¹ The following special abbreviations are used: Epirus = N. G. L. Hammond, Epirus (Oxford, 1967).

Hammond, Koina = id., ‘The Koina of Epirus and Macedonia’, Illinois Classical Studies, 16 (1991), 183–92.

Hatzopoulos = M. B. Hatzopoulos, Macedonian Institutions under the Kïngs (Meletemata. 22. 1, Athens, 1996).

HM 1 = N. G. L. Hammond, A History oj Macedonia, i (Oxford, 1972).

Detailed maps with ancient and modern names are available in Epirus (e.g. maps 7, 12, and 16) and in HM 1 (e.g. maps 6, 9, and 11); and a general map with ancient names in N. G. L. Hammond (ed.), Atlas aj the Creek and Roman World in Antiquity (Princeton, 1981), map 12. The indexes of these works include references to the maps for the names of places.

² My concern is with the instances of ἔθνος in inscriptions and in ancient authors and neither with the theories of what may have constituted ‘ethnicity’, nor with ‘the post-war circumspection concerning ethnicity’, of which J. M. Hall has given an account in PCPS 41 (1995), 83–100Google Scholar.

³ As I argue in Epirus, 447 55 and 702.

⁴ The form Ἄραθθος occurs in Tod GHI i no. 2, dated c. 600 BC.

⁵ There are no palaeographical grounds for emending these passages, as in the Loeb edition.

⁶ LSJ s.v. III translates Φῦλον as ‘clan or tribe according to blood or descent’ and cites Iliad ii. 362. If Strabo was quoting Hecataeus verbatim, Hecataeus believed that the Talares were a racial unit. The word fulon in this sense seems to be archaic.

⁷ See Epirus, 451 2 for this meaning of ‘Epirotes’.

⁸ As I found in the winter of 1943 4. It was then a dangerous task for a runner to cross the Pindus range from Epirus to Macedonia.

⁹ I have described it from my own experience in Epirus, 25–7, HM 1, 14–16 and more fully in my book Migrations and Invasions in Greece and Adjacent Areas (Princeton, 1976), 37–51Google Scholar, where references will be found 10 earlier works. To these may be added Winnifrith, T. J., The Vlachs (London, 1987)Google Scholar and Forbes, H., ‘The identification of pastoralist sites within the context of estate-based agriculture in ancient Greece’, BSA (1995), 325–58Google Scholar.

¹⁰ Anon., , Descriptio Europae Orientalis, ed. Górka, O. (Kraków, 1916), 25Google Scholar.

¹¹ So also Campbell, J. K., Honour, Family and Patronage (Oxford, 1964), 41Google Scholar, Who described a company of Sarakatsani as ‘a number of autonomous but related families joining together for the co-management of their Hocks … founded upon ties of kinship and marriage’.

¹² See Epirus, 467 8; Mt. Amyron is to be identified with Mt. Tomor, as in Epirus, map 1.

¹³ The majority of the tablets record the personal questions of people from this area. Sec the report of Dakaris, S., Christidis, A. Ph., and Vokotopoulou, J., ‘Les lamelles oraculaires de Dodone et les villes de l'Épire du Nord’, L'lllyrie méridionale et l'Épire dans l'antiquité, ii (Paris, 1993), 56–60Google Scholar, and especially n. 30 ‘les inscriptions oraculaires … fournissent un trésor linguistiquc du Nord-Ouest, datant du Vie siècle à la fin du me siccle’. The name of the magistrates in Seleucid Syria ἀδειγᾶνες, in Plb. v. 54. 10 has been confirmed by an inscription mentioning πελιγᾶνες in Syria, 23 (1942–1943), 21–2CrossRefGoogle Scholar. These words in the North-West dialect were evidently brought from Macedonia to Syria. So too Hesychius s.v. Πελιγᾶνες.

¹⁴ They were published by D. Evangelides in Eph. Arch. 1956, 1–2. They were discussed and dated by me in Epirus, 525–31; see also my article Koina, 183–4.

¹⁵ First published by D. Evangelides; cited and discussed in Epirus, 527–8.

¹⁶ Franke, P. R., Alt-Epirus und das Königtum der Molosser (Erlangen, 1954), 287Google Scholar, held that the Eurymenaei were a Molossian tribe. See Epirus, 526–7.

¹⁷ For instanee near Fiorina Keramopoullos excavated what proved to have been an open village (PAE 1931, 55–6). In a survey of Lynccslis L. Gounaropoulou and M. B. Hatzopoulos reported traces of thirteen ancient settlements in Les Milliaires de la Vow Egnalienue entre Héhaclee des Lyncestes el Thessalomque (Melctemata, 1; Athens, 1985)), 17 n. 3Google Scholar.

¹⁸ I have discussed in CQ. n.s. 49 (1999) 249–50Google Scholar various objectinons; e.g. those of of P. A. Bruont reaching down to his Locb edtion. He regarted the speech as an ‘epideictic display by Arrian’ with ‘exaggerations and absurdities’ and made comments in his notes at vol. ii. 228–9.

¹⁹ Vlachs in my experience wore athick hooded garment resching down to the feet. it was made of thick felt and rendered waterproof by a mixture of goat-wool and sheep wool. When I spent a day in pouring rain m such a garment, it felt like a tent and kept me dry. Known as a διΦθέρα of Alexander's apeech. A form of it, worn over the head, is shown in an ancient relief in Sakellariou, M. B. (ed.), Macedonia Athens, 1983), 85Google Scholar fig. 50. On the other hand Philip's ‘cloaks’ were short and open, Hying back from the shoulders m action, as ibid. 87, 129, and 139.

²⁰ In the Greek text the word ἀπέφηνε. here translated ‘made’, has the meaning rather of ‘declared’; for it expressed Philip's intention and not the extent to which he implemented it in his lifetime.

²¹ The site at Gradista near Petres was identified with Celle by M. B. Hatzopoulos and me (see AJAH 7 (1982), 137Google Scholar and Hatzopoulos 1, 94 n. 4); according to Blackman, D. J. in AR 44 (1997–1998), 88Google Scholar ‘it may have been founded m the second half of the fourth century BC’. For a summary of the excavations there see Adam-Veleni, P. in Δἐκα Χρὁνια Αρχαιολογικὀ Εργο στη Μακεδονἱα και Θρἀκη (Thessalonike, 1997), 6–8Google Scholar.

²² Listed in Epirus, 659–61 with approximate measurements.

²³ See the preceding note, and P. Cabanes in ΑΦιέρωμα στον N. G. L. Hammond (Makedomka, Parartema 7; Thessalonike, 1997): 99. The number of ethnies in the whole of Kpirus and Upper Macedonia must have numbered some hundreds.

²⁴ In BSA 48 (1953), 135–40Google Scholar = my Collected Studies 2 (Amsterdam, 1993), 69 74Google Scholar with plates.

²⁵ The best text is in Rizakis, A. and Touratsoglou, J., Επιγραφἐς Ανω Μακεδονἰας, 1 (Athens, 1985), no. 186Google Scholar with a full bibliography on p. 169, to whieh should be added Buraselis, K. in Ancient Macedonia, v (Thessaloinke 1993), 279 92Google Scholar and Hatzopoulos 1, 79–82.

²⁶ Spomenik, 71 (1931), 110Google Scholar. 63 and no. 88.

²⁷ Ibid. no. 437 (Dostoneis) and BCH 47 (1923). no. 277Google Scholar (Vitolište).

²⁸ Ibid. no. 339. See also HM 1, 70 and 89, and Hammond, Koina, 187–8.

²⁹ In ‘Sur les koina regionaux de la Haute Macedoine’, Ziva Antika, 9 (1959), 170Google Scholar, she wrote of ‘les grandes tribus’.

³⁰ These traditions were no doubt handed down by the priestesses and the priests ot Dodona, the oldest shrine in Greece. They were collected and probably reorganised by Rhianus and by Theagenes, writers of the Hellenistic period (for Rhianus see Epirus, 701 and for Theagenes see HM 1, 31–2). Theopompus named Helenus, son of Priam, as the ancestor oi the Chaonian royal family (FGIH 115 F 335).

³¹ The ancient evidence does not enable us to see how in other areas a collection of tribal communities became associated to form a state. In Upper Macedonia each canton was self-sufficient in that summer pastures on the surrounding mountains; for instance Pelagonia and Derriopus were enclosed by Mt. Plakenska, Mt. Babuna and Mt. Barnou (see HM 1 maps 1, 6 and 9). Hear the geographical conditions may have helped the association of the comunities to froma state. But that was not so in Epirus.