castellum in the same district. In that case the order in Polybius is not geographically consecutive. In 217 Philip was securing the western frontier of Dassaretis in taking Antipatrea (Berat) and some fortified sites to the west of it at Gerunium and Orgessus. In 200, when Apustius ‘ravaged the frontier territory of Macedonia’ (Livy 31, 27, 2 extrema Macedoniae populatus), he captured the castella Corrhagus, Gerunium and Orgessus and then reached Antipatrea. Of the places near Lake Lychnitis, Boii is related to the river Bevus (Livy 31, 33, 6) and to the town Beve of Steph. Byz. Βεύη, πόλις Μακεδονίας καὶ πρὸς αὐτῇ Βεῦος ποταμός, which I think lay south of the lake (see JRS 66, 43 with n. 19)

⁵⁶ One wonders indeed if Philip thought in early summer 216 of landing troops in Italy, as Polybius suggests and as Walbank, Philip 70 implies (‘so long as Philip had hoped to invade Italy himself, negotiations were better delayed until he could approach Hannibal as an equal on Italian soil’). Philip's fleet of 100 lembi could hardly have transported more than 5,000 to 7,000 infantry (and few, if any cavalry) across the Adriatic and it would have been utter folly to put this force into Apulia at a time when before the battle of Cannae (see Walbank, Comm. 439) there were in the coastal plain of Apulia Roman forces of anything up to 80,000 infantry and 6,000 cavalry and Hannibal's forces of some 40,000 infantry and 10,000 cavalry, with whom Philip had no understanding. On the other hand, if he landed his force on the coast of Lucania, it would be farther from its base and in great danger of being cut off. In view of these considerations I do not think that the suggestion of Polybius should be seriously entertained or amplified, as it has been, for instance, by Holleaux, Rome 177. ‘Peut-être remue-t-il de plus grands projets: si les dieux l'assistent, si la mer reste libre, passer au plus tôt dans l'Italie du sud; s'y faire accueillir comme un sauveur, comme un second Pyrrhos, par les villes helléniques etc.’

⁵⁷ These points are made by Polybius (5, 109, 1–2 οὐχ ὡς πρὸς ναυμαχίαν … ἀλλὰ μᾶλλον ἕως τοῦ παρακομίζειν στρατιώτας καὶ θᾶττον διαίρειν οὗ πρόθοιτο καὶ παραδόξως ἐπιφαίνεσθαι τοῖς πολεμίοις).

⁵⁸ His rearguard was at the island of Sason, while he himself was close to the mouth of the Aous (ἤδη…συνεγγίζοντος αὐτοῦ τοῖς περὶ τὸν ᾿Αῷον ποταμὸν τόποις); the message came to Philip ὑπὸ νύκτα. He evidently planned to begin this commando type of landing as dark fell.

⁵⁹ The criticisms made by Polybius are unduly harsh. They owe much to hindsight. Holleaux, Rome 178, n. 1 defends Philip. Badian, Studies 19, suggests that Philip may have intended to attack not Apollonia but Scerdilaïdas. His suggestion involves great difficulties. Philip might have had to sail as far as Kotor or Dubrovnik without any friendly harbour or even facility for shelter and watering, and he would have lost any possibility of contact with his forces in Dassaretis. Even if he deprived Scerdilaïdas of an island such as Pharos, and if he established a naval base there, he had no hope of keeping contact from there with Macedonia. If he wished to attack Scerdilaïdas by sea, the only realistic policy was to reduce the southern Ardiaei and open up a naval base at Lissus or near Scodra, with which contact could be maintained overland from Macedonia. This he did in 211.

⁶⁰ Holleaux, Rome 183 takes a different view; he thinks this was ‘la clause capitale des accords … Philippe s'engage à donner aide aux Puniques—c'est-à-dire à leur amener des renforts en Italie’.

⁶¹ Plu., Aratus 51 διαβαίνοντος εἰς Ἤπειρον αὐτοῦ, where the verb indicates a land force.

⁶² See Hammond, Epirus 127 and 609 for Oricum and Apollonia.

⁶³ The account in Livy 24, 40 is much better than the summary in Zonaras 9, 4; see Hammond, Epirus 609 for the status of Oricum, and n. 1 for comments on the views of Holleaux and Walbank on this campaign.

⁶⁴ Philip's gains in these years are known from Roman attempts to recover them by treaty or by force (Livy 27, 30, 13; 29, 12, 3; 29, 12, 13; 31, 27, 2). Selim Islami maintains in Studia Albanica 1966, 1, 227 f. that Philip's conquest was of Lissus alone and that his possession of it was short-lived; this does not, however, account for a passage such as Livy 27, 30, 13 (of 208). It is usually held that Lissus coined now for the first time, and payments for shipbuilding on a large scale for Philip may have prompted him to supply the metal. The circulation of the coinage, as that of Scodra, was purely local. Selim Islami, loc. cit., thinks Lissus may have started to coin earlier than in 211. See n. 32 above.

⁶⁵ Epirus and Acarnania being hostile to Aetolia and so to Rome and being allied with Macedon; see Hammond, Epirus 610.

⁶⁶ The motive imputed by Livy to Philip was a desire, before entering Greece, to secure in his rear ‘Illyrios finitumasque eis urbes’, that is the cities of Epidamnus and Apollonia in particular and the neighbouring Illyrians, namely the Dassaretii, the Parthini and those north of Epidamnus. It is unlikely that Apolloniatis and Oricia bordered on one another, as Amantia had a port on the Gulf.

⁶⁷ The suddenness of the attack on the extensive territory of Apollonia may indicate that it was seaborne. The annalistic account is often regarded with great suspicion, and certainly its slant is tendencious; but the facts and the timing especially may be correct. An inscription at Ambracia seems to confirm the capture of Ambracia by Philip, an event mentioned only in an annalistic account and therefore usually discredited (App., Mac. 3); see Hammond, Epirus 611. The attack on Corcyra in Zonaras 9, 6 may then be accepted. An earlier attack in 215 is mentioned in App., Mac. 1. It would indeed be surprising if Philip and his allies, the Epirotes, had left the rich island in peace.

⁶⁸ It is clear from the moderate wording of Livy that the Parthini did not come over to Rome and that Dimallum was not taken by Rome.

⁶⁹ It is possible but unlikely that he came from Aetolia (see Hammond, Epirus 612); for the sequence of events in Livy 29, 12, 1–2 seems to be speeded up to benefit Rome's reputation. The route through the Tsangon pass to Berat is described by W. M. Leake, Travels in Northern Greece 1, 335 f.

⁷⁰ Livy 29, 12, 13; App., , Mac. 3, 2Google Scholar; Zonaras 9, 11. Carthage could hardly blame Philip for making a separate peace as she had never sent him any help.

⁷¹ Holleaux, Rome 277 f. takes the opposite view; he thinks Rome was most generous, the treaty was very favourable to Philip, and Rome made peace because of ‘la faiblesse relative de l'adversaire’. Generosity was not conspicuous in Rome's treatment of weakened and feeble enemies at this time.

⁷² Scerdilaïdas was mentioned for the last time in Livy 27, 5, 7 and then together with Pleuratus. I believe these two were the rulers of the northern Ardiaei in the region of Dubrovnik. May, J. M. F. in JRS 36 (1946), 50 fGoogle Scholar. has argued, rather unconvincingly in the lack of any evidence, that Philip withdrew voluntarily ‘from Lissus and Ardiaea’ soon after 209.

⁷³ Holleaux, Rome 303, n. 1 puts Macedon's army in 199 at 20,000 infantry and 2,000 cavalry and in 197 with full conscription at 23,500 infantry and 2,000 cavalry.