¹ I am most grateful to the Albanian scholars who welcomed me so warmly at Tirana in 1972 and to Professor F. Papazoglou of the University of Belgrade for their great kindness in sending me copies of Iliria and offprints of their works; and in general to many Greek and Albanian friends who enabled me in 1930–9 to gain a detailed knowledge of the areas with which this article is concerned. The exceptions are Çermenike and much of Dassaretis. I cite the translations or résumés of Albanian articles in French. The following abbreviations are used:

Anamali = Anamali, Skënder, ‘Amantie’, Iliria 2 (1972), 67–165CrossRefGoogle Scholar.

Budina = Budina, D., ‘Antigonée’, Iliria 2 (1972), 269–378Google Scholar.

Cabanes = Cabanes, P., L'Épire de la mort de Pyrrhos à la conquête romaine (272–167) (1976).Google Scholar

Ceka = Ceka, N., ‘Le koinon de Bylliones’, Iliria 1984. 2, 79–89Google Scholar.

Dautaj = Dautaj, B., ‘La cité illyrienne de Dimale’, Iliria 2 (1972), 149–65Google Scholar.

Hammond 1966, I = Hammond, N. G. L., ‘The kingdoms in Illyria circa 400–167 B.C.’, BSA 61 (1966), 239–53Google Scholar.

Hammond 1966, 2 = id., ‘The opening campaigns and the battle of Aoi Stena in the Second Macedonian War’, JRS 56 (1966), 39–54.

Hammond 1967 = id., Epirus (1967).

Hammond 1968 = id., ‘Illyria, Rome and Macedonia in 229–205 B.C.’, JRS 58 (1968), 1–21.

Hammond 1971= id., ‘Antigonea in Epirus’, JRS 61 (1971), 112–15.

Hammond 1974 = id., ‘The western part of the Via Egnatia’, JRS 64 (1974), 185–92.

Hammond 1976 = id., Migrations and Invasions in Greece and adjacent areas (1976).

Hammond 1980 = id., ‘The hosts of sacred envoys travelling through Epirus’, Epeirotika Khronika 22 (1980), 9–20.

Hammond 1981= id. (ed.), Atlas of the Greek and Roman World in Antiquity (1981).

Hammond 1983 = id., ‘The Lettering and Iconography of “Macedonian” coinage’, Ancient Greek Art and Iconography, ed. W. Moon (1983).

Hammond, HM = id., A History of Macedonia 1 (1972); 11 with G.T. Griffith (1979); m with F.W. Walbank (1988).

Holleaux = Holleaux, M., Rome, la Grèce et les monarchies hellénistiques (1921)Google Scholar.

Papazoglou 1970 = Papazoglou, F., ‘Quelques problèmes de l'histoire épirote’, Ziva Antika 20 (1970), 115–36Google Scholar.

Papazoglou 1986 = ead., ‘Politarques en Illyrie’, Historia 35 (1986), 438–48Google Scholar.

Walbank, C = Walbank, F. W., A Historical Commentary on Polybius I–III (1957–1979)Google Scholar.

² Hammond, , HM 11 74Google Scholar f. and Hammond 1983, 247. The lettering was evidently made by Ionic-speaking Greeks of colonies of Chalcis in Chalcidice. ‘Atintanoi’ was the Greek form of an Illyrian name in the West-Greek dialect of Epirus or in the Doric dialect of Epidamnus.

³ Ichnae was Paeonian (see Thuc. 2. 99. 4) until it was conquered c. 510 by the Macedonians. Price, M., Coins of the Macedonians, 7Google Scholar is mistaken in calling Ichnae the capital of the Bottiaeans for the pre-510 coins.

⁴ For the mines see Hammond, , HM 1, 93Google Scholar f. with n. 4. The workings at ‘Starski dol’ were reported in Ziva Antika 3 (1953), 261, and tin was found at Velmej in the Saletska valley (Ziva Antika 12 (1963), 341). See also Davies, O., Roman Mines in Europe (1935), 239Google Scholar and his map 6. Silver is not one of the minerals found in North Albania, for which see Hammond 1976, 75.

⁵ See Filow, B., Die archaische Nekropole von Trebenischte (1927)Google Scholar and Hammond, HM II, 91 n. 2 for subsequent reports and in CAH IV (1988), 250 f.

⁶ Either because Glaucias included in his realm the ‘Illyrii proprie dicti’ or because Diodorus (cf. 16. 93. 6) was using a general but anachronistic term; see Hammond 1966, 1, 241 f.

⁷ For a day's march see Hammond, , Alexander the Great: King, Commander and Statesman (New Jersey, 1980; London, 1981), 58, 66, 119 and 127Google Scholar.

⁸ The best map of Albania is that of E. Nowack, Geologische Karte von Albanien, 1:200,000. From Bradashesh Cassander would follow the route of the later Via Egnatia, for which see Hammond 1974, 188 with pl. IX, identifying the site there as ‘Ad Quintum’, as did Ceka, N. later in Iliria 6 (1976), 287CrossRefGoogle Scholar ff.

⁹ Hammond 1967, 600.

¹⁰ See Hammond 1974, 185 f. and 1976, 29.

¹¹ It was shown to have been a tribal group by Hecataeus, FGrH 1 F 101: Ἄβροι ἔθνος πρὸς τῷ ᾿Αδρίᾳ Ταυλαντίων.

¹² The Greek seems to be a translation of a Latin original, perhaps written by Fabius (see Walbank, C 1, 153 and Hammond 1967, 599), in which case εἰς τοὺς εἴσω τόπους translated ‘interius’.

¹³ The source of Appian was a Latin writer but different from that followed by Polybius; see Hammond 1968, 5 n. 16.

¹⁴ Identified by tile-stamps ΔΙΜΑΛΛΙΤΑΝ (genitive plural); see Dautaj, 150 f. and Hammond 1968, 14 f., attributing its foundation to Pyrrhus or to Apollonia, and now supported by evidence of a prytanis as at Apollonia (Iliria 1984. 2, 85 n. 30). The ethnics were locally ‘Dimallitas’ and ‘Dimallas’ (Bul. Ark. (Tirana) 1974, 71 f.) and at Oropus ‘Dimallites’.

¹⁵ See Hammond 1981, map 12 and HM III, 409.

¹⁶ See Hammond 1967, 600 and more fully HM I, 46; Papazoglou 1970, ‘on lit d'ordinaire “Atintania”’; Kirsten, E. in Philippson, A., Die griechischen Landschaften 11, 1 (1956), 211Google Scholar proposed ‘Antania’, which has the defect of being three letters shorter than the corrupt text.

¹⁷ ‘Atintas’, as progenitor of the Atintani, was made a son of Macedon, most probably during the period of Macedonia's independence.

¹⁸ BCH 21 (1897), 161 f. = Spomenik 77 (1934), 32; see Hammond, HM 1, 76 f. with further references. ‘Antania Gemindou’ appears after Heraclea Lyncou in Hierocles, Synecdemus 639/2a, probably as a towncentre of the Antanoi (see Hammond, HM I, 86 n. 3).

¹⁹ Holleaux, 109 f. and n. 1 on 110–11.

²⁰ Lévêque, P., Pyrrhos (1957), 184 n. 7Google Scholar.

²¹ Cabanes, 78 ff., on which see JHS 97 (1977), 208.

²² Although I disagree with their views in this instance, I had the pleasure of their friendship in 1972, and I am full of praise for the advances they and other Albanian scholars have made in our knowledge of the ancient sites in Albania.

²³ See, for instance, Ceka, 67 fig. 8 and 80.

²⁴ Published in PAE 1965, 59.

²⁵ In Iliria 6 (1976), 132 f.

²⁶ In Iliria 1982. 2, 85.

²⁷ In Iliria 1984. 2, 83 f. and his maps on pp. 62 and 66.

²⁸ Examples of a koinon, for instance, in Eph. Arch. 1914, 239 no. 20 (Aterargi); SGDI 1346 (Amymni); 1334 and 1590 (Molossi); 1370 (Thesproti); SEG 24 (1969), 450 and 451 (Epirotes); Inscr. Magn. 32, 1. 50 (Epirotes). Papazoglou 1970, 126 doubted my belief (1967, 656) in a koinon of the Epirotes; but inscriptions have now proved me correct.

²⁹ See inscriptions cited in Hammond 1967, 528 and 564 and Cabanes 555, 568, 576 and 581, for instance.

³⁰ IG v, 1. 28 recorded the greetings of the ‘Bylliones’ the Lacedaemonians; Arch. Eph. 1925–6, 26 no. 140 for the ‘Byllion from Nicaea’, first century B.C. Nicaea (Klos) was placed by Steph. Byz. s.v. ‘in Illyris’ which, being a Hellenistic form, tends to confirm the statement in Strabo that the Bylliones were Illyrians. Byllis (Gradisht) had its own ethnic, Byllideus, and coined separately from the Bylliones (see H. Ceka, Questions de numismatique illyrienne (1972), 135); it was by contrast a Greek city, being visited by sacred envoys (see Hammond 1980, 13), and may have been founded by Pyrrhus.

³¹ In the order of words ἀπὸ Βουλινῶν goes not with Άμαντιεῖς but with μέχρι ἐνταῦθα Ίλλυριοί.

³² This is an embarrassment to Cabanes, who placed the Atintanes on the right bank of the Aous north of the junction with the Drin (his p. 80) and found himself forced to suppose that Molossia gained the whole Drin valley from Chaonia in 429 B.C., ‘momentanément’ (p. 13). Ceka, 83 n. 20 is confusing (it places the Atintanes west of the Chaones, i.e. in the sea; evidently the translation into French is at fault).

³³ Although disputed by N. Ceka, the evidence is decisive; for no one carries voting-discs to a place where they have no relevance. For the site, called Yerma, see Hammond 1967, 209 f., Budina, 276 (discs) and Hammond 1971.

³⁴ See Hammond 1967, 682 f. for this location athwart the main route from Molossis to Chaonia, and 218 for their fortified sites, of which one is near Lia, well known through the book of N. Gage, Eleni. It was maintained in Hammond 1980, 12 that ‘Artichia’, visited by sacred envoys c. 360–355, was in the area of the Atintanes (with map p. 17).

³⁵ See Hammond 1967, 117 f., 682 and fig. 29 c.

³⁶ See Hammond 1967, 468 f.

³⁷ Thucydides, Ps.-Scylax and Ps.-Scymnus are unanimous in calling the Epirotes ‘barbarians’; so too a Greek writer, or the Apolloniates themselves, called the (to me) Epirotic Amantini and the Illyrian Bylliones ‘barbarians’ (Pliny, HN 3. 23. 145), without implying that both were Illyrian, as Ceka, 80 n. 5 suggests. For the inscriptions, published by D. Evangelides in Arch. Eph. 1956, 1 ff. and Hellenica 1957, 247 ff., see commentaries by Hammond 1967, 525 ff. and Cabanes, 534 ff.

³⁸ For Brygi see Hammond 1967, 468 and for the modern analogy of the ‘admixture’ of Greek-speaking and Albanian-speaking peoples ibid., 20 map 2, based on travels in 1930–9.

³⁹ As in Strabo 324 one voyages from ‘the Chaones’ to the Ambraciote Gulf ‘towards the rising sun’.

⁴⁰ Names varied in form because they were reported often orally and then by Greeks speaking in different dialects. For the Dexari see Hammond, HM 1, 94 f.

⁴¹ Other references are given by Hammond 1967, 385 n. 1 with Paus. 5. 22. 3 and Ps.-Scymnus 442–3 in the text.

⁴² As is done by Anamali, 68 f.

⁴³ The tradition of the heroes was already recorded in the sixth century at the latest (see Hammond 1967, 383 f.).

⁴⁴ Anamali, 91 ff. (Amantia); Iliria 1984. 2, 113 and Papazoglou 1986, 438 f. (Olympe).

⁴⁵ BCH 45 (1921), If. and Hammond 1980, 13.

⁴⁶ The fragments of Hecataeus are discussed in Hammond 1967, 451 f. and 471 f.

⁴⁷ Müller, C., Geographici Graeci Minores (1855–1861) and Hammond 1967, 513 and 516 f.Google Scholar (for the dates of what the geographers described).

⁴⁸ An idea of their travels is conveyed in Hammond 1980 for the periods c. 360–355, c. 330 and c. 220–189 B.C.

⁴⁹ Hammond 1967, 522. Papazoglou 1970, 134 found this emendation ‘tres convaincante’.

⁵⁰ See Hammond 1967, 679 (cf. 123 and map 2) for the head of the valley. This valley differs widely in character from the Acroceraunian coastal range, which belonged to the Chaones (124f. and compare Quayle, A., Eight Hours from London (1945)Google Scholar, whose ‘Grama Bay’ is ‘Grammata’, so called from the Greek inscriptions on the rock faces).

⁵¹ A connection with ‘Kemara’ (now Himarrë) is suggested in Hammond 1980, 14 n. 1, and the ethnic ‘Kariopos’ implies that the text is sound.

⁵² An alternative reading in the MSS is ‘Ηδωνία. This reading is supported by an inscription, not seen by me but reported in Hammond 1967, 737 no. 19 with a probable reading ‘Ηδ[ω]νεσατης. If so, ‘Edonia’ was the region named now after Margariti, south of the river Thyamis (mod. Kalamas). Comparable ethnics in-ates are Oriatas in Molossis and Geneates in Pelagonia (Hammond, HM I, 91). Papazoglou 1970, 134 n. 46 disapproved of the suggestion, but offered no other explanation of the word.

⁵³ In 1967 I placed Amantia at Klos, but later evidence and second thoughts made me move it to Ploçë in Hammond 1981, map 12. Its fertile area was up the Shushicë valley; so too Kirsten (n. 16 above), 211, ‘das Shushica-Tal das “Siedelungsfeld” der Amanten oder Abanten war’.

⁵⁴ This interpretation supports the identification of the chief spring at St Nicolas church, some five miles from Apollonia city.

⁵⁵ For instance, Callimachus, καὶ ‘Αμαντίνην ᾤκισαν Ώρικίην, cited in Steph. Byz., s.v. ‘Abantis’ fin.

⁵⁶ Apollonia was sometimes described as ἠ Ήπείρωτις, i.e. in Epirus (Strabo 764), but not as in Illyris; it was intermediate between the two.

⁵⁷ The minerals in the Mati basin and farther north do not include silver; see map 16 in Hammond 1976.

⁵⁸ Likewise Cabanes, 80, having the villages fired apparently at night and presumably expecting the cavalry and infantry to make the fifty miles next day to catch the Epidamnians in the fields. Even in the last war German troops avoided night operations in occupied Greece; they burned hundreds of villages in Macedonia, all by day.

⁵⁹ For instance, Ptol. 3. 12. 2 placed Apollonia and Epidamnus within the sphere of the Taulantii.

⁶⁰ I had this experience travelling in the 1930s and saw in 1972 pictures of the great floods of the pre-Hodja days.

⁶¹ See Hammond 1967, 224, ‘forded the river at the second attempt with much difficulty’ at Kutë upriver from Klos.

⁶² Cabanes, 80 has a simple method: his map 5, purporting to show the Epirote camp, leaves the route between the Illyrian fleet and the Illyrians in Phoenice open, and he does not mention my plan (Hammond 1967, 117 ff. and fig. 29 c, based on autopsy). Why did he send the defeated Epirotes northwards? His answer is simple: ‘leur retraite étant coupée vers le Sud’, which is inconsistent even with his own map 5.

⁶³ Papazoglou 1970, 133 found it unacceptable that Polybius should mention the Atintanian name at 2. 5. 8 (the flight after defeat at Phoenice), and at 2. 11. 11 (the Roman march north in 229 B.C.), and yet fail to warn his readers that it was being used of different peoples.

We must remember that Polybius did not know the area, had no maps and probably gathered material for events from Greek internees for 230 B.C., and from Roman friends and narratives for 229 B.C., while he himself was interned (see Walbank, C 1, 33 f.) and writing the early books of his history. He may not have realized the difference himself; and if he did, he may not have thought a digression on the subject was desirable.

⁶⁴ For details see Hammond 1968, 2 f.

⁶⁵ E. Badian in PBSR 20 (1952), 93 = Studies in Greek and Roman History (1964), 24 and Hammond 1967, 601 and 1968, 9.

⁶⁶ ‘Illyris’ in Hellenistic times seems to have extended up to the Drilon (Drin) from the Bylliones in the south.

⁶⁷ See Hammond 1968, 1 f. for the coastal plain, and Hammond 1981 map 12, based on 1974, 189 f., for the ancient road system. Map and satellite picture in 1976, 8–9.

⁶⁸ This area is described in Ancient Macedonian Studies in honor of Charles F. Edson (1981), 201 ff. and especially 205.

⁶⁹ For these last episodes see now Hammond, N. G. L. and Walbank, F. W., A History of Macedonia III (1988), 537 f. and 567Google Scholar f. The Molossians had been in Central Epirus by Dodona since early times (in epic saga, Hecataeus F 108, Pindar, Paean 6, no and Aeschylus, PV 829). Similarly the Abantes, i.e. Amantes, were next to Apolloniatis from early times (in epic saga, Paus. 5. 22. 2 and Pliny, HN 3. 23. 145), the Taulantii next to Epidamnus (Thuc. 1. 24. 1 and Eratosthenes in Steph. Byz., s.v. ‘Dyrrachium’), and Encheleae next to Dassaretae (Hecataeus F 103 with the form ‘Dexaroi’ and Polyb. 5. 108. 8 with the form ‘Enchelanes’), and so on. The pattern of tribal settlement seems to have been fixed from at least the seventh century in Illyris and Epirus.