¹ Herodotus, vi. 108.

² The Topography of Athens (2nd ed. 1841), vol. ix. pp. 89–92, 211.

³ Athenische Mitteilungen, vol. i. pp. 71–6, 88–9.

⁴ Similarly modern Capua represents, not ancient Capua, but Casilinum.

⁵ Pliny, , Hist. Nat. iv. 7 (11)Google Scholar, describes Marathon as a mere speck (‘locus’).

⁶ Ross, , Erinnerungen, p. 186Google Scholar.

⁷ Iwan-Müller's, Handbuch der klassischen Altertumswissenschaft, vol. iii. (1889), p. 119Google Scholar.

⁸ Karten von Attika: Erläuternder, Text. Heft iii–iv. pp. 40Google Scholar, 52.

⁹ Op. cit. p. 52.

¹⁰ Two leading German critics, Delbrück and Ed. Meyer, have pronounced this argument to be fatal to Lolling's view.

¹¹ Loc. cit.

¹² Lieut. Rhediades, (Ἀρχαιολογικὴ Ἐφημϵρίς, 1906, pp. 239–44)Google Scholar locates the Heracleum whence Xerxes watched the battle of Salamis hard by a church of St. George. Other such transferences of Heracles' sanctuaries to St. George might be discovered by further investigations on the spot.

¹³ Miltiades, ch. 5: regione non apertissima proelium commiserunt: namque arbores multis locis erant rarae (so the archetype: not ‘stratae’).

A telling illustration of this important feature of the battle-site will be found in Grundy, 's Great Persian War, p. 187Google Scholar.

¹⁴ The difficulty of manœuvring hoplites on such obstructed ground led Milchhöfer to transfer the entire battle to quite another site (loc. cit. pp. 51–3). But the solution of the problem as given above is more in accordance with the data of ancient texts.

¹⁵ Delbrück, Ed. Meyer and Busolt follow Nepos' account in this matter.

¹⁶ The Persians were at perfect liberty to shift their position, and a more advantageous fighting ground would have been worth more to them than the eventual accession of a Spartan reinforcement would be to the Athenians.

The absence of the Persian horse is attested by the tradition embodied in Suidas' gloss on the proverb πλημοχόη the historic value of which has been successfully upheld by Milchhöfer and Macan.

Those who assume the presence of cavalry at the scene of action are quite at a loss to explain why it failed to assist the hard-pressed Persian infantry or how it managed to re-embark in face of the Athenian pursuit. The attempt of Delbrück, (Gesch. d. Kriegskunst im Altertum, i. p. 59–62)Google Scholar to explain away these difficulties verges on the comic.

¹⁷ Cf. Plutarch, , Aristides 5Google Scholar: ἤνυσαν αὐθήμϵρον. Lysias, , Epitaphius 76Google Scholar, and Isocrates, , Panegyricus 86Google Scholarsq. preserve the same tradition in an inverted form when they make the Athenians march out and defeat the Persians on the same day. From these passages it may be inferred that the extreme rapidity of the Athenians' march was one of the most persistent elements in the current story of the battle.

¹⁸ i. 15. 3.

¹⁹ Hist. Nat. xxxv. 57.

²⁰ De Natura Animalium, vii. 38.

²¹ Herodotus iv.-–vi., vol. ii. pp. 228–230.

²² Prof. E. A. Gardner has kindly brought to my notice the parallel case of the statue of Nemesis at Rhamnus, which is variously ascribed to Alcamenes and Agoracritus, but like many other such unattributed sculptures can be referred with absolute certainty to the school and age of Pheidias.

²³ Pausanias, i. 27. 1.

²⁴ Pausanias in this context ignores the portrait of the poet Aeschylus contained in the same picture (Harrison, and Verrall, , Mythology and Monuments of Ancient Athens, p. 137Google Scholar Pausanias, i. 21. 2).

²⁵ The arrangement of the picture in the Stoa may readily be inferred from the vase paintings of the so-called ‘fine r.-f. style,’ or from the reconstruction of Polygnotus' work in Robert, Die Nekuia des Polygnot.

²⁶ Griechische Geschichte, iii. p. 323, n. 3.

²⁷ Klio, 1908, pp. 190–1.

²⁸ Hdt. viii. 24–5.

²⁹ This route was subsequently used by a Persian column (Hdt. viii. 31). Xerxes' failure to use it sooner can hardly be explained save on the above hypothesis. Munro's suggestion (J.H.S. 1902, p. 313) that a garrison of Locrians at Heracleia at first barred the Persian advance is invalidated by the same author, who shows that the position at Heracleia could be circumvented.

³⁰ Klio, 1908, pp. 477–486.

³¹ Berl. Phil. Wochenschrift, 1909, pp. 60–3.

³² Themistocles, ch. 14.

³³ viii. 96.

³⁴ The Mediterranean Pilot, vol. iv. p. 59.

³⁵ The swell even on ordinary days is sufficient to raise crests on the waves.

³⁶ Persae, ll. 441–4.

³⁷ viii. 76.

³⁸ Persae, ll. 382–3: σμηγματοθήκη σμῆγμα ἀγγεῖον μέλαν μετὰ πλατείας βάσεως ἔχον τὰ Of all modern critics Prof.Goodwin, (Harvard Studies in Classical Philology, xvii. (pp. 81–2)Google Scholar alone seems to have firmly grasped the fact that the Persian squadrons were kept moving through the night. This fact is important as showing that the Persian fleet after sailing up-channel in the afternoon (Hdt. viii. 70) issued back into the open water and did not re-enter the straits till the morning of the battle. It also helps to account for the defeat of the Persians: the exertion and mental strain of this cruise must have been quite exhausting.

³⁹ Herodotus, who is our best witness on this matter (viii. 97), represents this work as mainly if not wholly a bridge of cargo boats. The ‘mole’ version is a later improvement.

⁴⁰ Ctesias places the work at the narrowest point of the straits. This is probably a guess, for no traces of the building seem to have remained even to Herodotus' time.

⁴¹ Persica (ed. Gilmore), § 57.

⁴² See especially the damaging criticisms by Wright, H. B., The Battle of Plataea, p. 55Google Scholar.

⁴³ Meyer, Ed. found the Asopus waterless in early June (Gesch. des Altertums iii. p. 409Google Scholar, n.).

⁴⁴ The rise from the Asopus level amounts to about 200 ft. in a mile.