¹ Str. 330 fr. 22; Plut. 16.5; Zonaras ix 23. It was primarily these passages which led me in HMac. 129 f. to put the battle not by Katerini as Kromayer and Pritchett had done but in the plain farther north. I had not then visited the site of Pydna and the plain between Pydna and Korinos.

² For instance by Edson, C. F. in Hesp. xviii (1949) 84 f.CrossRefGoogle Scholar, followed by Papazoglu, F., Makedonski gradovi u rimsko doba (Skopje 1957) 101Google Scholar, and by myself in HMac. 128 f, before I visited the site.

³ My wife and I spent October 24–26, 1981, walking over the area which is bounded by Makri Yialos, Kitros, Sevasti and the coast.

⁴ This identification was suggested by Heuzey, L. and Daumet, H., Mission archéologique de Macédoine (Paris 1876) 241Google Scholar, with a confused description, but it went out of favour. Pritchett 153 ff. revived it strongly, when he and R. Stroud visited the site, and his Pls 134–6 show Heuzey's plan, a view south from the southern part of the acropolis, and a view of the elevation 56 from the south. However, on his map of the battle area (p. 157) he placed Pydna inland east of Kitros and showed nothing at elevation 36. Petsas, P. M. in The Princeton Encyclopedia of Classical Sites (Princeton 1976)Google Scholar supported Heuzey's identification; but he put the port of Pydna far away from its walled circuit at Cape Atheridha, where there is a modern artificially constructed harbour. It was essential to go over the ground again.

⁵ The natural meaning of these words would be from the beach near the Achilleion Hotel to the beach by the mouth of the Karagats river, in which I was told by a villager that large stakes have been found. The space between the Hotel and the river mouth is dominated by high cliffs. If Cassander had aimed to enclose only the acropolis, he would have made a much smaller circuit from cliff to cliff.

⁶ I was shown some from this place; they were plain, but I was told that others were inscribed with Greek letters, e.g., AN and MEPNA.

⁷ Perseus had used Dium as his base in 169; see Zonaras ix 22, reading Δίῳ for ἰδίῳ.

⁸ In 1978 I travelled by bus over this pass, starting from Katerini and ending at Elassona. See Kromayer Map 7.

⁹ The force of Scipio Nasica—or any supporting force—could have proceeded from the Petra pass to the region of Katerini and appeared in the rear of Perseus' new position, as it is sited by Kromayer and Pritchett.

¹⁰ Such sacrifice was normal after any unusual natural phenomenon, such as an eclipse, or before joining battle. Aemilius began sacrificing at dawn (Plut. 17. n). Perseus sacrificed presumably at dawn, as Alexander the Great did daily, and it seems he had ridden to Pydna for the purpose. Polybius, being Plutarch's source (Plut. 19.4), relates that Perseus rode off to the city as the battle was beginning, in order to sacrifice to Heracles; one wonders whether Polybius has moved the time of Perseus' sacrifice in order to make a contrast between Perseus being absent and Aemilius praying as he went into action (Plut. 19.6). The point, however, is that Polybius thought that Perseus was within easy riding distance of Pydna.

¹¹ The contrast is between morning (not dawn) and afternoon in ἕωθεν and περὶ δείλην, when in fact the battle did begin.

¹² This being a general rise in the Mediterranean (see HMac. 145 with n. 2).

¹³ Not, I take it, at once but in concert with Aemilius. The fleet could act as a go-between.

¹⁴ Pritchett, who, like Meloni 369, adopts Kromayer's positions, differs in that he brings Scipio through the Petra pass, that is past Petra itself, and then has him slip past Perseus' entire army to rejoin Aemilius (as if that was Scipio's task). This assumes a most improbable inactivity on Perseus' part. Livy is clearer: he makes Aemilius say that he had driven the enemy garrisons out of the pass and opened a new way into Macedonia (xliv 39.9 saltu deiectis hostium praesidiis novum iter aperui).

¹⁵ In this position Perseus' army could be by-passed by the Roman army advancing along the wide coastal plain with the Roman fleet in support, or it could be taken in the rear by a Roman force using the Petra pass, whether that of Scipio or another. See Kromayer Map 9.

¹⁶ And behind his camp in Kromayer, –Veith, , Schlachtenatlas, Röm. Abt. 10.3.Google Scholar

¹⁷ See Pritchett's photographs of the Mavroneri (pls 132, 133); and Meloni's plan on p. 394, showing its tortuous course.

¹⁸ Kromayer 315–16, Meloni 380 n. 3 and Pritchett 152 n. 33 seem unaware of the difficulty.

¹⁹ Pritchett 158 offers only one ridge beyond one wing of the position which he gives for Perseus' army. Meloni 372 translates λόϕοι συνεχεῖς as ‘una catena di colli’; he does not explain where they were.

²⁰ Walbank 384 f. has some objections to Kromayer's location of the battle but he accepts it as ‘the most likely’. He repeats an argument by Pritchett 161 that Kromayer's position for Perseus might have been on the boundary between the territories of Pydna and Dium (improbably if one looks at a map). This argument is stultified by the use of Πύδνης and not Πυδναίας (cf. Πελλαίας in Str. vii fr. 20) and by the express mention of πόλις in Polybius as cited by Plutarch and in Zonaras.

²¹ Kromayer 311 f. referring to Leake, W. M., Travels in Northern Greece (London 1835) iii 433Google Scholar; and Bursian, C. in RhMus xvi (1861) 424Google Scholar, reviewing Heuzey 156.

²² Walbank 378 makes a good assessment of the sources. The primary sources had access to the participants, and Polybius in particular knew Aemilius.

²³ The distance from the foothills to the coast might then have been some two km. The foothills themselves were suitable for the deployment of the Macedonian cavalry and the light-armed. That the camp ot Aemilius was well back from the sea is apparent from the comment in L. xliv 45.4 that after the battle Aemilius moved his camp ‘nearer to the sea’.

²⁴ διὰ μέσου δὲ ποταμοὶ ῥέοντες. We may compare X. An. i 4.4 διὰ μέσου δὲ ῥεῖ τούτων ποταμός, in between them ‘i.e. two fortifications] flows a river’, Here we have to understand ‘the two armies’. See LSJ s.v. μέσος III d.

²⁵ When Aemilius decided to move his army off the plain, he had a choice of two ridges (see MAP 2). He may well have chosen ridge 51 in order to secure this important source of water.

²⁶ Zon. ix 23 has irepl περὶ δείλην γὰρ ὁψίαν ἡ μάχη ἐγένετο.

²⁷ See Meloni 375 n. 1.

²⁸ Kromayer gave this length for Perseus' line, and again in his Schlachtenatlas, Röm. Abt. 49 ‘almost 4 km long’. However, he based it on different calculations; for he postulated in the phalanx proper a depth of 32 men on the strength of Front, ii 3.20, who calls it a double phalanx (phalangem duplicem). In my opinion this passage in Frontinus refers to Aemilius' first arrival on the scene and not to the battle itself. Plutarch describes the phalanx in some detail. He never says that it was double, and he does say that its length (τὸ μῆκος τῆς παραταξέως) contributed to its disorder in the battle. He can hardly have been thinking of a double phalanx.

²⁹ The name Λεῦκος may be explained by the whitish soil which we noticed at one point in the valley of the lower Ayios Yeoryios in walking from the ridge-point 51 to the toumba at 42. This river contained some water in 168 BC but its sluggish flow had failed to wash the bloody water clear by the next day (Plut. 21.6).

³⁰ Across the plain; not, as in Kromayer's reconstruction, across a river-bed full of water and up a slope, more rapidly than the Romans advanced.

³¹ Livy had described this retreat in a passage now lost. He referred back to it at xliv 41.9. The terrifying aspect of the advancing phalanx is described in Plb. xxix 17 and Plut. 19.2.

³² For such a disparity in casualties one should compare this battle with the battles of Marathon and Plataea and those of Alexander against the Persians. It is a mistake to dismiss the numbers as ‘incredible’ or ‘propagandist’. One should take into account the nature of the body-armour, the effectiveness of the weaponry and the advantage of men in formation over broken or leeing opposition. All these aspects are stressed by Plutarch in his account of the battle.

³³ Zon. ix 23 adds a march by Aemilius through Tempe, a very narrow pass. This should be included in the four days I have suggested.

³⁴ For the discrepancy with Plut. see Walbank 381.

³⁵ Kromayer 305 n. 1 discusses this matter; see also Walbank 383 f.

³⁶ Meloni 371 has him retire by night.

³⁷ Plutarch's remark that Scipio ‘followed safely’ the fleeing troops of Milon suggests that he did not meet any further resistance at Petra.

³⁸ Those who fail to notice the lacuna in Livy's text make Aemilius reach his goal at noon of the day on which he left his position south of the Elpeüs river. For instance, Leake supposed that Aemilius could not have gone farther than Korinos in half a day and so he located the battle near Korinos. Pritchett 160 asserts that Perseus reached his goal at midday; I know no evidence for that, and it is perhaps a slip, Perseus for Aemilius.

³⁹ So Kromayer 323 with n. 1.

⁴⁰ My positioning of the armies is such that the two back lines could be withdrawn and the camp could be fortified without the Macedonians being able to see from their battle line what was happening.

⁴¹ See Kromayer loc. cit.

⁴² Plutarch's account is best understood in the light of Frontinus. For Plutarch has those in the Roman rear ‘turn’ and make an encampment, and ‘those contiguous to the last’ (i.e. to the last of those in the rear) ‘move off in withdrawal’ (for ὑπαγωγή cf. Thuc. iii 97.3). These are the first two parts of Frontinus' triple formation, Plutarch failed to explain how the front part moved off.

⁴³ This part of Frontinus' account can hardly be applied to the battle which ensued some days later; for then it was the cavalry of the Roman right wing which distinguished itself, and there is no mention of action on the Roman left wing.

⁴⁴ Some have identified the horse of the frieze at Delphi with the unbridled horse of Plut. 18.1, and a man hurling a javelin with the heroic Paelignian commander, Salvius (20.1); see Meloni 384 n. 4 with references.

⁴⁵ This question was much debated by E. Meyer and Kromayer; see Kromayer iv 600 ff. and Schlachtenatlas, Röm. Abt. 47 f., and Meloni 380–2.

⁴⁶ Livy called these forces on both sides praesidia and one Roman force specifically pro castris stativum praesidium, that is a standing outpost in front of the camp.

⁴⁷ Although ϕησί is in a subordinate clause (Plut. 18.5), it indicates that Plutarch was drawing on Scipio or this account.

⁴⁸ This colourful word is probably in the style of Scipio's writing.

⁴⁹ There was at least one other agema among the phalangite forces, as the plural is used at Plut. 19.1, τοῖς ἀγήμασι.

⁵⁰ L. xlii 51.3 pars ferine dimidia.

⁵¹ See Zon. ix 22, Polyaenus iv 21 and L. xliv 41.4.

⁵² Kromayer 343 f. and his conclusion at 348, and Meloni 375 n. 1. Pritchett 158 is more positive: ‘the Roman force numbered 38,000’. If the Macedonians were markedly superior in light-armed troops, we can understand how they hoped to make use of the ‘continuous ridges’ for their manoeuvres (Plut. 16.8).

⁵³ Front, ii 3.20 had Aemilius make a feigned retreat in order to dra w the Macedonian phalanx on to broken ground (confragosa loca); but that was at the time of Aemilius' arrival.

⁵⁴ For cavalrymen being prominent on the frieze see Meloni 393. Th e representation of horsemen is of course traditional in this type of art.

⁵⁵ Meloni 390, for instance, supposed that Aemilius, having got his army out in battle line and then seen the defeat of his right wing, managed to move the troops of his left wing—namely the Numidian elephants and the allied infantry, 10,000 or so in number—to the right wing by marching them along behind the battle line, When this operation was complete, he had them line up and defeat the successful Macedonian left wing, which had apparently failed to follow up its own victory. The march alone over 4 km of broken ground with elephants and so large a force of men would have taken an hour. Meanwhile it appears that Perseus failed to get his cavalry even into position, let alone exploit the total lack of opposition to his own wing! Time and space, let alone probability, do not allow for such suppositions, which rest in this case on nothing more than L. xliv 41.3, where Aemilius ‘brings the elephants and the allies' squadrons onto the right wing where the battle had been joined by the river’ (alias being emended to alas).

⁵⁶ One feels that a more capable king would have led his phalanx, as Philip II did against Bardylis and at Chaeronea, and as Alexander III did at Issus in my opinion (see Hammond, , Alexander the Great: King, Commander and Statesman [London 1981] 104)Google Scholar. If Perseus planned to make a devastating charge with his excellent cavalry, his plan did not materialise. Some of the questions about Perseus and the Macedonian tactics are considered by Gyioka, P. K., Περσεύς: ὁ τελευταῖος βασιλεὺς τῶν Μακεδόνων (Thessaloniki 1978) 397 f.Google Scholar; he accepts Kromayer's location for the battle.

⁵⁷ One man in Pydna hid his treasure successfully; for a hoard of Perseus' silver coins was found in excavating the foundations of a house outside the acropolis but inside what I have suggested was the walled area of the city. The find was made shortly before I arrived, and the coins had been dispersed.