¹ Besides the common abbreviations I use the following: Th. Axenidhis, D., PL = Ἡ Πελασυὶς Λάρισα (1947–9)Google Scholar; Beloch, K. J., GG = Griechische Geschichte ² (1912–27)Google Scholar; Burn, A. R., PG = Persia and the Greeks (1962)Google Scholar; Ehrenberg, V., SS ² = From Solon to Socrates ² (1973)Google Scholar; Hammond, N. G. L., HM i = A History of Macedonia i (1972)Google Scholar; Hignett, C., XIG = Xerxes' Invasion of Greece (1963)Google Scholar; How, W. W. and Wells, J., CH = A Commentary on Herodotus (1912)Google Scholar; Kromayer, J., AS ii = Antike Schlachtfelder in Griechenland ii (1907)Google Scholar; Larsen, J. A. O., GFS = Greek Federal States (1968)Google Scholar; Meyer, E., FAG = Forschungen zur Alten Geschichte (1899)Google Scholar; id., TH = Theopomps Hellenika (1909); Obst, F., FX = Der Feldzug des Xerxes (1913)Google Scholar; Pritchett, W. K., AGT ii = Studies in Ancient Greek Topography ii (1969)Google Scholar; Sordi, M., LT = La Lega Tessala (1958)Google Scholar; Westlake, H. D., TFC = Thessaly in the Fourth Century B.C. (1935)Google Scholar.

² According to Larsen, , GFS 117Google Scholar, n. 1, ‘Plutarch has preserved a correct tradition to the effect that Themistocles was opposed to the Thessalian venture from the outset’; others too lay weight on Plutarch's statement. Plutarch however recounts the expedition and its consequences in just over 25 words, beginning ‘because of great opposition he led out a great force’, and this off hand way of speaking suggests to me that the biographer has chosen the simplest way of mediating between Themistocles' celebrated naval policy and the Thessalian expedition.

³ Ch. 31–3 of Plutarch's essay deal with Herodotus' treatment of the Thebans, and it has often been conjectured that much of the material comes from Aristophanes of Boeotia. Jacoby, on FGH 379Google Scholar F 5–6 sceptically remarks that ‘Plutarch's fluent polemic operates with very few facts’, but the 500 troops and the general Mnamias are among the few; they were not invented by Plutarch, nor yet, one would think, by his source.

⁴ The Suda makes Damastes ‘a pupil of Hellanicus’. It is sometimes said that Damastes ‘agrees’ with Hellanicus (so D.H. i 72.2 and Plin. h.n. vii 154, whence V. Max. viii 13 ext. 6), and in several catalogues of geographical authorities Damastes comes after Hellanicus. On the other hand Porphyry declared that Hellanicus' treatment of Scythia was taken from Herodotus and Damastes (Eus. p.e. × 3). On this evidence it is safer to conclude only that the works of Damastes and Hellanicus showed similarities which suggested indebtedness to later writers, and that priority was usually awarded to the more famous name. The likeness between the Scythian portions of their respective works may be due to common use of Aristeas' Arimaspeia (cf. Steph., Hyperboreioi = FGH 4 F 187 and 5 F 1Google Scholar), a widely current work which stands behind Hecataeus too and probably accounts for Agathemerus' statement that Damastes ‘copied out’ Hecataeus.

⁵ Diotimus (PA 4386) was general in 433/2, commanding at Corcyra. His embassy to Susa almost certainly antedates the Peloponnesian War: it may be Aristophanes' target at Ach. 61–90, where the date 438/7 is given, perhaps capriciously, for the departure of the ambassadors; and it may be the occasion of Pyrilampes' obtaining the peacocks that became a spectacle at Athens for many years (‘more than 30 years’, if we believe Antiphon fr. 57 Thal., delivered before 411 at the latest). The victory which Diotimus gained at Naples while commanding an Athenian fleet probably fell in the 430's, during the Sabellian invasion of Campania (so Meyer, , FAG II 321–2Google Scholar, overlooked by Jacoby, at FGH 566Google Scholar F 98, who rejects several inferior suggestions), unless it was even earlier, perhaps during the raids of Etruscan pirates, as J. K. Davies suggests to me.

⁶ F 1, cited from Peoples and Cities, lists the fabulous neighbours of the Scythians—the same as in Herodotus. Other likely fragments of this work are F 2, 5, 8, 9, 10. Damastes is scouted by Strabo for loose talk, λῆροι, but Strabo's own citation proves him captious: Damastes explicitly queried Diotimus' geography.

⁷ Meyer, E., Geschichte des Altertums iii ² (1937) 365Google Scholar; Toepffer, , RE i 1 (1893) 1373Google Scholar, s. Aleuadai.

⁸ Munro, , CAH iv 282Google Scholar.

⁹ Westlake, H. D., JHS lvi (1936) 16–17Google Scholar; Hignett, , XIG 102Google Scholar; Larsen, , GFS 115Google Scholar.

¹⁰ The formula is Burn's, , PG 341Google Scholar; cf. Ehrenberg, , SS ²153Google Scholar.

¹¹ Sordi, M., RIL lxxxvi (1953) 309–10Google Scholar; cf. Beloch, , GG ii 1²41–2Google Scholar; Axenidhis, , PL i 92–4Google Scholar.

¹² Meyer, , FAG ii 212Google Scholar; Munro, , JHS xxii (1902) 305Google Scholar.

¹³ So too at Ctes. FGH 688 F 13, §27.

¹⁴ A life-long Tageia is argued by Meyer, , TH 220–2Google Scholar; Beloch, , GG i 2²200–1Google Scholar; Gschnitzer, F., AA vii (1954) 191–2Google Scholar; Sordi, , LT 334–9Google Scholar; and Larsen, , CP lv (1960) 238–9Google Scholar and again GFS 14–16, 19. On the other hand Cary at CAH iii 602, Westlake, , TFC 25–6Google Scholar, Morrison, , CQ xxxvi (1942) 59Google Scholar, and Axenidhis, , PL i 88Google Scholar, envisage a short-term Tageia, but Cary mistakenly relies on SIG ³ 55 (on which see Chadwick, J. in Studi Ling. V. Pisani i [1969] 231–4)Google Scholar, and Westlake and Axenidhis adopt the paradoxical view that the constitutional term was ‘usually’ prolonged for an indefinite period or even for life.

¹⁵ In discussing the Aesymneteia as an ‘elective tyranny’, Aristotle says that whereas some such leaders were appointed ‘for certain periods or certain undertakings’, others ‘held this office for life’ (pol. iii 14, 1285a 33–5). Presumably the life-long Aesymneteia was always a desperate resort: it is hard to believe that any Greek state made constitutional provision for electing a life-long, absolute ruler.

¹⁶ Daochus' Tageia ended in or before 404, when Lycophron aimed to dominate the whole of Thessaly (X. Hell. ii 3.4) : on any possible view, then, it overlaps the first part of the Peloponnesian War. including the campaign of 431 which saw a Thessalian contingent fighting in Attica (Th. ii 22.3, where Gomme errs). If the outbreak of the War is allowed to be the likeliest occasion for Daochus' appointment, his 27 years' tenure shows that the office was conferred for the duration of hostilities. The epigram describes Thessaly as ‘teeming with peace and wealth’ under Daochus—the best compliment for his military inactivity.

¹⁷ Hypothetical Tagi are mustered in large numbers by Meyer, , TH 237–49Google Scholar; Beloch, , GG i 2²197–210Google Scholar; Axenidhis, , PL i 84–91Google Scholar; and Buck, R. J., CP lxiii (1972) 95–6Google Scholar. Of their instances, apart from Daochus, I admit only Eurylochus (as a fictional Tagus), Cineas, and Lattamyas; the only name I add is Cleomachus.

¹⁸ Hdt. viii 27.2–28; Plb. xvi 32.1–2; Plu. mor. 244A–D; Paus. x 1.3–11, 13.4, 6–7; Polyaen. vi 18.2, viii 65.

¹⁹ Beloch, , GG i 2²205Google Scholar, was wary of claiming Lattamyas as a Tagus, but according to Buck, , CP lxvii (1972) 95Google Scholar, ἄρχοντα ‘should mean that he was a tagos, not a subordinate official, a ruler, not simply a local commander’. It is hard to see why.

²⁰ A late grave-stone from Crannon addresses Λάτταμϵ Ἐχϵκρατίδου (IG IX 2, 469), perhaps the only other instance of the name. Echecratides is not rare, and no connexion with the ‘Echecratids’ of the fifth century can be entertained.

²¹ Robert, L. in Bernard, P.et al., Fouilles ď Ai Khanoumi i (1973) 219–22Google Scholar.

²² The details given by Plutarch do not favour Jacoby's assertion (on FGH 427 F 5) that the name Cleomachus ‘was obviously invented along with the story’.

²³ So Meyer, , TH 242, 254Google Scholar; Beloch, , GG i 2²201Google Scholar; Axenidhis, , PL i 41Google Scholar.

²⁴ So Meyer, , TH 237Google Scholar; Beloch, , GG i 2²200Google Scholar; Axenidhis, , PL i 85, 102Google Scholar; Gomme on Th. ii 22.3; Sordi, , LT 335Google Scholar.

²⁵ In pol. iii 14–17 the professed subject is Basileia, and Aristotle uses this term consistently except when he comes to discuss Aesymneteia as the third type— not of Basileia, but of Monarchia (iii 14, 1285a 30); after Aesymneteia he effects the transition to the fourth type, which is again true Basileia, by the phrase μοναρχίας βασιλικῆς (1285b 4). Aristotle could not bring himself to apply the term Basileia to an elective, and in most cases temporary, magistracy.

²⁶ A good example is Arrian, , Ind. 5.3Google Scholar, ‘Sandracottus, the greatest Indian ruler’, etc.

²⁷ Or rather, exists only in the imagination of writers who recount the novella of Thargelia. The story in all its bearings was properly discounted by Hiller, at RE VI A 1 (1936) 118Google Scholar, s. Thessalia, but some details continue to be accepted at face value, as by Sordi, , LT 330Google Scholar n. 3. Those who believe Antiochus King of Thessaly will presumably also believe that Thargelia survived him as Queen of Thessaly for thirty years, which comprehended Xerxes' invasion (Phot, and Suda Thargelia).

²⁸ Of the events of 511–10 Herodotus says that the Thessalians, acting in concert, ‘sent a thousand horse and their king, Cineas a man of Conia’ (v 63.3). The emphatic phrase τὸν βασιλέα τὸν σφέτϵρον, which seems to have gone unnoticed by commentators and translators, is chosen because the Spartans, by contrast, did not at first depute a king to lead the Spartan force (v 63.2, 64.1). Of course the phrase need not mean that Cineas was the only king in Thessaly. Cineas, as we have seen, may have been a Tagus, but we cannot know whether Herodotus called him a king qua Tagus: if he did, he showed his usual ignorance or indifference concerning constitutional forms.

²⁹ The passages of Pindar and Herodotus often adduced in this connexion do not in the least imply that the Tageia belonged to the Aleuads. In describing the whole Aleuad family as ‘Thessalian kings’ Herodotus vii 6.2 plainly means hereditary rulers: he could scarcely speak thus even if the Tageia had devolved on two or three Aleuads in succession. The words which Pindar addressed to his Thessalian patrons in 498—‘the skilful piloting of cities is the patrimony of good men’ (P. x 71–2)— mark their power as regional, not federal (so Hiller, at RE VI A 1 [1936] 118Google Scholar). The opening lines of P. x, and also lines 69–71, merely show that the nobles of Thessaly, including the Aleuads, claimed descent from Heracles and of course ruled their domains by hereditary right; we need not infer that Pindar puts the Thessalian Tageia on a footing with the Spartan kingship, still less that the Aleuads then held or aspired to the Tageia. All the wrong conclusions were drawn by Wilamowitz, , Pindaros (1922) 123Google Scholar—who thought that the Tageia passed to the Aleuads after the Scopads collectively came to grief beneath a collapsing house!

³⁰ In the fifth century Larisa minted an abundant coinage (surpassing Pharsalus as well as all other centres), and the types were widely imitated in northern Thessaly. This state of affairs has been taken to reflect the primacy of the Aleuads and of Larisa in the Thessalian federal state (so Herrmann, F., ZN xxiii [1923] 33–43Google Scholar and xxv [1925] 1–69). Herrmann's interpretation of the numismatic evidence was dictated by the prevailing view of Thessalian history and (it should be clearly recognised) is by no means mandatory. Larisa always had close ties with Macedon and the north (as we shall see below) and hence access to the silver supply in Thrace; it was therefore natural that she should take the lead in minting coins.

³¹ Since the port of Halus was Pharsalian (Str. IX 2, 5.8, 433c), Itonus, only sixty stades away, must have been Pharsalian too. The federal worship of Athena Itonia happens to be first attested in Hellenistic inscriptions (Boesch, P., Θϵωρός [1908] 28Google Scholar, line 2; Segre, , RFIC [1934] 172Google Scholar, B2, line 6; SEG xxv 653, lines 20–1), but undoubtedly existed long before.

³² A fifth-century decree of Thetonium (IG IX 2, 257 = SIG ³ 55, lines 7–8) provides the sole mention of a local Tagus before the Hellenistic period, when Tagi are attested for various parts of Thessaly. The site of Thetonium at Kupritsi lies about 15 km. SW of Pharsalus. At Delphi the title Tagus was used of the head of a phratry.

³³ In Pharsalus itself the federal office-holders are not the Meno's, famous in literature as the great landowners of the area, but the family of Daochus, whom we know mainly from their several monuments at Delphi and Pharsalus.

³⁴ The reason given by Morrison, (CQ xxxvi [1942] 61–3Google Scholar) for the Athenian attack on Pharsalus, namely that the city lay on their road, is not sufficient in itself, for we can hardly suppose that the Athenians meant to reduce, in geographical order, every important town in Thessaly; and had their main objective lain elsewhere, they could have skirted Pharsalus by starting from the Gulf of Pagasae.

³⁵ The phrase might also denote all Thessalian authorities, whether dynasts or magistrates, who remained in power during the invasion. In either case we cannot determine whether the federal apparatus continued to function after the Greek withdrawal. The charge of medism which Herodotus brings against the Thessalian leaders at ix 1 is of course consistent with his account at vii 174 of the mass reaction when the Greeks left.

³⁶ The Theban speaker at Th. iii 62.3–4 excuses the Thebans for medising on the grounds that a few powerful medisers controlled the state.

³⁷ Beloch, , GG ii i²42 n. 1Google Scholar, not unreasonably dismissed the story as a ‘slander’, but added a false argument, namely that the Aleuads already ruled Thessaly and hence had nothing to gain from collaborating with Xerxes.

³⁸ So Westlake, , JHS lvi (1936) 12Google Scholar; Sordi, , RIL lxxxv (1953) 297–8Google Scholar.

³⁹ In place of Εὐαίνϵτος the mss. of D.S. give Συνϵτός, a most unlikely name; and since proper names are so often corrupted in Diodorus, no one will hesitate to accept the obvious correction.

⁴⁰ Ehrenberg, , SS ²153–4Google Scholar. Cf. Obst, , FX 52–5Google Scholar; Munro, , CAH iv 281–2Google Scholar; Larsen, , GFS 117Google Scholar.

⁴¹ Hignett, , XIG 103Google Scholar.

⁴² De Sanctis, , RFIC viii (1930) 339–42Google Scholar; Sordi, , RIL lxxxvi (1953) 299–310Google Scholar, 323, and LT 92–6.

⁴³ Cataudella, M. R., Athenaeum xliii (1965) 389–90Google Scholar.

⁴⁴ Schachermeyr, , JOAI xlvi (1963) 169Google Scholar.

⁴⁵ Diodorus asserts that reinforcements were sought, but he cannot be trusted. In his maundering way he tells us first that ‘the Greeks’ dispatched 10,000 hoplites to Thessaly under Euaenetus and Themistocles, and then that ‘these’, sc. Euaenetus and Themistocles, sent ambassadors to other Greek cities requesting further levies (xi 2.5). Whether or not this sequence of events is derived from Ephorus, it scarcely merits the attention which it receives from Larsen, , GFS 115Google Scholar.

⁴⁶ Larsen, , GFS 17–18Google Scholar, considers 6,000 a paper figure and suggests 3,000 as the reality; Westlake, , TFC 108–9Google Scholar, says 4,000; but Wade-Gery, , JHS xliv (1924) 62Google Scholar thinks it ‘not impossible’ that Jason's estimate was sound. In fact Jason as Tagus subsequently had under arms ‘more than 8,000 cavalry including the allies’ (X. Hell. vi 1.19): the reputed size of Jason's army is roughly vindicated by Westlake, , TFC 105–12Google Scholar, and the allies would hardly account for more than a quarter of the cavalry force. The Thessalian cavalry who accompanied Alexander or who sided with the Greeks in the Lamian War numbered 2,000, but there is no reason to think that they were a general levy, and in any case Thessaly was not then the power that it had been in Jason's day, to say nothing of the late Archaic period. Moreover, in the fourth century a large proportion of Thessalians served as hoplites (Hell. vi 1.8 and 1.19); these were the class of smallholders settled round the cities which had grown up since Archaic times.

⁴⁷ This would still be true if we restricted our view to the 10,000 hoplites. The only larger figure that Herodotus reports before Plataea is the paper strength of the Siceliot Greeks, as expounded by Gelo to the allied ambassadors: 20,000 hoplites (vii 158.4). Although this figure was repeated by Timaeus (Plb. xii 26b = FGH 566 F 94)—who was capable of imagining a Greek army of 50,000 foot at Himera (D.S. xi 21.1)—Ephorus reduced it at one stroke to 10,000 (sch. Pi. P. i 146a, b = FGH 70 F 186: Brunt, P. A., Historia ii [1953] 159Google Scholar n. 2 regards 10,000 as a ms. corruption of 20,000, but the figure is given twice). It is in the nature of things that such estimates should be grossly inflated: the paper strength of 8,000 ‘shield’ for Naxos (Hdt. v 30.4) is simply a physical impossibility. Casualty figures may be doubted too, but certain instances are instructive. The 6,000 Argives fallen at Sepeia (Hdt. vii 148.2: Paus. iii 4.1 says 5,000 and an Argive tradition cited by Plu. mor. 245D says 7,777) represents the full citizen levy of one of the most populous mainland states fighting on native soil. In a calamitous defeat at the hands of the Iapygians Rhegium lost 3,000 men and Tarentum many more, ‘the greatest slaughter of Greeks of all that we know of’ (Hdt. vii 170.3): thus the army fielded by two leading states of Magna Graecia might have numbered 10,000. On this showing the force which went to Thessaly is extremely impressive. In their maximum effort at Plataea the Greeks were serving much closer to home—especially the Athenians, whose contingent of 8,000 must have included every available man. Larson, , GFS 114–16Google Scholar, correctly observes that the army deployed in Thessaly was superior in some respects to the army of Plataea.

⁴⁸ Westlake, , JHS lvi (1936) 18 n. 28Google Scholar, thinks that ‘since the Greeks would hardly wait idly for some two months at Tempe’, they did not arrive until Xerxes was in Thrace.

⁴⁹ The Persians reached Attica three months after crossing the Hellespont (Hdt. vii 51.2): the date of their arrival in Attica is disputed, however, and to argue the matter fully would take more space than is warranted, since the precise date of the Thessalian expedition has small importance.

⁵⁰ Westlake, , JHS lvi (1936) 17Google Scholar, says that the story of the Thessalian expedition ‘takes the reader by surprise’; ‘casual and almost parergic’ was Macan's label, which Westlake quotes with approval: I can see no basis for these judgments, except so far as Herodotus' narrative is continually surprising and has the air of artless reminiscence. It is certainly wrong to argue from such impressions that the actual expedition was ‘a hastily improvised scheme’ (so Westlake).

⁵¹ From 170 to 31 B.C. we hear of five successive Roman armies that were quartered and provisioned in Thessaly (Westlake, , TFC 6Google Scholar).

⁵² Hignett, , XIG 103 n. 1Google Scholar, following Grundy, holds that the allies went by sea only because the Boeotians could not be relied upon. But if ships were to hand, the sea passage through the sheltered waters of the straits of Euboea was preferable to the land route in every way—shorter, quicker, easier.

⁵³ Larsen, , GFS 115Google Scholar, wrongly assumes that contingents marching from states close to Thessaly, such as Boeotia, Phocis, and Locris, were additional to a sea-borne force of 10,000.

⁵⁴ Buck, , CP lxix (1974) 48Google Scholar.

⁵⁵ Pritchett, W. K., AJA lxv (1961) 373–5Google Scholar; Helly, B., Gonnoi (1973) i 10–11Google Scholar.

⁵⁶ The Roman route is traced by Pritchett, , AGT ii 164–76Google Scholar, and more convincingly by Helly, , RP xlvi (1972) 276–82Google Scholar.

⁵⁷ Kromayer, , AS ii 270Google Scholar (a brief description); Pritchett, AGT ii pl. 146 (a photograph taken within the pass).

⁵⁸ Students of Xerxes' invasion almost invariably describe this route as reaching the Thessalian plain by way of the Melouna pass, which crosses a ridge of Olympus southeast of Olosson. In fact the normal route in ancient times, and certainly the route to be preferred by the Persian army, led through the Europus valley: see Wace, A. J. B. and Thompson, M. S., Prehistoric Thessaly (Cambridge 1912) 7Google Scholar.

⁵⁹ Hammond, , HM i 117Google Scholar, 430 n. 2.

⁶⁰ Either the col above the Haliacmon valley, known as Stena Portas, which forms the northern entrance to the pass, or the pass itself, now called Stena Sarandaporou, is easily defensible (Hammond). It is also possible to get from Macedon to Thessaly by ascending the Haliacmon even further and then striking south to Aeginium and Tricca or southeast to Perrhaebia (Westlake, , TFC 18–19Google Scholar and Hammond, , HM i 109, 118)Google Scholar; but these routes are so long and difficult that they can be ignored for the present purpose.

⁶¹ Macan, R. W., Herodotus, Books vii, viii and ix (London 1908) i 164–5Google Scholar; How and Wells, , CH ii 175Google Scholar; Munro, , CAH iv 2921Google Scholar; Hignett, , XIG 110Google Scholar.

⁶² Pritchett, , AJA lxv (1961) 369–75Google Scholar. Although Pritchett's avowed aim is to save Herodotus from his critics, Herodotus is in fact misrepresented; the key passages vii 128.1 and 131 are not even mentioned in Pritchett's article, and the treatment of ‘Upper Macedonia’ also ignores viii 137.2. Ehrenberg, , SS ²153Google Scholar and 425 n. 49, professes to accept Pritchett's conclusions, but his narrative shows that he has not envisaged the route correctly.

⁶³ Pritchett argues at length that the coast road as far as Lower Olympus, followed by the track over the mountain, is easier as well as shorter than either the Petra or the Volustana routes; but he never compares the mountain track with Tempe.

⁶⁴ How and Wells, , CH ii 175Google Scholar, suppose that Herodotus' account of ‘the great labour involved in cutting a road’ points to the track over Lower Olympus; but Herodotus clearly thinks of this operation as taking place far from Tempe and over a much longer stretch than Lower Olympus (‘the Macedonian mountains’, vii 131). Moreover, the passage which How and Wells adduce from Livy as describing the difficulty of the track (xliv 3.3) actually refers to an area much further west—the initial stage of the Roman march from Tripolitis past Mount Otolobus to Lower Olympus—and so has nothing to do with any route used or in prospect of being used by the Persians in 480.

⁶⁵ Pritchett, , AJA lxv (1961) 375Google Scholar, contends that the term signified something entirely different to Herodotus, namely the coastal plain of Dium north of the massif of Heracleium, whereas Lower Macedonia is the narrow coastal plain south of Heracleium beside Lower Olympus. The latter is a miniscule region which could scarcely be treated as a geographic entity and may not even have belonged to Macedon in the fifth century; and the notion of Upper Macedonia as a coastal zone is excluded by Herodotus vii 128.1 and viii 137.2.

⁶⁶ It is no objection that Herodotus, when thinking of the Peneius as the sole outlet from the Thessalian basin, describes Olympus and Ossa jointly as ‘Thessalian mountains’ (vii 128.1). If Herodotus thought that a route starting in Upper Macedonia could emerge at Gonnus, he plainly had no inkling of the geography of the Macedonian interior.

⁶⁷ Hammond, , having walked from Verria to Servia and again from Servia to Elassona, firmly excludes this route for the Persian army (HM i 430 n. 2)Google Scholar.

⁶⁸ From a close reading of Hdt. vii 183.2–3, 192.1, 193.1, 196, it would appear that Xerxes and his army got from Therma to Malis in 14 days; but this reckoning cannot be accepted on any view of Xerxes' route.

⁶⁹ The European contingents of Xerxes' army listed at vii 185.2 include Eordians, Bottiaeans, Brygians, Pierians, and Macedonians proper, but not Elimiots (first mentioned in literature at Th. ii 99.2).

⁷⁰ How and Wells, , CH ii 370–1Google Scholar; Munro, at CAH iv 282Google Scholar; Westlake, , JHS lvi (1936) 19Google Scholar.

⁷¹ In the previous year A. Hostilius Mancinus had led his army through the Volustana pass to Elimeia. Livy's report of this episode has been lost after xliii 3.7, but we can be sure that the pass was not defended.

⁷² To the original force of perhaps 37,600 Marcius added another 5,000 in spring 169 (Kromayer, , AS ii 340–8)Google Scholar.

⁷³ Livy hereabouts has simplified Polybius without understanding the topography, and at xliv 2.6–8 writes as if Tripolitis would also be a natural startingpoint for a route over Lower Olympus (‘past Lake Ascuris’). It was much too far to the northwest. Of course Marcius had to evade the Macedonian post at Gonnus, but there was no need to go further north or west than Olosson at the outside. Pritchett, , AGT ii 170Google Scholar, Helly, , RP xlvi (1972) 277Google Scholar, and Hammond, , HM i 137Google Scholar, all seem to accept Livy's view of Tripolitis as a cross-roads leading equally to Lower Olympus.

⁷⁴ How and Wells, , CH ii 370Google Scholar; Burn, , PG 342–3.Google Scholar

⁷⁵ If we credit a story of Polyaenus (iv 3.23), Tempe could also be turned by way of Mount Ossa. Alexander cut steps in the vertical rock and brought his army over the peak while the Thessalians were holding Tempe: ‘even today as you go through Tempe you can see the rocks of Ossa worked into steps’. Not very plausible (despite Westlake, , TFC 217–18)Google Scholar; though of course the story could conceivably rest on the circumstance which scholars deduce from it, namely that Thessaly resisted Alexander in 336.

⁷⁶ That the Ziliana is the Sys has been generally assumed. But the site of the town Leibethra, which was once overwhelmed by the Sys in spate (Paus. ix 30.11), is contested by Hammond, , HM i 135–6Google Scholar, who would identify it with some rather shabby Hellenistic remains lying above the Ziliana between two torrent-beds; in consequence he makes the Sys one or other of the torrents, and suggests (p. 138 n. 1), on very slight grounds, that the Ziliana was anciently called the Lapathus.

⁷⁷ Marcius' descent to the coast north of Heracleium was a terrible ordeal that lasted four days and exhausted his army, even though he was unhampered by the enemy. The area through which he passed is debatable: see Kromayer, , AS ii 281–5Google Scholar; Pritchett, , AGT ii 170Google Scholar, 174–5; Helly, , RP xlvi (1972) 276–82Google Scholar; Hammond, , HM i 138Google Scholar. In Helly's reconstruction, which seems to me the likeliest, the first two days of Marcius' ‘descent’ were in fact spent on the east side of Lower Olympus.

⁷⁸ In 1941 the Germans, by crossing Lower Olympus from Skotina to Gonnus, were able to circumvent defenders who held both Platamona-Heracleium and Tempe (Pritchett, , AJA lxv [1961] 375)Google Scholar; but the defenders were very few, a mere battalion, and the Germans a whole division, so that the episode does not show what would have happened in 480.

⁷⁹ On Brasidas' route see Edson, , CP xlii (1947) 97–8Google Scholar, and Gomme on Th. iv 78.6. Heracleium was assessed for tribute in 425 and 421; Edson affirms, Gomme doubts, that it was effectively controlled by Athens. If it was not, it must have been in the hands of Thessalians unfriendly to Brasidas.

⁸⁰ Down to the moment when the battle of Salamis and its issue took both sides by surprise, the task of the Greek fleet was merely to support the main defence by land. This needs to be said, because the importance of offensive action at sea is hugely overrated by some recent critics, especially those concerned to vindicate ‘Themistocles’ decree' as inscribed in the third century. It is obvious that no naval action was envisaged during the Thessalian expedition: Themistocles commanded the Athenian hoplite contingent, and the allied fleet, or some considerable part of it, lay beached at Halus. After the withdrawal Thermopylae, not Artemisium, became the key to the Greek defence. Herodotus' narrative leaves not the slightest doubt. The council at the Isthmus resolved to defend Thermopylae (vii 175.1); the ordering of the fleet to Artemisium was a corollary (viii 175.2), since enemy troops could otherwise be landed behind the Greek position (a threat which did not arise in Thessaly). The soldiers at Thermopylae were full of confidence (indeed they were over-confident, and subsequently sacrificed themselves out of chagrin), the sailors at Artemisium were timid and even panicky. Herodotus describes the ground at Thermopylae in great detail (viii 1762.—177, 198–201), while the site of Artemisium receives passing mention (viii 176.1). The fighting at Thermopylae is so magnified with anecdote and retrospect (vii 201–39, viii 24–5) that it only falls short of Salamis; the much more varied action at Artemisium seems brief and blurred by comparison (viii 1–21). This despite the fact that Herodotus gathered his information at Athens, which sent levies to Artemisium but not to Thermopylae. Finally, the news of Thermopylae caused the fleet to retire from Artemisium (viii 21.2) and was received with shock and alarm throughout Greece (viii 40–1). It follows that Themistocles could never have urged a naval defence as a substitute for a defence by land.

⁸¹ Although no consensus exists, understandably, about the size of the Persian army, there is a certain disposition to speak of the range 150,000–200,000, either as representing three army corps of about 60,000 each, or as the tenth part of Herodotus' total, with chiliads substituted for myriads, or as the maximum number that could be sustained by the water-supply en route. None of these conjectures has any value, and they certainly do not corroborate each other. Hignett, , XIG 355Google Scholar, reckons 80,000, including 9,000 cavalry; Larsen, , GFS 116Google Scholar, suggests a single infantry corps of 50,000 plus a few thousand cavalry. More to the point is Tarn's remark, apropos of the figure 600,000 which Arrian gave for Darius' army at Issus, ‘the greatest force raised by Antigonus when king of Asia west of Euphrates was 88,000 men, partly Europeans, and … in 302–301, when every state was making a supreme effort, Macedonia, Greece, Thrace, Egypt, and Asia west of India, with mercenaries, pirates, and Illyrians, had some 230,000–240,000 men under arms of whom probably half were Europeans' (CAH vi 367).

⁸² Edson, , CP xlii (1947) 97Google Scholar, and in Ἀρχαία Μακϵδονία (Salonica 1968), 37–38, holds that the name Heracleium betrays a Macedonian foundation and a ‘propagandist’ intent, inasmuch as Heracles sired the Argead dynasty. But the name would suit Thessalian pretensions just as well, for the magnates of Thessaly also harked back to Heracles.

⁸³ In view of this passage we may suspect that the Thessalian magnates who had intrigued with Perdiccas during the previous year (Th. iv 132.2) were partly or chiefly Aleuads of Larisa.

⁸⁴ Beloch, , GG iii 2²69Google Scholar. No one should be misled by historians of Alexander who vilify Philinna as a ‘dancer’, a ‘harlot’, and the like.

⁸⁵ Helly, B., Gonnoi (1973) i 75–6Google Scholar.

⁸⁶ When Larisa sided with Macedon against southern Thessaly, the Macedonians would be admitted through Tempe and their adversaries would have little warning. ‘Alexander's Tower’, at which Philip V encamped on the first night after Cynoscephalae, before proceeding to Gonnus the next day (Plb. xviii 27.2), is evidently a fortress or watchtower on the road south of Gonnus, and such a place would only be needed if Gonnus were not available; the builder was probably Alexander of Pherae, who early in his reign fought with Macedon while meeting opposition in Larisa.

⁸⁷ Wallace, M. B., Phoenix xxiv (1970) 199 n. 13 (on p. 200)Google Scholar.