¹ Jahreshefte des Öst. Arch. Inst. xxvii. (1932), 64–105Google Scholar.

² VI. xxiv. 2.

³ Leroux, G. (Origines de l'Édifice Hypostyle, p. 185)Google Scholar called the stoa (in its common form with interior columns) an elongated megaron with a central row of interior supports, of which one side has been replaced by a colonnade. But the essential openness of the stoa makes it different in principle from any kind of megaron. Possibly its simplest form (as in the stoa of the Athenians at Delphi) was suggested by simple lean-to shelters placed against a wall, and other forms were developed from this.

⁴ The German writers constantly use the term ‘Hufeisen’; the arrangement could be compared more aptly to goal-posts.

⁵ von Gerkan, A., Griechische Städteanlagen, p. 94Google Scholar.

⁶ V. i.

⁷ Cf. Tritsch, , Die Siadtbildungen des Altertums und die Griechische Polis. Klio, xxii. (1929), 1–83Google Scholar; see p. 76.

⁸ Cf. Demosthenes (?) xlix. 22; see Judeich, W., Topographie von Athen (1931), p. 452Google Scholar.

⁹ Von Gerkan, op. cit., p. 40. Fabricius (Pauly-Wissowa, II Reihe, Halbb. 6, pp. 1928 ff., section 11) thinks the southern part a somewhat later extension.

¹⁰ Milet, I. vi (Nordmarkt), p. 87Google Scholar.

¹¹ Ibid., I. vi, p. 89.

¹² Ibid., pp. 4, 90; for the gradual return of prosperity in the fifth century (cf. Athenian tribute lists) and the fourth, see Dunham, A. G., History of Miletus (London U.P., 1915), pp. 107, 108, 117Google Scholar.

¹³ Milet, I. vii.

¹⁴ Ibid., p. 47.

¹⁶ Op. cit., p. 100.

¹⁵ Milet, I. ii.

¹⁷ Milet, I. vi. 91–3Google Scholar.

¹⁸ Von Gerkan's restoration (Milet, I. vi. Taf. xxvii) gives a good idea of the appearance of a large part; it includes one feature which is a rather later development— the wall across the east side of the north agora.

¹⁹ Wiegand, T., Priene (agora, ch. vi, especially pp. 214–17)Google Scholar; M. Schede, Die Ruinen von Priene (agora, ch. v).

²⁰ Aristotle, (Politics, VII. xii. 2, 3, 1331a, b)Google Scholar recommends that, as in Thessaly, there should be an ἐλευθέρα agora, free or ‘pure’ of trade, and of a religious (and, it appears, rather snobbish) character; and another and separate agora τῶν ὠνἱων. This is hardly normal or natural in a Greek city—Aristotle goes to Thessaly for his example—and is not, I think, characteristic of the Ionian planned towns. There is something comparable to it in temple courts adjacent to the agora as at Priene and Magnesia, but these are not rival agoras.

²¹ Contrast the east stoa of the south agora at Miletus, with its shops and stores.

²² Humann, C., etc., Magnesia am Maeander (agora, pp. 3 ff., 22 and 107 ff.Google Scholar).

²³ Though possibly steps connected the east and south stoas even before the gate (see below) was built (see Magnesia, p. 110).

²⁴ Classical Quarterly, Jan. 1937, p. 29Google Scholar.

²⁵ The way in which the agora and its stoas are fitted into the street plan is a very important question; see von Gerkan, op. cit., p. 95.

²⁶ Athenische Mitteilungen, 1902, pp. 16 ff.Google Scholar; and 1904, p. 114. The shops on the south side belonged to a storey below agora level and faced outwards.

²⁷ Altertümer von Pergamon, III. i. 93 ff.Google Scholar

²⁸ Fyfe, T., Hellenistic Architecture, p. 170Google Scholar.

²⁹ BCH xxvi. 480 ffGoogle Scholar. and xxxv. 271–2; Délos, VII. (R. Vallois, Le Portique de Philippe).

³⁰ BCH xxix. 6 ff.Google Scholar

³¹ Délos, ii (Leroux, G., La Salle Hypostyle), p. 51Google Scholar.

³² BCH viii. 112 ffGoogle Scholar.; xxxvi. 117.

³³ Marktplatzanlagen der Griechen und Römer, Dresdener Diss., 1916. Wymer says (p. 22)Google Scholar that the forum of Caesar, in the middle of which was the temple of Venus, embodied the Ionic ideal regular enclosed place; apart from the question of enclosure, we have seen that a temple was not allowed to dominate the real Ionian agora.

³⁴ Op. cit., p. 94.

³⁵ Hesperia, VI. iii (1937). 354Google Scholar.

³⁶ Forschungen in Ephesos, iii. 89 ffGoogle Scholar. (the agora of Ephesus, of which fanciful pictures have been drawn, is only known in a late form; the Hellenistic form cannot be determined); Arch. Anz., 1938, p. 749Google Scholar, and Rev. Arch., 1938 (2), p. 228Google Scholar (Aphrodisias); van Diest, W., etc., Nysa ad Maeandrum, pp. 33 ffGoogle Scholar.; Lanckoronski, K. Graf, Städte Pamphyliens und Pisidiens (e.g., Kremna, ii. 161Google Scholar, which shows influence of forum); the small ‘agora’ of Cnidus (Antiquities of Ionia, III. ch. ii. p. 42)Google Scholar, cited by Wymer (p. 19), according to von Gerkan (p. 94) was not large enough for the purpose, and may have been the court of a gymnasium, but in position and form it has some resemblance to the ware-houses of Delos.

³⁷ Magnesia, pp. 5, 109 and 110.

³⁸ Milet, I. vii. 51 ff.Google Scholar

³⁹ Ibid., I. vii. 47.

⁴⁰ Ibid., I. vi. 94.

⁴¹ Ibid., I. v.

⁴² Gardner, E. A., etc., Excavations at Megalopolis, p. 102Google Scholar; G. C. Richards also found that the excavations confirmed in every point Curtius' old restoration, which, while placing the monuments with accuracy, seriously misrepresented the character of the agora by making the colonnades all continuous.

⁴³ AJA xxxvii. 1933, 555 ffGoogle Scholar.; the stoa was much reconstructed and complicated in Roman times. A shorter stoa on the north side of the area, immediately south of the temple hill, was built possibly at the end of the fourth century (AJA xxx. 1926, 47Google Scholar). At Orchomenos in Arcadia (BCH 1914, pp. 71 ff.Google Scholar) a long stoa, possibly of the fourth century, was placed on the north side of the agora, and at right angles to it on the east, though separate and at a lower level, was another long narrow building. The agora of Mantinea was given its fairly regular enclosed form only by the building schemes of Epigone in the first century A.D. (Fougères, G., Mantinée, pp. 179 ff.Google Scholar).

⁴⁴ Hesperia, V. i (9th report). 4–6Google Scholar.

⁴⁵ Hesperia, VI. iii (12th report). 357Google Scholar. Slight remains have also more recently been found of a late Hellenistic stoa in the eastern part of the north side of the agora, extending to within a short distance of the Attalus stoa (Hesperia, viii. 213Google Scholar).

⁴⁶ Hesperia, VI. i. 172 ffGoogle Scholar. Plan IX, facing page 360, gives a clear idea of the effect of the new buildings, if one thinks away the later Odeion.

⁴⁷ Von Gerkan, op. cit., pp. 102 and 137; two storeys are found in the lower agora, the shrine of Athena on the Acropolis, and elsewhere at Pergamon, and under Pergamene influence in the north stoa of the agora at Assos (Bacon, F. H., etc., Investigations at Assos, pp. 33 ffGoogle Scholar.).

⁴⁸ The south side of the lower agora at Pergamon may have been like this (Athenische Mitteilungen, 1902, p. 25Google Scholar); the arrangement was found in the south building of the agora at Assos.

⁴⁹ Hesperia, VI. i. 216Google Scholar.

⁵⁰ Alt-Athen, I. pp. 40–42 and 130Google Scholar, and Taf. III. Dörpfeld attributes ambitious designs to Kimon for replanning the agora. Kimon, he says, while in Ionia had seen and admired the magnificent public places of his allies there— large squares surrounded by colonnades—and came back with the idea of building a fine new agora at Athens: in attempting to carry out his plan he may possibly have had the co-operation of Hippodamus; but the scheme was never completed; along the south side was built a great stoa, the Poikile, which is to be identified with the south stoa of the Americans; ‘the stoa planned by Kimon on the north was neither then nor later carried out.’ In reality, all that Kimon is said to have been responsible for is the grave of Theseus and the planting of trees, while his brother-in-law Peisianax had the Poikile built; and, while Dörpfeld's theories contain many bewildering contradictions of the carefully considered conclusions of the excavators themselves, his identification of the Poikile is particularly arbitrary, since the south stoa is attributed to the second century, and, with open colonnades all round, had no suitable field for the great paintings. What concerns us most at the moment is that, as we have seen, there is no reason to believe that his assumed Ionian models existed at all in Kimon's time.