Acknowledgements. I offer my best thanks to Mr. J. Boardman, who as my supervisor for a thesis on Early Corinth has been generous with his time and knowledge ever since I began work on this subject; he and Professor G. L. Huxley both read an earlier draft of this article and suggested many improvements. Professor M. Robertson made many helpful comments on the chapter of my thesis from which this article grew; Dr. L. H. Jeffery and Mr. V. R. d'A. Desborough examined my thesis and gave me the benefit of their criticism; and Professor A. E. Astin has sustained me with his advice and encouragement. I am grateful to them all, and it would be a poor return for the help I have received not to emphasize my own responsibility for all remaining errors.

In Greece, Mr. P. Kalligas at Corfu, Mrs. E. Touloupa at the Athens National Museum, and Mr. C. K. Williams, Dr. J. K. Perlzweig, and Miss K. Butt in Corinth all provided me with facilities for study and gave freely of their time to answer my questions. My thanks are due to Mr. A. H. S. Megaw and the authorities of the British School at Athens for arranging permits for my visits to Greek museums at an awkward time; and for permission to study finds in Greece I must record a special debt to the Greek Archaeological Service in general, and to Miss O. Alexandri, Mr. B. Kallipolitis and Mrs. E. Protonotariou-Deilaki in particular. For permission to mention unpublished vases I am much indebted to Professor P. Amandry, Dr. J. L. Benson, Dr. Keith DeVries, and Professor L. Lerat.

My visit to Greece in 1968 was financed by the Research Fund of The Queen's University of Belfast; my gratitude is due to the administrators of the Fund for their generosity.

Abbreviations. Articles and Books:

AAA = Athens Annals of Archaeology.

Coldstream = Coldstream, Greek Geometric Pottery.

Courbin, CGA = Courbin, La Céramique géométrique de l'Argolide.

Dunbabin = Dunbabin, The Oracle of Hera Akraia at Perachora, BSA xlvi (1951) 61 ff.

Hammond = Hammond, The Heraeum at Perachora and Corinthian Encroachment, BSA xlix (1954) 93 ff.

Hanell = Hanell, Megarische Studien.

VS = Johansen, Les Vases sicyoniens.

Weinberg = Weinberg, Corinth vii pt. I, The Geometric and Orientalising Pottery.

Vases, etc.

Objects found at Perachora are referred to in the following way: those published in Perachora i have first the plate number and then the number on the plate, e.g. i.12.I (= Perachora i, pi. 12, vase I); those published in Perachora ii have merely the serial number, e.g. ii.205.

The following three abbreviations are also used regularly; in each case the letter precedes the catalogue number of the publication concerned:

Aetos B = Benton, BSA xlviii (1953) 255 ff.

Aetos R = Robertson, BSA xliii (1948) 9 ff.

W = Weinberg, Corinth vii pt. I.

¹ This point, however, is disputed; see below.

² It is a matter for discussion whether Limenia was a full cult title for Hera at this upper sanctuary and the worship of the goddess here is therefore in this sense distinct; or whether Limenia is merely a descriptive title and the worship of the goddess here is an adjunct to the main cult down by the harbour. I argue for the latter view below, but use the normal name for the upper sanctuary throughout, for ease of reference.

³ Fifth century, i.27.2: . Fourth century, i.29.1–2 (also i.131.1–2): .

⁴ On Acraea as a cult title, see below, p. 195.

⁵ The circumstances of the find are reported in Perachora i. 27 ff.

⁶ Coldstream's abbreviations for the various phases of Corinthian Geometric will be used throughout; see Coldstream 91 ff. On the absolute chronology of the Geometric period, see Coldstream 302 ff.

⁷ The closest is Corinth xiii. 16–10Google Scholar (fig. 2) from grave 16, one of the earliest graves in the North Cemetery (Corinth xiii. 15Google Scholar); but the Perachora vase is noticeably shallower.

⁸ Cf. Coldstream 97. Two good examples of the deep, late MG II skyphoi were found in the North Cemetery: Corinth xiii. 18–7Google Scholar; 20–2 (fig. 3). Young has claimed that the decoration, as well as the shape, of i.12.1 places it earlier than Corinth xiii. 16–10Google Scholar, on the ground that the vase from the North Cemetery has uprights bounding the chevron zone while that from Perachora does not (Corinth xiii. 41Google Scholar); but Ridgway has shown that there is little chronological significance in this criterion (Studi Etruschi xxxv (1967) 313 ff.Google Scholar; cf. Coldstream 96).

⁹ They are very similar to W 21 (LPG) and to two unpublished cups from the Anaploga well at Corinth, which is to be dated to MG I (Corinth C–63–617, 645; for a preliminary report on the deposit, cf. ADelt xix (1964) B 101Google Scholar). I am very grateful to Dr. Keith DeVries for permission to mention these pieces and for information about the date of the deposit in which they were found. Dunbabin (JHS lxviii (1948) 64Google Scholar) dates the Perachora cups c. 850, and Coldstream (p. 102) believes that the shape was absent from the Corinthian repertory from EG to MG II; but the cups from the Anaploga well now show that the shape was current in MG I, and we do not have enough Corinthian MG II material to be sure that such vases were not made during that phase.

¹⁰ For this term, cf. Coldstream 98.

¹¹ Corinth xiii. 14–2Google Scholar (FIG. 4); 15–2. These vases were found in graves belonging to Group A in the North Cemetery. Group B, which is to be dated later than Group A (Corinth xiii. 15Google Scholar), included Grave 17, in which was found a similar protokotyle Corinth xiii. 17–4Google Scholar (FIG. 4); but this example is deeper, has a finer fabric, and has smaller handles.

¹² No protokotylai decorated in this way have been found in dated grave-groups; but is is reasonable to suppose that the shape of the elaborately decorated protokotylai followed the same development as that of the less ambitious, darkglaze examples.

¹³ Cf. Coldstream 95, 97; Corinth xiii. 42Google Scholar.

¹⁴ Coldstream 96. The only published example with glaze at the bottom which belongs to the MG II period is W 76; but an unpublished piece (Corinth KP 166) from Grave I in the Potters' Quarter (for the grave, cf. Corinth xv pt. 1, 7) is similar. I am grateful to Dr. J. L. Benson for permission to mention this vase.

¹⁵ The body shape is still similar to that of the examples with glaze on the body quoted above, n. 14. Later pieces with a flatter shape, lines on the whole body, and glaze at the lip are Corinth xiii. 17–2Google Scholar and Goldstream pl. 18c (Corinth KP 164); Corinth xiii. 20–1Google Scholar is even flatter, and the lines on the neck reach right up to the lip.

¹⁶ It is to be noted that there was little, if any, stratification in the Geometric Deposit (Perachora i. 31 fGoogle Scholar). That may mean that all the votives from the temple were thrown out together into the area south of the building when it went out of use, and that the period of use of the temple is therefore the same as the period between the first and the last votives; but Payne noted that the nature of the terrain may have been responsible for the lack of proper stratification (Perachora i. 31Google Scholar).

¹⁷ Perachora i. 92 ff.Google Scholar; Coulton, , BSA lxii (1967) 365Google Scholar.

¹⁸ Perachora i. 93Google Scholar; 98 ff.; see further Coulton, loc. cit.

¹⁹ The fragment is not complete enough to make this identification certain. But it is likely that the row of uprights, to the right of the zone with wavy lines, represents the legs of a row of birds, since the tail of the bird at the extreme left seems to appear to the left of the last upright. Straight legs seem to be a feature of early examples of such bird-files (below, p. 172); an early date is also indicated by the wavy lines on the left of the fragment (Coldstream 105 for this criterion).

²⁰ On i.12.2 see below, p. 169.

²¹ This fragment is misnumbered in the text of Perachora i as i.26.12 (Perachora i. 95Google Scholar). It is certainly earlier than the kotylai with meanders from the Limenia Deposit, ii.468–9; its meander is much heavier. It is not certain that the drawing to the left is a bird; it could be a badly painted dot rosette.

²² Cf. Aetos R 161, B 944 (both with rays). The same decoration also occurs in LG (Aetos B 798–9; Corinth xiii. S–5Google Scholar); but the shape of i.25.9 and its rays make it clear that it must be EPC.

²³ Payne, (Perachora i. 93 ff.Google Scholar) suggests that i.25.6, 9 and i.26.19, 20 (misnumbered as i.26.18, 19) are eighth-century.

²⁴ There is one piece from the Geometric Deposit which seems to date from a much later period than the vast majority of the finds; the conical oinochoe(?) handle i.13.16, although not closely identifiable, is not likely to be earlier than the end of the seventh century. At the western end of the deposit there were a few seventh-century sherds; but all these fragments were originally part of a later deposit which has become mixed with the Geometric in a very small area (Perachora i. 31Google Scholar). See also below, p. 172 (i.122.6).

²⁵ i.23.10–12 were found as good as whole; they date to MPC–LPC (Perachora i. 94Google Scholar), and are therefore of relevant date, since Dunbabin supposed that there was no temple in the harbour area in the seventh century.

²⁶ See above, p. 163, with n. 18.

²⁷ We should, however, expect the number of dedications to fall in the classical period, since that was general in all Greek sanctuaries; cf. Perachora i. 93Google Scholar. In this connection it ma y be noted that if a new temple was built by the harbour in the late sixth century where none had existed since the late eighth, we should have expected the number of dedications at the Limenia Deposit to decrease because of the competition; but no such decrease is apparent.

²⁸ Below, pp. 175 ff., especially pp. 177 f.

²⁹ Cf. most recently Plommer, and Salviat, , BSA lxi (1966) 207 ff.Google Scholar, modifying Perachora i. 89 fGoogle Scholar.

³⁰ Strabo 380: .

³¹ Latte, , RE xviii (1942) 837Google Scholar; Dunbabin 70 n. 70.

³² Revue de l'histoire des religions cxliii (1953) 145 ffGoogle Scholar. Will brings a number of objections against Dunbabin's view; but he does not dismiss it entirely, and shows that his own solution is compatible with that of Dunbabin (op. cit. 164 f.).

³³ Will, op. cit. 146.

³⁴ Op. cit. 146 f.

³⁵ I am grateful to Mr. Boardman for pointing this out to me. If this had happened, presumably the priest would have been able to control the answer; but that was regularly so with an oracle.

³⁶ Perachora i. 78Google Scholar; Dunbabin 62.

³⁷ ‘Cults of Corinth’ (unpublished Ph.D. thesis, Johns Hopkins University, Baltimore, 1955) 10 fGoogle Scholar. I am grateful to Dr. J. Perlzweig for drawing my attention to this work and for allowing me to read the copy in the library of the Amerian School at Corinth.

³⁸ Payne, , Perachora i. 80 f.Google Scholar, suggests that they may have been used to secure a beam which supported a screen for the cult statue. If that is so, the stones that covered the holes when they were found must have been added after the beam was removed; that is perfectly possible, since the stones were irregularly fitted.

³⁹ Perachora i. 121 (Payne)Google Scholar; op. cit. 152 f. (Dunbabin). Many clay phialai mesomphaloi were found in the Archaic sanctuary at Solygeia in the Corinthia; cf. Verdelis, , PAE 1958Google Scholar, pls. 112γ, 113α; Archaeology xv (1962) 191, figGoogle Scholar. at bottom. The majority of the phialai were miniatures, but there were a few full-sized examples. Verdelis suggested that the sanctuary at Solygeia was sacred to Hera, probably under the title Acraea (Archaeobgy xv (1962) 192Google Scholar); that might provide a parallel for the use of phialai at Perachora, but there is no way of demonstrating that the Solygeia sanctuary was a Heraeum.

⁴⁰ It might be suggested that the models were dedicated by the architects or builders of the temple when their work was completed; but this is unlikely, since the models seem to cover a considerable period of time (below, p. 187).

⁴¹ At Delphi there is no clear evidence for the existence of a cult until well into the eighth century (Amandry, , La Mantique apollonienne à Delphes 201 ff.Google Scholar); the evidence for the oracle is only a little later—the series of replies given to prospective colonists (Forrest, , Historia vi (1957) 164 f.Google Scholar; Parke, and Wormell, , The Delphic Oracle i. 49 ff.Google Scholar).

⁴² Perachora ii. 395, no. 17Google Scholar: .

₄₃ i.43·5–7: cf. Perachora i. 136Google Scholar; Perachora ii. 398, no. 101Google Scholar.

⁴⁴ Aetos B 630–3 all have nicks in the rim, but they are decorated with herons facing over water (630) or with herons in metopes separated by a panel of horizontal wavy lines (631–3), which is a still later stage in the development. For the series of LG kotyle decoration, cf. Coldstream 101.

⁴⁵ I have placed W 75 in mid MG II partly on grounds of shape: it is much less shallow than i.12.1 and Corinth xiii. 16–10Google Scholar, which both come early in the phase (see above, p. 161, for this chronological criterion); and partly on grounds of its context: the grave in which it was found (Morgan, , AJA xli (1937) 543Google Scholar; Weinberg 25) contained a conical oinochoe (W 76) which comes early in the MG II series (above, p. 162 with n. 14) and a pedestalled krater which is appreciably later than the beginning of MG II (cf. Corinth xiii. 43Google Scholar).

⁴⁶ Other kotylai outside the series were found; but they cannot be used for direct comparison, since no counterparts were found in the Limenia Deposit. i.11.8 should be early; it has the same decoration (upright zigzags enclosed by uprights in handle panel) as W 107, which has a very shallow shape (section: Weinberg 36, fig. 11). i.11.9, decorated with glaze and added white lines, is probably also early, with its comparatively heavy fabric and its shallow shape (section: i.121.12). i.13.13, which is glazed and has two added white circles between the handles, has few parallels; but the closest is still MG II (Coldstream pl. 18f), and therefore the Perachora vase should be no later than early LG. Coldstream appears to place it still within MG II, despite making all kotylai LG (97 n. 1; 101).

⁴⁷ There may have been a heron between the wavy lines to the right and the single wavy line in the centre; if there was it will have been balanced by a second to the left. The decoration cannot have been merely a single panel of wavy lines (which would be without parallel); the one in the centre differs from those on the right, and would not join them if produced in their direction. Th e scheme may have been similar to that on the krater Corinth xiii. S–3Google Scholar, though without the hour-glass pattern.

⁴⁸ There appear to be eight; this is many fewer than on the kotylai with the chevron scheme from both the Acraea and the Limenia Deposits (see above; cf. also Aetos B 630, Coldstream pl. 19k). This fact would not be important if it were the only early feature, since if the panel decoration were like that of Corinth xiii. S–3Google Scholar (above, n. 47) there would have been little room at the ends for uprights; but the handle is also early (below).

⁴⁹ Ladder handles do not normally appear on kotylai; cf. Coldstream 101; Benton, , BSA xlviii (1953) 279Google Scholar. But they are found on MG II skyphoi (W 75; KP 170, unpublished, from Grave V in the Potters' Quarter: for the grave, cf. Corinth xv pt. 1, 8 f.; my thanks are due to Dr. J. L. Benson for his permission to mention this vase). I have not noted any LG skyphos or kotyle which has the same handle decoration.

⁵⁰ There are certainly no direct comparisons that can be made between i.12.2 and any vases from the Limenia Deposit; but that cannot be used to argue that this kind of kotyle ceased to be mad e before the Limenia Deposit was opened, since no similar vases have been found elsewhere either.

⁵¹ For the term, cf. Coldstream 99; the class, op. cit. 102 ff.

⁵² Geometric Deposit: perhaps i.122.1–2. Limenia Deposit: ii.682–6, 695 (some without panels).

⁵³ The Limenia kyathoi are ii.629–39 (a selection, FIG. 8). Four of these fragments (629, 632, 635, 637) might have had more than one bird; they are the only examples in which the bird does not certainly face the edge of its panel. Dunbabin suggests that ii.634 may have belonged to the same kyathos as ii.629 (Perachora ii. 73Google Scholar); if so the vase had a single bird, since on ii.634 the creature faces the edge of its panel.

⁵⁴ References in Perachora ii. 73Google Scholar.

⁵⁵ Cf. Coldstream pl. 19d. MG II vases with an undercut base, however, usually have a ring foot, which is not present on ii.205.

⁵⁶ It is similar to i.14.1 from the Geometric Deposit, for whose date in mid MG II see above, p. 162. The Limenia vase, however, has a rather wider neck.

⁵⁷ They are in a quite different position from the dots that do appear in MG II; cf. the krater W 73, with dots just below the rim.

⁵⁸ The scheme is only suitable for the decoration of a shoulder, and there are therefore few parallels. Cf. however Aetos B 908 (floating three-barred sigmas in a metope), 909 (the same in one metope, and diagonal wavy line in another), 910 (metopes, but the motifs that float in them not clear). Only 910 is preserved whole; its shape is not distinctive, and it cannot be dated.

⁵⁹ ii.205 came from the Egyptia n Pit (Perachora ii. xvii, 35Google Scholar), whose closely datable objects include nothing from the eighth century; it is therefore unlikely that the vase was made much before c. 700.

⁶⁰ i.14.2 is only a neck, but it presents some difficulty. It has very heavy fabric, and the curious incised eye on the lip seems equally to be comparatively early; but such elaborate decoration (‘N’ pattern in the upper frieze and dicing in the lower) does not normally occur on the neck of the conical oinochoe until the seventh century. Our neck, however, is from a vase of larger size than usual, and the two motifs do appear on a pomegranate vase in LG (VS pl. 8, 1); i.14.2 should probably be of that period, but it cannot be used for comparison, since it stands outside the regular series.

⁶¹ For the date of this krater early in LG, cf. Corinth xiii. 44Google Scholar.

⁶² The series has been fixed by Callipolitis–Feytmans, : BCH lxxxvi (1962) 117 ffCrossRefGoogle Scholar. The order in which Dunbabin placed the examples from the Limenia Deposit must now be slightly altered.

⁶³ ii.731 has a ring foot, while i.123.16 does not; the ring foot is likely to be a later feature.

⁶⁴ The groups of bars on the rim do not often appear after early LG at the latest on drinking-vessels, but they seem to have survived into EPC on other shapes (cf. W 116: krater); they probably lasted for some time on plates, since the rim is a conspicuous part of such a shape. The groups of diagonally disposed chevrons appear on the shoulders of two closed vases in Ithaca (Aetos B 864, 864a); 864 seems to be early in LG (lines covering much of the body, but a frieze of chevrons below the shoulder), but it is perhaps hazardous to compare two such different shapes. Most of the Limenia plates show later features—snakes (ii.726; cf. below), hooks (ii.727) and hatched triangles with rays (ii.728); only ii.729–30 (groups of sigmas, zigzags) appear to be as thoroughly Geometric as i.B.1 (cf. Dunbabin in Perachora ii. 85Google Scholar).

⁶⁵ From the photograph the plate appears to have a convex section; but this is far from certain.

⁶⁶ ii.208, 212: both are conical oinochoai.

⁶⁷ Perachora ii. 104Google Scholar, fig. 7 for the section.

⁶⁸ Coldstream, 107, places the invention of this shape early in EPC; but the decoration of this vase cannot be so late.

⁶⁹ The herons on kraters perhaps remained neater for longer than on other shapes, since they were larger. A krater similar to ii.1275 was found in the North Cemetery at Corinth (Corinth xiii. S–3Google Scholar); it too has a tall shape, and should be late LG.

⁷⁰ It did not have the hour-glass pattern, so it may be earlier than ii.1275 (Perachora ii. 124Google Scholar); on kotylai the hour-glass pattern is not found with the earliest herons (Coldstream 101).

⁷¹ Above, p. 169; the vases are ii.1277, 1279. ii.1006 also belongs to the Thapsos Class; it is probably from a pyxis (Perachora ii. 109Google Scholar), but might be from a krater.

⁷² But see above, n. 68.

⁷³ Geometric Deposit, early LG: i.13.22 (kyathos), i.11.7 (conical oinochoe). Geometric Deposit, middle LG: i.122.9 (kotyle-pyxis), i.122.13–14 (snake pattern). Limenia Deposit, middle LG: ii.819 (bird; uncertain shape), perhaps ii.906 (kotyle-pyxis), perhaps ii.1046 (globular pyxis). Limenia Deposit, late LG: ii.629–39 (kyathoi; perhaps a few middle LG), ii.940 (flat pyxis), ii.1275, perhaps ii.1276 (kraters; the latter perhaps middle LG). In order to complete the list of possible middle-LG vases from the Limenia deposit, mention should be made of the scraps ii.145. 1074–5, 1299. None of them can be dated with certainty, since they are so small; but they could all belong among the earliest dedications in the Limenia sanctuary, although they might equally well be later.

⁷⁴ Geometric Deposit, early to middle LG: i.123.16, and probably i.14.7. Geometric Deposit, middle LG: i.B.i. Limenia Deposit, middle LG: ii.731. Limenia Deposit, late LG: ii.726–30 (above, n. 64; 729–30 perhaps middle LG).

⁷⁵ Vallet, and Villard, , Mégara Hyblaea ii. pls. 2.6Google Scholar; 3.9.

⁷⁶ The dating of Thapsos Class skyphoi from Megara Hyblaea and nearby Thapsos is of course still a matter of controversy. I hope to discuss the class as a whole elsewhere; meanwhile, see Coldstream 324 f. for a brief refutation of the view of Vallet, and Villard, (BCH lxxvi (1952) 289 f.Google Scholar) that some of the Thapsos Class skyphoi from these sites are to be dated early in Corinthian LG. As for other Corinthian LG fragments from Megar a Hyblaea, the heron on the globular pyxis Vallet, and Villard, , Mégara Hyblaea ii pl. 3.10Google Scholar is certainly middle LG, if not late: it is very schematic. There is another heron that is rather like that on i.122.9 (Vallet and Villard, op. cit. pl. 3.11). Cf. in general Coldstream 324 f.

⁷⁷ The white-on-black oinochoe i.122.6 from the Geometric Deposit dates to the early years of EPC; but it is likely that it was one of the few fragments that strayed into the deposit at a later date, since it comes some time after the great majority of pieces from its context (cf. above, n. 24).

⁷⁸ The building was 9·5 m. × 5·6 m. (Perachora i. 110Google Scholar).

⁷⁹ Nilsson, , Archaic Greek Temples with Fire-Places in their Interior, in Från stenålder till rokoko: Studier tillägnade Otto Rydbeck den 25 augusti 1937 (Lund, 1937) 43 ffGoogle Scholar.

⁸⁰ In the original publication of the hearth and kerbstones, three of which were inscribed (Perachora i. 111 f., 256 ff.Google Scholar), it was argued that the hearth must have been made, at the latest, towards the middle of the seventh century; but this view has been rightly rejected by Jeffery, , Local Scripts of Archaic Greece 122 fGoogle Scholar. Archaeologically it is quite possible that the hearth does not even belong to the Archaic period; the few Archaic sherds that were found there could easily have strayed into the pit long after they were made (Jeffery, loc. cit.). Similar hearths were made in the Classical period: references in Jeffery, op. cit. 122 n. 5.

⁸¹ There is evidence for cult celebrations somewhere in the temenos of Hera Limenia from the earliest years of the building. The pool just outside the temenos in which the phialai were found was not a natural feature, but was deliberately excavated during the early years of the Limenia Deposit (Perachora i. 120Google Scholar). The phialai indicate very plainly that the pool had some ritual significance—probably in connection with ablutions or libations (see above). Some rites were therefore performed in the temenos from the earliest times; they are likely to have been carried out in the building that certainly existed at that early date.

⁸² Two smal l bases were found in the south-west corner of the building (Perachora i. 112Google Scholar); but they are much too small to have served this purpose, and they are in any case not central in the building, as a cul t statue would have been. For a suggestion that they supported objects used in the cult (perhaps spits), cf. Perachora i. 187Google Scholar. Payne, (Perachora i. 133 f.Google Scholar) suggests that a bronze dove found in the Limenia Deposi t (i.40.3–4; 41) was originally held in the hand of a cult statue of Hera Limenia, on the ground that it is irregularly made; the irregularities would fit well with this view, as would the holes for attachment at the sides of the bird, but there is nothing to show that the dove did not come from a cult statue of Hera Acraea down by the harbour rather than from one of Hera Limenia. A discarded statue from the Acraea temple could well have been dedicated at the Limenia Deposit; it is argued below that the two sanctuaries belonged to exactly the same deity.

⁸³ Perachora i. 116Google Scholar; the temenos had gone out of use by the time the next building (a Roman house) was constructed in it.

⁸⁴ Hammond (p. 98) appears to follow Dunbabin's view that the late-sixth-century temple of Hera Acraea was the first successor of the Geometric building; but there are difficulties for his view even if we adopt this improbable solution to the problems of the building history of the Acraea sanctuary (for the two alternatives, see above, pp. 164 f.). It is just possible that there was no major sanctuary in the harbour area until the building of the late-sixth-century temple to Hera Acraea; but in that case it is difficult to see why the Corinthians should have built a temple then for the goddess whose cult they had suppressed some two centuries earlier.

⁸⁵ I hope to discuss the question of early Corinthian contacts in the west elsewhere; meanwhile, see Coldstream 353 f.

⁸⁶ The Limenia Deposit begins in middle LG; that is also the archaeological date for the foundation of the first Sicilian colonies (cf. most recently Coldstream 322 ff., especially 324–5). The date for the foundation of Corcyra is not yet archaeologically fixed (Coldstream 367; but cf. also 370); but Corinthian interests in the west will have been quickened by the foundation of Syracuse alone, so the doubt over the date of Corcyra matters little.

⁸⁷ See LS⁹ s.v.

⁸⁸ Cf. Tomlinson, , BSA lxiv (1969) 235Google Scholar.

⁸⁹ Strictly, there is no firm information about the temple, since the conclusion is based entirely on the votive deposit.

⁹⁰ But see below, n. 95. On Hera Epilimenia at Thasos, Bon and Seyrig, , BCH liii (1929) 345 ffGoogle Scholar.

⁹¹ Dunbabin 62 with n. 8. He used the point to argue that the Limenia sanctuary could have been the only one dedicated to Hera Acraea on the site between the disuse of the Geometric temple and the building of the late-sixth-century temple to Hera Acraea by the harbour. This is not a likely view (see above), but his more limited conclusions on the nature of the title Limenia remain unaffected. The point is not that Limenia could not have been a cult title, since plainly it could; merely that it might not have been.

⁹² A temple was of course built in the area in the late sixth century; but a good deal of quarrying had to be done in order to provide a suitable site for it, and even then the ground was not entirely satisfactory (Perachora i. 78Google Scholar; Dunbabin 62).

⁹³ A parallel may be found in the cult of Artemis at Sparta; cf. Dunbabin 63 n. 16 (with references). The goddess was probably known there both as Orthia and as Limnaea or Limnatis; the former was the cult name—indeed, Orthia seems originally to have been a distinct goddess (cf. Rose, in Artemis Orthia 401 f.Google Scholar)—and the latter alternatives referred to the location of her sanctuary in a marsh. The case is not precisely the same as that at the Heraeum, since there was only one sanctuary in the Artemis cult; but the definition of a goddess by reference to the place where her sanctuary was to be found is exactly parallel.

⁹⁴ Perachora ii. 397, no. 57Google Scholar: . Dr. Jeffery, however, now suggests to me that is a possible, though less likely, restoration.

⁹⁵ I am most grateful to Professor Robertson for drawing my attention to a passage in Pausanias which conceivably provides a parallel for the situation at the Heraeum. At ii. 34. 11 that author reports a temple of Aphrodite at Hermione at which the goddess seems to have been known both as Pontia and as Limenia. Pontia was a regular cult title for Aphrodite (cf. Kruse, , RE xxii (1953) 29Google Scholar), and Limenia is not known elsewhere as such; but no excavations have been undertaken on the site, so the truth about the cult is not attainable.

⁹⁶ As is supposed by Hammond (p. 98)—somewhat surprisingly, in view of the admitted construction of a temple to the same goddess in the late sixth century. Payne, (Perachora i. 25Google Scholar) suggested that the Limenia cult went a long way towards ousting that of Hera Acraea even if it did not fully succeed.

⁹⁷ For clay koulouria from the Geometric and Limenia Deposits, see below, 180 ff.; above, 177. They were also found in the later deposits by the harbour (i.29.11; 33.11, 16).

⁹⁸ The earliest and most useful reference in the literary sources is Xen., Hell. iv. 5. 1 ff.Google Scholar; cf. Perachora i. 16 ffGoogle Scholar.

⁹⁹ Dunbabin, , JHS lxviii (1948) 59 ffCrossRefGoogle Scholar. I hope to argue elsewhere that Corinth is not likely to have been founded by the Argives; but meanwhile it is worth pointing out that Dunbabin's case rested heavily on the conclusion of Payne which is at issue here—that the Argives played a significant part in the foundation of the Perachora Heraeum when the site itself was under Corinthian control, and that the Argives continued to show a strong interest in this cult in Corinthian territory. The case for an Argive foundation of Corinth itself is therefore seriously weakened if I am correct in arguing below that the finds from the Perachora Heraeum do not support Payne's view. Another argument used by Dunbabin was that if the Argives founded both Sicyon and Megara, as is almost certain (see below, pp. 192 f.), they are very likely to have founded Corinth too on geographical grounds; but this is hardly persuasive.

¹⁰⁰ In fact, that was soon after 800 (above, pp. 161 ff.); but Dunbabin, (JHS lxviii (1948) 64Google Scholar) believed that it was in the ninth century, and Payne, (Perachora i. 31)Google Scholar thought that a possible date.

¹⁰¹ It is hardly worth while to list the numerous variations that occur. The pots from Corinth and the Corinthia which have fabric similar to that of finds from Perachora will be noted in the following discussion. All parallels which are quoted I have studied at first hand except where it is specifically stated that I have not: there may therefore be some discrepancy between my description of the colour of the fabric and that found in the original publication, but I hope to have preserved consistency.

¹⁰² Some of the parallels for the fabric of finds from Perachora were found in the Corinthia outside Corinth; it is therefore possible that some Perachora pieces were not made in Corinth itself, but in some other part of the Corinthia. Where such a possibility exists it will be mentioned in the notes; no specific provincial Corinthian fabric can yet be identified, but there may have been local workshops in or near Crommyon and Tenea, near the modern Athikia (below, n. 159).

¹⁰³ This seems to have been the generic word for flat cakes or loaves of all kinds; cf. Ath. 644 b. There were, however, very many special names for flat cakes of different kinds; see the long discussion in Ath. 643 e ff.

¹⁰⁴ References in Courbin, CGA 249.

¹⁰⁵ Verdelis, , PAE 1958, 136Google Scholar, with pl. 113α; Archaeology xv (1962) 191, fig. at bottomGoogle Scholar.

¹⁰⁶ Stroud, , Hesperia xxxiv (1965) pl. 11 d–e, gGoogle Scholar.

¹⁰⁷ At Solygeia the earliest finds are the conical oinochoe and the kotyle Verdelis, , PAE 1958Google Scholar, pl. 111β; Archaeology xv (1962) 188, figGoogle Scholar. at bottom. A very similar kotyle was found at Perachora, and probably dates to the mid seventh century (Perachora ii. 53, no. 380); the conical oinochoe is less easy to place, but must belong to the seventh century (cf. ii.221–3, 231–2, 236). For the date of the Demeter and Kore sanctuary, cf. Stroud, , Das Altertum xi (1965) 10Google Scholar; material dating back to LG was found on the site (Stroud, , Hesperia xxxiv (1965) 2Google Scholar; Hesperia xxxvii (1968) 300)Google Scholar, but the earliest finds clearly associated with the cult date to the last quarter of the seventh century.

¹⁰⁸ The examples from the sanctuary of Demeter and Kore are unlikely to be earlier than the late seventh century, but those from Solygeia could be slightly earlier (above, n. 107).

¹⁰⁹ Waldstein, , Argive Heraeum ii. 117Google Scholar, pl. 58.13; FIG. 10.

¹¹⁰ The Perachora koulouria are illustrated in Perachora i, pl. 16; here, PLATE 38.

¹¹¹ A representative selection of Corinthian birds is reproduced here, FIG. 11. The slight body of the example on the koulouri places it late in the series and the cursory drawing agrees with such a date. The birds found in Argive are usually of a quite different style, and are never so skilfully drawn; cf. Courbin, CGA pls. 126–9, (some reproduced here, FIG. 12).

¹¹² The Corinthian parallels are mainly in the fine hand-made vases of the Geometric period; that is to be expected, for the koulouri was made by hand and then painted. Cf. W 13, 89; KP 979 (unpublished, from Potters' Quarter Grave V; cf. above, n. 49); Charitonides, , AJA lix (1955) 125, no. 2CrossRefGoogle Scholar.

¹¹³ One has been illustrated by Kalligas, , ADelt xxiii (1968) B, pl. 252βGoogle Scholar; there were many other fragments found in the same excavation (op. cit. 312). Kalligas, however, rejects the view that the objects are koulouria (loc. cit.: he uses the word ), and suggests that they should be identified as parts of circular handles that were once attached to clay copies of metal tripods. Such an identification, however, seems improbable. In the first place, if the objects were from clay tripods it is strange that only fragments from the circular handles have been found; a significant number of examples turned up, but if they had originally belonged to tripods one would have expected other parts of the tripods to have been preserved as well—perhaps decorative handle-attachments, the point where the handles were joined to the bowl, the bowl itself, or the legs. Four fragments of legs from clay tripods were found in the Limenia sanctuary at Perachora (ii.3390–3 pls. 126–7); but there were no fragments of clay handles recognized, although that proves little in view of the small number of leg-fragments found. Secondly, the size of the objects from Corcyra is so large that it is not easy to accept that they once belonged to tripods. The width of the illustrated fragment is over 0·060 m., and the diameter of the original circle must have been at least 0·300 m., and may well have been over 0·400m.; the size of all the other fragments is very similar, and such immense handles would have looked foolish on the comparatively small clay copies of tripods which have been found at Perachora (see above). The Corcyra fragments are far larger than the handles of almost all the metal tripods that have been preserved from the Geometric and Archaic periods. In the most recent study of such tripods (Willemsen, , Olympische Forsckungen iiiGoogle Scholar) there are a very few handles of comparable size (e.g. op. cit. 45, Br 4314, 5684, 1265, 2332: width between 0·049 and 0·064 m.; diameter 0·262–0·406m.), especially among the hammered pieces (e.g. op. cit. 139, Br 4311 + 6597, the unpreserved fragment found in 1938/9, Br 9694, Athens NM 7483, Br 10861, B 2041, 2187: width 0·058–0·076m.; diameter 0·307–0·536m.); but these are exceptions, and most examples are much smaller: the width rarely exceeds 0·035 m., and the diameter is usually less than 0·200 m. In metal tripods there is a much greater variety in size than in the fragments from Corcyra, and if the latter were from tripods one would have expected a similar variety. Finally, although the tripod-handle Kalligas quotes as a parallel for the decoration of his objects is very close (Benton, , BSA xxxv (1934/1935) pl. 24. 3Google Scholar, from the Idaean Cave), such handles are not common on metal tripods; there are a few comparable pieces (cf. Willemsen, op. cit. 159 f., pl. 93 bottom), but I have been unable to trace anything else which is as close as the example from the Idaean Cave. That parallel makes it possible that clay copies of metal tripods might have been given handles of this type; but if all the objects from Corcyra are tripod-handles it is very strange that there is no representative of any of the numerous other handle-types, which are far more common in metal tripods. In view of these difficulties, and the strong similarities between the Corcyra finds and the koulouria from Perachora (for a comparison see below, n. 131), there seems little doubt that the Corcyra fragments should be identified as koulouria. The Corcyra excavation also revealed objects of a similar shape but more elaborate decoration, which consisted of stamped lotus-flowers and palmettes (ADelt xxiii (1968) B 312, pl. 252α)Google Scholar. Kalligas himself identifies these objects as (i.e. koulouria), and compares the very similar pieces from the Potters' Quarter at Corinth and from the Limenia sanctuary at Perachora (Corinth xv pt. 2, Class xxxv, 7 and 9; Perachora ii.3464–6 pl. 130)Google Scholar. It is therefore certain that some koulouria were offered at Corcyra (though see below, n. 114), and since the less elaborately decorated fragments are to be identified as koulouria as well, there is probably a chronological distinction between the two types: the examples with stamped decoration will be the later of the two (for a date in the sixth century, cf. Perachora ii. 328)Google Scholar.

¹¹⁴ For references to the preliminary reports on excavations at this sanctuary, cf. Kalligas, , ADelt xxiv (1969) A, 53, n. 12Google Scholar; there is a brief summary by Dontas, , AAA 1968, 66 fGoogle Scholar. For the identification of the deity to whom the sanctuary belonged as Hera Acraea, see now Kalligas, op. cit. 51 f. If I am right in arguing below that the koulouria were originally dedicated in this sanctuary they will provide a strong additional argument in favour of the identification; they were plainly special to the worship of Hera Acraea at Perachora. The koulouria were actually found in the area of a small shrine of Apollo immediately to the south of the temenos wall of the Acraea sanctuary (ADelt xxiii (1968) B, 305Google Scholar, fig. 1 (Section V); the Apollo shrine, op. cit. 309 ff.), and it is therefore not immediately obvious that they were dedicated to Hera Acraea; in fact Kalligas (op. cit. 312) appears to associate them with the Apollo shrine. When I was in Kerkyra in 1968 Mr. Kalligas was kind enough to show me his finds and to discuss them with me; and he suggested then that the koulouria were part of the debris that had been washed down into the Apollo shrine from the Acraea sanctuary on top of the hill (see the plan, op. cit. 305 fig. 1), and that they were originally dedicated in the Hera Acraea votive deposit. It is clear that some of the material found in the Apollo shrine belonged to the Acraea sanctuary (cf. op. cit. 308: kalathoi); and there seems little doubt that the stratigraphy will allow the koulouria to have been dedicated to Hera Acraea. In that case, it is more than likely that they were, in view of the very similar objects that were offered to the same goddess at Perachora. Nothing even remotely comparable has turned up anywhere else in the Greek world in an Apollo sanctuary. The koulouria decorated with lotus-flowers and palmettes (see above, n. 113) which were found at Corcyra can have no special connection with Apollo, despite the suggestion of Kalligas (ADelt xxiv (1969) A, 57)Google Scholar; for similar koulouria were dedicated to Hera in the Limenia sanctuary at Perachora (above, n. 113). Kalligas also refers to a poem from the Greek Anthology (A. P. vi. 155) which demonstrates that cakes covered with cheese (line 4: ) were proper offerings for Apollo; and he argues from this that the koulouria with lotus-flowers and palmettes are to be associated with the Apollo shrine (ADelt xxiii (1968) B, 312)Google Scholar. But this argument does not of course exclude the possibility that they were originally offered to Hera Acraea, and the fact that both types of koulouri found at Corcyra are very closely paralleled at Perachora makes it next to certain that it was to Hera Acraea that the Corcyra koulouria were dedicated.

¹¹⁵ The Geometric Deposit of Hera Acraea at Perachora was closed towards the middle of LG (see above); the koulouria cannot of course be dated except by the rest of the deposit. For the date of the Corcyra deposit, see below, p. 202.

¹¹⁶ The likelihood is that the koulouria found at Corcyra were made locally, and they have therefore not been used as parallels in the following discussion of the origin of the Perachora koulouria.

¹¹⁷ Perachora i. 69Google Scholar. I was fortunate enough to be able to see all the koulouria from the Geometric Deposit in the Athens National Museum. I am most grateful to the Director of the Museum, Mr. Kallipolitis, and especially to the Epimelitria Mrs. Touloupa, for allowing me facilities to study them.

¹¹⁸ In W 77 the attempt was made to chase the added particles away from the surface, but it was not fully successful; that it was made, however, is clear from a comparison of the number of particles on the surface with that of those in the break. The same can be seen in the photographs of a vase from Klenies (Charitonides, AJA lix (1955) 125Google Scholar, pl. 39, figs. 1–2, no. 1); compare the small number of particles on the outer surface (fig. 1; many of the visible marks are divots on the surface, and not added particles) with the greater number inside the rim (fig. 2). The same effect was sometimes gained by the application of a slip to the surface of a hand-made vase; cf. KP 984 (unpublished, from Potters' Quarter Grave I; cf. Corinth xv pt. 1, 7. Dr. J. L. Benson kindly allowed me to mention this vase); Lawrence, Hesperia xxxiii (1964) 90 (M 3), 91 (A 2)Google Scholar.

¹¹⁹ On this technique, cf. Perachora i. 41Google Scholar; Payne calls it typically Argive, but he noted it only in the koulouria and the architectural model Group D from Perachora, and the origin of both is in dispute.

¹²⁰ W 53 has black and red added particles; W 86 has mainly black, but a few red. Both have pale surfaces and are red in the break.

¹²¹ The marks that appear in the photograph are not particles, but divots; cf. above, n. 118.

¹²² For the fabric, cf. KP 177, 181 (unpublished, from Potters' Quarter Grave V, cf. above, n. 49; I am grateful to Dr. J. L. Benson for permission to mention these pieces); both have fired pinkish in parts, while 181 is greenish and 177 greyish in others. There are numerous wheel-made parallels.

¹²³ For the Argive Heraeum piece see above. I was unable to see the greenish tinge mentioned by Payne, (Perachora i. 69)Google Scholar.

¹²⁴ These two examples with striations were identified as possibly Corinthian by Payne; cf. below.

¹²⁵ It was only in i.16.7 that Payne noted this technique, and it is certainly most striking on this example; but he gives the impression that others shared the same peculiarity (Perachora i. 41)Google Scholar.

¹²⁶ Cf. W 86 (above, n. 120).

¹²⁷ Once again I was unable to see the green colour that Payne mentions (Perachora i. 69)Google Scholar; cf. above, n. 123.

¹²⁸ There is more grit in i.16.11 than Payne's description (‘fine’) indicates (Perachora i. 69)Google Scholar.

¹²⁹ i.16.14 is a photograph of the striations; the same pattern appears on i.16.6.

¹³⁰ The same technique is, however, found on the architectural model Group D (below), which is probably Corinthian.

¹³¹ The majority of the Corcyra koulouria had a finer clay than those from Perachora, but a few had added particles; none had them sprinkled on the base as in i.16.1, 7–9. There were no added spirals at Corcyra; the examples from the colony were all decorated with degenerated scallops round the outer edge which had often become mere nicks. The closest of the Perachora koulouria is i.16.13–14; but that piece must be earlier than those from Corcyra, since its scallops are well formed. The Corcyra fragments never have striations on the base.

¹³² Courbin (CGA 249) distinguishes two types of koulouri, and he claims both as Argive. The first includes the Perachora examples and that from the Argive Heraeum, while the second is represented at Argos and at Tiryns. It seems that there is one Corinthian and one Argive type.

¹³³ I was unable to see any of these models in the Athens National Museum except Group A, which is on general exhibition. I have therefore been forced to discuss the others on the basis of Payne's descriptions.

¹³⁴ Groups A and D (below).

¹³⁵ Group A.

¹³⁶ For the step-meander, cf. Courbin, , CGA 551 n. 3Google Scholar. Examples: op. cit. pls. 42 (C 210), 43 (C 201), 48 (C 239). Multiple-brush squiggles: op. cit. pls. 47 (C 915), 80 (C 3823), 125 (C 3846).

¹³⁷ Perachora i. 41Google Scholar; I was able to see the object in the National Museum.

¹³⁸ There is a small fragment from Athens that may come from a similar model (BCH lxxvi (1952) 202 fig. 1)Google Scholar. The fragment is so tiny that it is not certainly from such a model, but it is undoubtedly Argive (see Courbin, , CGA 551 with n. 7)Google Scholar. Another model comes from Gela; it was probably made in Sicily (Orlandini, , NS 1956, 273 f.Google Scholar, fig. 1).

¹³⁹ Aetos R 600; Robertson, , BSA xliii (1948) 101 f.Google Scholar, Pl. 45.

¹⁴⁰ Op. cit. 101.

¹⁴¹ Robertson remarks that Argive rows of figures in LG often have short arms as on fr. g of the model; but most Argive examples still have longer arms, and Argive figures are usually done in silhouette and not in outline as on the model. For a selection of Argive fragments with rows of figures, cf. Courbin, CGA pls. 145 ff.

¹⁴² Professor Robertson points out to me that Payne wondered if the picture on VS pl. 20.1 might not be a copy of a Geometric wall-painting in a temple (Necrocorinthia 8 n. 1). I would not stress the claim that the model and the aryballos could be from the same hand; but the similarity is none the less striking.

¹⁴³ Cf. W 77 (above, n. 118). There are two similar vases in Delphi (Mus. Inv. 5968, 7695) that are almost certainly Corinthian; I am grateful to Professor P. Amandry and Professor L. Lerat for permission to mention these vases.

¹⁴⁴ For the two methods—chasing away the particles, which was done all over the visible surface of Group A, and applying a slip, which was done only where the meander was to be painted—cf. above, n. 118.

¹⁴⁵ Cf. Lawrence, , Hesperia xxxiii (1964) 90, M 3 (EG hand-made aryballos)Google Scholar; C–63–650 (unpublished large hydria from the MG I well at Anaploga; for the well, see above, n. 9. Dr. Keith DeVries kindly allowed me to mention this vase).

¹⁴⁶ Cf. W 52; Corinth xiii. 14–1, 15–1, 17–5Google Scholar.

¹⁴⁷ Corinth xiii. 70–1Google Scholar; but in this vase (of the PC period) there are added particles throughout. A finer vase is KP 177 (above, n. 122), whose clay is greyish in parts and pinkish in others.

¹⁴⁸ Cf. above, n. 118. There is also a slip on the surface of the model Group A where the meander hooks are painted. It is natural for the whole surface of Group D to be slipped, since the decoration covers the whole of the surface.

¹⁴⁹ Cf. VS pl. 3. 3. But examples are also known in Argive; cf. Cook, , BSA xlviii (1953) 80Google Scholar, pl. 28d.2 (though this fragment from Mycenae might be Corinthian); Courbin, CGA pls. 25 (C 11), 49 (C 244), 64 (C 235), 124 (C 4652).

¹⁵⁰ Perachora i. 35Google Scholar. The find-place, however, is not a good indication of date, since the deposit was not properly stratified (Perachora i. 31 f.)Google Scholar.

¹⁵¹ The roof of the model Group A is not fully preserved, and cannot be reconstructed with any great accuracy; but that it was curved and had a comparatively steep pitch is certain (Perachora i. 35)Google Scholar. Model D: Perachora i. pl. 120. Fr.a of the Ithaca model (Robertson, , BSA xliii (1948) pl. 45)Google Scholar shows that the pitch of its roof was similar to that of Group D; the chequers of the roof of the Ithaca model must surely represent tiles (Robertson, op. cit. 102; contra, Cook, R. M., BSA lxv (1970) 17 n. 1)Google Scholar, which go with a less steeply pitched roof.

¹⁵² Payne rightly remarked that the technical peculiarity shared by this model and the koulouria i.16.1, 7–9 must be significant; but he used the point to argue from an Argive origin for the koulouria to the same origin for the model (Perachora i. 41)Google Scholar.

¹⁵³ Cf. also W 66; I have, however, been unable to see Group E.

¹⁵⁴ Group A is compared with some of the Perachora koulouria, and with the hand-made krater i.12.5 (Perachora i. 42)Google Scholar; for the latter vase, see below. For the parallels noted by Payne for Group B (relief ware) and Group D (koulouria), see above.

¹⁵⁵ I was able to see a only few of the fragments examined below in the National Museum; once again, therefore, I have been obliged to discuss the rest on the basis of Payne's descriptions. For i.15.1, see Perachora i. 65Google Scholar. Courbin, (CGA 551 n. 1)Google Scholar does not mention the fabric; it may therefore be assumed that it is not distinctively Argive.

¹⁵⁶ The strange disposition was noted by Courbin; CGA 551 n. 1.

¹⁵⁷ The diamond-chain, Corinthian: Aetos R 44, 112, B 752; i.11.6–7; ii.488–9, 958, 975. The diamond-chain, Argive (always very roughly done): Courbin, CGA pls. 36 (C 645), 37 (C 2362), 63 (C 2441). In Argive there is one example of vertically disposed groups of chevrons: Courbin, CGA pl. 116 (C 3074). In Corinthian there are no certain examples; but it is possible that on Aetos B 813 there were more groups than the single one that is preserved, since such a pattern would have balanced the horizontal zig-zags to the left. The normal Corinthian disposition is diagonal within the panel (i.B.i; Aetos B 864, 864a), and there is a Corinthianizing piece from Argos with the same disposition: Courbin, CGA pl. 121 (C 4366). There are examples from the Limenia Deposit at Perachora with horizontally disposed groups (ii.118) and with single horizontal columns (ii.565, 658, 689).

¹⁵⁸ In Aetos B926 there are two different motifs one on top of the other without the normal horizontal dividing line; but they are separated by a horizontal wiggly line instead, and that line should be seen as doing the duty of the regular straight one rather than as a separate motif (Benton, , BSA xlviii (1953) 342, fig. 33)Google Scholar. There is undivided decoration on some drinking vessels in the handle zone (ii.452–5, 653–6), but the parallel does not convince, for the motifs are much simpler than in i.15.1. Similar decoration is found on fragments from Megara Hyblaea (Vallet, and Villard, , Mégara Hyblaea iiGoogle Scholar. pls. 1.6–7; 4.5; 147.1; 153.2). The first three are published as Corinthia n (op. cit. 16, 26), and the others as local without Corinthian influence (op. cit. 152); but the last two are clearly very similar to the first, and probably followed Corinthian models (but see op. cit. 152 f.). However, in all these cases from the Sicilian colony the motifs are simple and their panels small, just as in the drinking vessels from the Limenia Deposit; there is no real comparison with i.15.1.

¹⁵⁹ The dark fabric is similar to that of two vases from a grave-group at Athikia (W 71–72; for the group, cf. Weinberg 19) and to that of some of the vases of the MG II period from the cemetery at Crommyon (not yet fully published; for a preliminary report, cf. ADelt xvii (1961/1962) B, 52 ffGoogle Scholar. I am most grateful to the excavator of the cemetery, Miss O. Alexandri, and to Mrs. Protonotariou-Deilaki, Ephor of the Argolid and the Corinthia, for their kind permission to study the vases from Crommyon in the museum at Corinth). The decoration of i.15. 1 could have been developed in a provincial Corinthian workshop, but the suggestion is very much a matter of conjecture; provincial Corinthian kraters are much more likely to have followed the Corinthian pattern.

¹⁶⁰ There are vases decorated in a similar manner to the examples quoted above, n. 158, in Euboean and from Arcadia. Euboean: cf. Boardman, , BSA xlvii (1952) 3, fig. 1.2Google Scholar; 5, fig. 3.6. Arcadia: BCH lxxxvii (1963) 765Google Scholar, fig. 10, top left (Gourtsouli near Mantinea; cf. op. cit. 766).

¹⁶¹ In the absence of parallels the fragment cannot be securely dated; but the light-ground style is most unlikely to be MG, whatever the origin of the piece.

¹⁶² Once again Courbin does not mention the fabric of the piece, so it is not likely to be distinctive (CGA 551 n. 1). There seem to be Corinthian parallels; cf. W 106 (but see below, n. 188); Corinth xiii. S–8 (I was, however, unable to see this vase in the Corinth museum). Cf. also the dark fabric of some vases from Athikia and Crommyon (above, n. 159).

¹⁶³ Corinth xiii. 44Google Scholar. In more detail, Courbin, , CGA 204 ff.Google Scholar; the high rim is not quite universal, but it is normal.

¹⁶⁴ Corinthian: Aetos R 59; B 789, 794, 887; i.123.4; Corinth xiii. S–7. Argive: Courbin, CGA pls. 40 (C 240), 100–1 (C 209), 139 (C 33).

¹⁶⁵ See above, n. 159.

¹⁶⁶ The method of hatching the meander does not help to determine the origin of the piece. In i.15.3 there is a change of direction in the hatching at the corners; but this occurs in both Corinthian and Argive. Argive: Courbin, CGA pls. 2 (C 833), 9 (C 891). Corinthian: W 80,83. Other examples in each fabric have no such change in direction. Argive: Courbin, CGA pls. 2 (C 834), 4 (C 28). Corinthian: W 37, 58.

¹⁶⁷ There are also later examples of purplish paint in Corinthian: ii.81, 826.

¹⁶⁸ Cf. Courbin, , CGA 551 n. 1Google Scholar; Courbin does not commit himself on the origin of the vase.

¹⁶⁹ Perachora i. 65 for the descriptionGoogle Scholar; the fabric is thick, the clay leather-brown, and the paint black.

¹⁷⁰ Argive: Courbin, CGA pl. 58 (C 617, 1075). Corinthian: Aetos B 619, 663a; W 108 (section, Weinberg 36 fig. 12).

¹⁷¹ The foot: Constantinou, PAE 1952, 161 fig. 8, rightGoogle Scholar. The fabric: Boardman, , BSA lii (1957) 2Google Scholar; cf. Coldstream 190.

¹⁷² W 182; Weinberg, , Hesperia xvii (1948) 213–14, pl. 77, C 25Google Scholar; Orsi, , MA xxv (1918) 558 fig. 147Google Scholar.

¹⁷³ Brann, , Agora viii. 59 pl. 13, no. 243Google Scholar. The vase has numerous red added particles, and the clay is a reddish buff on the surface; the core is grey. Corinth xiii. 70–1Google Scholar (PC) is a close Corinthian parallel from the North Cemetery; the model Group D from Perachora has a similar clay, but there are no added particles apart from those that were sprinkled on the base. Brann (op. cit. 59) states that the vase is possibly Argive but probably Corinthian. Corinthian hand-made vases were certainly exported to Athens during the latter part of the Geometric period; Brann, op. cit. 59, no. 242 is, without doubt, of that origin.

¹⁷⁴ Cf. Perachora i. 42, 65Google Scholar. The attempt to chase the particles away from the visible surfaces was less successful in the koulouria and the krater than in Group A. There is possibly a fragment from a vase like i.12.5 from the sanctuary at Aetos on Ithaca. It has the same fabric, and the handle and the rim are similar; but it has not been published.

¹⁷⁵ Corinthian: W 53, 82, 86 (Stillwell, , AJA xl (1936) 43 fig. 21)Google Scholar; Corinth xiii. 14–1, 15–1, 16–9, 20–3Google Scholar; Weinberg, Hesperia xvii (1948) 212 f.Google Scholar, C 16–19; W 171 (PC). Argive: Courbin, CGA pls. 91 (all examples), 94 (C815, 2460, 2451).

¹⁷⁶ W 173 (PC).

¹⁷⁷ References above, n. 175 (Corinth xiii). There are vertical handles in the same position as the horizontally set examples on the Perachora vase on Weinberg, , Hesperia xvii (1948) C 17Google Scholar.

¹⁷⁸ Courbin, (CGA 551 n. 1)Google Scholar describes the objects as ‘couvercles trilobés’; he does not claim them as certainly Argive.

¹⁷⁹ For i.124.4 (pale red clay) Payne compares the koulouria (Perachora i. 66)Google Scholar; see above, pp. 183 f. For the greyish-green clay of i.124.2–3, 5, cf. the model Group B and the parallels quoted above, n. 145.

¹⁸⁰ Perachora i. 61Google Scholar; cf. also Courbin, , CGA 550 n. 5Google Scholar.

¹⁸¹ Geometric: C–63–651, an oinochoe from the Anaploga well (see above, n. g; my thanks are due to Dr. Keith DeVries for his permission to mention this vase); W 43, 56–7, 59, 85. The last vase is claimed as possibly Argive by Weinberg (p. 30) and certainly not Corinthian; but the other examples make it clear that it could well be Corinthian. PC: W 136; ii.10, 17, 54, 116, 280 (I have not been able to see these vases from the Limenia Deposit).

¹⁸² KP 172, from Potters' Quarter Grave V: illustrated by Coldstream, pl. 18g. The Perachora fragment, however, is not complete enough to make it certain that the two vases were of the same kind.

¹⁸³ It is instructive that Courbin did not claim this piece as Argive despite the cross beneath the handle (CGA 550 n. 5); Benton, (BSA xlviii (1953) 262)Google Scholar believes it to be Corinthian.

¹⁸⁴ Perachora i. 62 with n. 7 (cf. Courbin, , CGA 550 n. 5Google Scholar); the vase is described under the number i.14.4 (Perachora i. 62Google Scholar).

¹⁸⁵ Cf. Courbin, CGA pls. 49 (C 1019), 124 (C 3391).

¹⁸⁶ ii.743. The fact that the leaves are solid on the Limenia piece and done in outline on the other matters little.

¹⁸⁷ Courbin, , CGA 550 n. 5Google Scholar; Perachora i. 58, 59Google Scholar.

¹⁸⁸ Corinth xiii. 22–1Google Scholar; Aetos B 1029, R 356 (both local, but doubtless derived from Corinthian models). W 106 is a similar vase, but was also claimed as possibly Argive by Courbin (CGA 550 n. 3); but such vases were certainly made at Corinth, so it is more than likely that the examples from Perachora and Corinth are of the same origin. There is another Corinthian example of the shape from Leporano near Tarentum (Porto, Lo, Bollettino d'Arte xlix (1964) 76, 77, fig. 23Google Scholar; NS xviii (1964) 224, 223, fig. 44.2Google Scholar).

¹⁸⁹ Corinthian: i.13.18; Aetos B 620; W 46 (earlier and more elaborate). Argive: Courbin, , CGA 311 n. 2Google Scholar.

¹⁹⁰ Still LPG are Williams, , Hesperia xxxix (1970) 20, pl. 9, nos. 33–4Google Scholar. Later examples are Corinth xiii. 16–5 to 8Google Scholar from the Nort h Cemetery; Corinth xii. 280Google Scholar, nos. 2258–9; 281, no. 2262. Some of these examples have three globes. Corinth xiii. 17–6, 7Google Scholar are complete spits similar to the type Perachora i. 70 fig. 11.3. There are similar pins from a grave of the MG I period at Klenies in the Corinthia (Charitonides, , AJA lix (1955) 126 fGoogle Scholar., pl. 40, fig. 14, no. 12).

¹⁹¹ Some analysis of the Argive Heraeum and Argos spits has been made (Courbin, , Annales xiv (1959) 209 ff.CrossRefGoogle Scholar); but it was for a quite different purpose, and the Perachora examples were not analysed.

¹⁹² Perachora i. 34, 73Google Scholar. The objects concerned are i.17. 11–12, 19. For the Argive Heraeum material, cf. Perachora i. 73Google Scholar (on i.17.19).

¹⁹³ i.18.20 (Perachora i. 75Google Scholar); cf. Boardman, , Island Gems 112 ffGoogle Scholar. The type is identified by Boardman as of island origin, later copied in Argos (see further Boardman, , JHS lxxxviii (1968) 7)Google Scholar.

¹⁹⁴ i.18.30–1; cf. Boardman, , Island Gems 110, n. 1Google Scholar. Payne catalogues four stone seals (Perachora i. 75)Google Scholar, but mentions only three in the list of objects from the Geometric Deposit (Perachora i. 34Google Scholar, no. 7); the piece i. 18.32 is probably not a seal, but a spindle-whorl (cf. i.18.33–4).

¹⁹⁵ i.13.8–9; 15.4; 122.3–4, 7; 12.4 (CGA 550 n. 5) and i.14.3; 124.2–5 (CGA 551 n. 1).

¹⁹⁶ Cf. Coldstream 353 n. 2. The only vase which may be earlier than LG and which was claimed as Argive by Payne is i.15.3; but the piece could equally well be Corinthian, and it is so small that there can be no certainty about the date.

¹⁹⁷ The stone seal i.18.20 is certainly Argive; and some of the metal finds may be of the same origin.

¹⁹⁸ Cf. most conveniently Coldstream 399 ff. (Site Index).

¹⁹⁹ It is possible that offerings of perishable material, again perhaps real koulouria, were made to Hera Acraea at a later stage in the history of the cult (see above, p. 178).

²⁰⁰ On Apollo Pythaieus, whose worship Hanell (84 f.) argues came from Argos to Megara, see now further Barrett, , Hermes lxxii (1954) 421 ffGoogle Scholar. See also Geisau, von, RE xxiv (1963) 170 fGoogle Scholar.; Gomme, , Dover, , and Andrewes, , Historical Commentary on Thucydides iv. 71Google Scholar.

²⁰¹ The argument is exactly parallel to that of Dunbabin, who supposed that the Heraeum was Corinthian at the time of its foundation; he used the objects from the site that were claimed as Argive as an argument in favour of an Argive foundation of Corinth (cf. above, n. 99).

²⁰² See however below, pp. 193 ff. It is argued there that there was a time when the Perachora peninsula was controlled by Megara, and that the cult of Hera Acraea was probably brought to the site by the Argives when they founded Megara. Independently of the evidence from the Heraeum, it is certain that the Megarians held Hera in special honour (below, p. 194); since Hera was a specifically Argive goddess, that fact can still be used to support the argument in favour of an Argive foundation of Megara.

²⁰³ Paus. ii. 6. 7; cf. Dunbabin, , JHS lxviii (1948) 63Google Scholar.

²⁰⁴ Hanell 76 n. 1. The passage on which he relies, however (Xen. Hell. iv. 5. 1), does not prove that Corinth was an Argive colony, as Hanell claims; cf. Dunbabin, , JHS lxviii (1948) 63Google Scholar.

²⁰⁵ Hanell was followed in this by Dunbabin, op. cit. 64 f.

²⁰⁶ See briefly above, n. 99.

²⁰⁷ If the genealogical evidence is to be followed without question, Sicyon will have been the earlier foundation, for it was in the generation after Temenus that Sicyon was Dorized (Paus. ii. 6. 7), while Megara was one generation later (below); such arguments, however, hardly inspire confidence.

²⁰⁸ I hope to discuss the date of the foundation of Corinth in detail elsewhere; see meanwhile Dunbabin, , JHS lxviii (1948) 62 fGoogle Scholar. (c. 900 B.C.).

²⁰⁹ An early excavation at the sanctuary of Zeus Aphesius in the Megarid uncovered what could possibly be a PG sherd; it is described as a small fragment of a phiale-shaped vase decorated with concentric circles (Philios, , EA 1890, 45Google Scholar). The shape is rather strange for the PG period, although there are possible parallels (Kübler, , Kerameikos iv pl. 23, nos. 1092, 2112Google Scholar; neither is decorated like the present fragment); but the concentric circles are typical of that style, and the piece may belong to the tenth century. One sherd, however, can hardly be used to date the foundation of a city; there is in any case nothing to prove that it is connected with Dorian Megara if it is PG, and it was of course not found at Megara itself.

²¹⁰ Plut. Qu.Gr. 17. Tomlinson, (BSA lxiv (1969) 235)Google Scholar has questioned the validity of this evidence, although he gives no reason besides the lack of material evidence for Megarian connection with either the Heraeum or its promontory (see below). He does, however, suggest that ‘the passage may … derive from a Megarian attempt to justify a claim to this region, made perhaps in Hellenistic times’. Plutarch must of course always be treated with some care as a source for early Greek history. But he did sometimes have access to respectable information, and there is some reason to believe that much of the material in the Quaestiones Graecae derives from Aristotle's Politeiai (cf. Ziegler, , RE xxi (1951) 862 fGoogle Scholar., with references); the Megarian matter in the work (Qu.Gr. 16–18, 59) ma y well rest on Aristotle's Constitution of the Megarians (cf. Halliday, , Plutarch, Greek Questions 92 ffGoogle Scholar.). In the particular case of Qu. Gr. 17 there is some independent corroboration in the fact that there seem to have been five constituent parts of the Megarian population for administrative purposes (cf. below, with n. 211); it is more than likely that this fivefold division goes back well beyond the Hellenistic period (Hanell 138 ff.), and it is in any case unlikely to have been based on a fiction of the kind Tomlinson suggests.

²¹¹ Cf. Hammond 96. The strongest argument, however, is not used by Hammond. The five divisions still seem to have been separate for some administrative functions in the Classical period (Hammond 95; Hanell 138 ff.); if that is so, they must have been parts of the Dorian state.

²¹² There is a long bibliography of discussions of the identifications of the five regions. Cf. Halliday, , Plutarch, Greek Questions 97 fGoogle Scholar. (with earlier references); Meyer, E., RE xv (19.31) 168 fGoogle Scholar.; Perachora i. 20 fGoogle Scholar.; Hammond 98 f. The Megareis must have been centred on Megara itself; the Tripodiskioi had their κώμη on the eastern slopes of Geraneia (cf. Thuc. iv. 70. 1–2).

²¹³ Hanell 76. It may be added that the cult of Hera was especially widespread in Megarian colonies; cf. Hammond 96 (but see also below, p. 200 with n. 242).

²¹⁴ This has been the view of all commentators known to me; see the works cited above, n. 212.

²¹⁵ Strabo states this of Crommyon: see below.

²¹⁶ Hammon d (pp. 96, 98) also claims that the goddess was worshipped under this title at Phlious, Megara Hyblaea, Chalcedon, and Byzantium; but his references (96 n. 12) merely speak of sanctuaries of Hera (sometimes on ἄκραι), and do not mention her title. Argos: Paus. ii. 24.1. Corinth: below, p. 200 with n. 241. Corcyra: above, pp. 181–2 with n. 114.

²¹⁷ Nilsson, , Geschichte der griechischen Religion 428Google Scholar; Hanell 75 ff.

²¹⁸ See above, n. 99. It should, however, be noted that some difficulty remains in connecting the Perachora cult of Hera Acraea with that in Argos whatever we suppose the precise nature of the connection to have been; for in Argos her sanctuary was on the acropolis (Paus. ii. 24. 1), and the meaning of Acraea will therefore have been ‘of the Heights’, while at Perachora the meaning must have been ‘of the Promontory’ (see above, p. 176). But we do not know enough about the precise significance of the title Acraea in the cult to be able to say that there could have been no connection between the Argive sanctuary and that at Perachora, and it seems unlikely that the two were completely independent. There is some evidence that at least in the Corinthia the title Acraea could cover both kinds of ἄκραι, for the Perachora sanctuary was on a promontory while that in Corinth itself was not (below, n. 241); these two sanctuaries can hardly be independent of each other (see further below, pp. 202 f.).

²¹⁹ It was certainly founded by the time of the first preserved dedications c. 800 B.C.; but the first use of the site for the worship of Hera could have taken place well before that.

²²⁰ References above, n. 212.

²²¹ Hammond 99. It is, however, just possible that the name derives from some other word—perhaps a village-name, as the Megareis were named after Megara; and the form Πιραεῖς may not be correct (see below, n. 224).

²²² Not necessarily Corinthians, as Payne suggested (Perachora i. 21Google Scholar); cf. Hammond 99 n. 29. Hammond supposes (loc. cit.) that the name could have been given by any Peloponnesians to men who lived opposite them outside the Peloponnese; but it was in fact used by Megarians for their fellow citizens.

²²³ Theoretically Salamis is a third possibility; but the word is never used of an island opposite the mainland, and it is most unlikely that Salamis did not have a more individual name.

²²⁴ Professor Huxley has very kindly suggested to me that there may be a connection between Plutarch's Piraeis and Cape Spiraeum, a promontory in the south-eastern Corinthia mentioned by Thucydides (viii. 10. 3). He points out that the form Πιραεῖς in Plutarch is strange if it is to be connected with the adjective Πιραεῖς, ‘opposite’, for it is only in Attic (and then rarely) that the form πειρ- occurs (πειραιεύς has this etymology: cf. Judeich, , Topographie von Athen 47Google Scholar; Schmidt, J., RE xix (1937) 100Google Scholar); the Doric form is περ- (Corinthian: IG iv. 329Google Scholar, ), and it is not easy to see how that could have become πιρ-, as in Πιραεῖς. This gives some ground for suspicion of the form Πιραεῖς, but there remain difficulties in connecting it with Cape Spiraeum. In order to accept the connection we must either emend Plutarch to read Σιραεῖς despite complete agreement in the manuscripts in the reading Πιραεῖς; or accept the reading Πειραιὸν in the manuscripts of Thucydides even though an inscription of the third century refers to a landmark in just this area of the Corinthia as Σπίραιον (IG iv. 926Google Scholar. 18), and a papyrus copy of this passage in Thucydides probably read Σπείραιον (P. Oxy. 1247. 42; for a discussion cf. Grenfell, and Hunt, , The Oxyrhynchus Papyri x. 129)Google Scholar. There seems little doubt that the form Σπείραιον is correct for the cape mentioned by Thucydides, and emendation of Plutarch to read Σπιραεῖν is therefore the more likely alternative; but it must remain doubtful, and if Πιραεῖς is to be emended at all we should perhaps read Περαεῖς.

²²⁵ Strabo 380: .

²²⁶ It is possible that the period to which Strabo refers is that between the destruction of Corinth by Mummius and its refoundation by Caesar; but that is hardly likely.

²²⁷ Corinth viii, pt. 3, nos. 1, 5. Each inscription has the significant Corinthian Iota (σ), for which the Megarians used the simple 1; and no. 1 also has the Corinthian San (M), for which the Megarians wrote Sigma (σ). Less important, but still probably significant, are the forms of Lambda (nos. 1 & 5) and Chi (no. 5), which are Corinthian ( and ) rather than Megarian (Λ and +). The Pi () in no. 1 is intermediate between Corinthian (Π) and Megarian (Г), and only the Alpha of no. I () follows the Megarian and not the Corinthian form (A, found in no. 5). For Corinthian and Megarian abecedaria, cf. Jeffery, , Local Scripts of Archaic Greece 114Google Scholar (Corinthian), 132 (Megarian).

²²⁸ Plut., Qu.Gr. 17Google Scholar: τῶν δὲ Κορινθίων πόλεμον αὐτοῖς ἐξεργασαμένων πρὸς ἀλλήλους (ἀεὶ γὰρ ἐπεβούλευον ὑφ᾿ αὑτοῖς ποιήσασθαι τὴν Μεγαρικήν), ὅμως δι᾿ ἐπιείκειαν ἡμέρως ἐπολέμουν καὶ συγγενικῶς.

²²⁹ F.Gr.Hist. 327 F 19. The version reproduced by Jacoby is taken from Schol. Pind., Nem. iii. 155bGoogle Scholar; it appears in very much the same form in Ar., Schol.Ran. 443Google Scholar.

²³⁰ Pind., Nem. vii. 155Google Scholar; Ar., Ran. 443Google Scholar; Ar. fr. 509.

²³¹ The context in which Pindar uses the words does not fit well with the story in Demon: their reference is not clear, but they seem to have been especially connected with children (τέκνοισιν).

²³² The same episode is mentioned by a number of other sources; but except for one case discussed below, p. 198 (Zenobius), there is no significant divergence from Demon. A named fragment of Ephorus (F.Gr.Hist. 70 F 19) repeats the same story but omits the detail of the demands made upon Megara to send mourners to Bacchiad funerals; but in the second part of the story there is a surprising verbal agreement between Demon and Ephorus. This fragment of Ephorus comes from Plat., Schol.Euthyd. 292Google Scholar e, where the historian is only quoted along with Aristophanes (above, n. 230) and Plato (Euthyd. 292 e); but neither of these authors, where their full text is preserved, does more than refer to the words , so Ephorus too need not have explained the proverb. It is therefore only certain that Ephorus used the words , not that he explained them in the same way as Demon (cf. Jacoby ad loc.); the scholiast may merely have taken the full explanation straight from Demon, and that will explain the strange verbal agreement. It is not easy to see why Hammond supposes that the whole story was not in Ephorus Book i (Hammond 97); there is no evidence for that view, and the historian certainly used the phrase . The Suda s.v. Διὸς Κόρινθος does not give the full story as in Demon, but only the episode of the ambassadors and the stoning; Hesychius s.v. does not explain the proverb with a story at all, but merely states its meaning; Zenobius v. 8 has a quite different story, and does not (as Hammond claims) offer the version found in Demon (see below, p. 198). Add to Hammond's references Apostolius vi. 17 (heralds and stoning); Zenobius iii. 21 (meaning of alone, with no story).

²³³ Hdt vi. 58. 2–3; Paus. iv. 14. 5 (Tyrt. fr. 7 West).

²³⁴ Even if the information in Zenobius is invented (as is perfectly possible), that does not restore our confidence in that of Demon, which will then be a distortion of an invention. Other paroemiographers (Diogenianus vi. 34; Apostolius xi. 10) explain the phrase by the story that the wife of a Megarian king forced the Megarians to mourn her husband's death; that too seems like a garbled version of the tale that appears in Zenobius, since it was hardly unususal for the death of a king to be mourned. The phrase was also explained by the fact that Megara grew garlic (cf. Zenobius v. 8; Apostolius xi. 10).

²³⁵ IG vii. 52Google Scholar:

.

²³⁶ The only evidence Hammond adduces for the Corinthian appropriation is from the passages discussed above, pp. 197 f.

²³⁷ It has been generally accepted that the exploits of Orsippus were against Corinth (cf. Hanell 72 n. 3); Payne, (Perachora i. 24 n. 1)Google Scholar claimed that the view cannot be seriously questioned. That is certainly going much too far, since it cannot be supported by a scrap of evidence.

²³⁸ The epitaph of the Athenians who fell at Artemisium (Plut., Mal. Her. 34)Google Scholar gives the impression that the battle was a Greek victory, and that the Persian fleet was destroyed; that is of course very far from the truth. Plutarch's method in that work is often to contradict Herodotus on the evidence of such epitaphs.

²³⁹ The epitaph is perhaps by Simonides; cf. Hammond 97 n. 19. But there can be no certainty.

²⁴⁰ All that we know is that there was a time in the history of the cult at the Heraeum when it was controlled not by Corinth but by Megara; that is not to say that Dorian Corinth existed at that time, since we do not know when Megara was founded. Megara could well have controlled the cult before the foundation of Corinth (cf. above p. 193).

²⁴¹ It has been argued that the cult of Hera Acraea which was thought to have existed at Corinth itself is in fact that at Perachora (Picard, , RA Series 5, xxxv (1932) 218 ff.Google Scholar). Payne, however, demonstrated that there must have been a sanctuary of the goddess in the city (Perachora i. 19 f.Google Scholar); it is very likely that it was to be found on the roof of the spring Glauke (Corinth i, pt. 2, 149 ff.Google Scholar), and that is hardly an inland height. There was a sanctuary of Hera on the way up Acrocorinth; but she was known there as Bounaia (Paus. ii. 4. 7).

²⁴² See the references in Hammond 98 n. 24; in no case is there any evidence in the sources he quotes for the title of Hera at the sanctuaries concerned, but some of them are stated to have been situated on ἄκραι. Cf. above, n. 216.

²⁴³ For the position of the sanctuary at Corcyra, cf. ADelt xviii (1963) B, 161 f.Google Scholar, with figs. 1, 4. For another significant point of similarity between the cult of Hera Acraea at Perachora and that of the same goddess at Corcyra, see below, pp. 202 f.

²⁴⁴ See however above, with n. 242.

²⁴⁵ There is a Corinthian oinochoe from Megara illustrated by Weinberg pl. 11, top left (National Museum 13679). It is dated by Coldstream (353 n. 3) in EG; but the four reserved bands on the body surely make it MG I (cf. Coldstream 94 for this criterion). The vase was found with Nat. Mus. 13678, a Corinthian MG I amphora; the two vases were probably from a single grave, although the circumstances of the find are not recorded. For Corinthian LG from the Megarid, cf. PAE 1934, 55Google Scholar, fig. 15. Coldstream (p. 353 n. 3) also argues for regular use of Corinthian ware by Megarians from the presence of Corinthian MG II vases in the cemetery at Crommyon; but it is argued here that Crommyon was part of the Corinthia at the time of these graves (on Crommyon, see below, n. 256).

²⁴⁶ It might be argued that the graffiti on two vases from the Geometric Deposit (i.123.10–11; Perachora i. 66Google Scholar demonstrate Corinthian control of the site; both are published as reading M, and that is a Corinthian letter (Jeffery, , Local Scripts of Archaic Greece 114Google Scholar). But we might also read Σ, the Megarian sigma (op. cit. 132); there are marks on either side of the M on i.123.10, but they are not certainly letters, so they cannot determine whether we are to read M or Σ. It is also possible that the marks which appear to be sans or sigmas are not letters at all; if they are they must be among the earliest known graffiti (Jeffery, op. cit. 12 ff. for the date of the introduction of the alphabet (c. 750); for the date of the closing of the Geometric Deposit (c. 735), see above). See further Jeffery, , Local Scripts of Archaic Greece 120 n. 3Google Scholar.

²⁴⁷ It is noticeable even within the Geometric period, before the Limeni a Deposit received the vast bulk of its votives; on the few early vases from the Geometric Deposit, see above, pp. 161 f., and on the LG vases from the two deposits, above, pp. 168 ff.

²⁴⁸ See above, p. 176; the over-estimation was shared by Payne, (Perachora i. 24 f.Google Scholar).

²⁴⁹ This geographical identification depends on the location of the Heraeis and the Piraeis; in fact it is likely that not only these regions once belonged to Megara, but the area immediately to the south of the Isthmus as well (see above).

²⁵⁰ It can hardly be argued with any confidence that the dedication of the Corinthian model at Aetos (see above) demonstrates that models were a typically Corinthian form of dedication. That may be true; but one Argive model was offered to Hera at the Argive Heraeum (see above), and another non-Corinthian model has turned up at Gela (see above, n. 138).

²⁵¹ In view of the close Corinthian connections with koulouria of this type it is perhaps worth pointing out that Athenaeus, in his long discussion of flat cakes (above, n. 103) mentions a type known in the Corinthian colony of Syracuse and called ὲπικύκλιος (Ath. 645 e–f); the name probably indicates that it was shaped like a ring, as the Perachora and Corcyra koulouria were, rather than like a disc.

²⁵² For the date of the sanctuary at Corcyra, cf. ADelt xx (1965) B, 391Google Scholar; Dontas, , AAA i (1968) 66 f.Google Scholar

²⁵³ I assume that the Perachora cult was, so to speak, the parent of that in Corinth rather than vice versa, although there is of course no direct evidence. If the Perachora sanctuary was the earlier, the appearance in the Corinthia of the Argive cult of Hera Acraea can be satisfactorily explained by supposing that the sanctuary at Perachora was Megarian in origin; but no such explanation can be offered if the first Corinthian sanctuary of Hera Acraea was in the city itself.

²⁵⁴ Payne proved that the temenos of Hera referred to by Euripides must have been in the city itself, and cannot have been that at Perachora (see above, n. 241).

²⁵⁵ Euripides seems to have mentioned the burial of the children of Medea in the temenos of Hera Acraea at Corinth in order to explain the ceremony that was current there in his own time (cf. Page, Euripides: Medea xxviii f.) that will mean that the ceremony itself must be dated significantly earlier than Euripides' explanation of it. The sanctuary of Hera appears in the story of Medea as early as Eumelus (Paus. ii. 3. 10 f.: Eumelus fr. 3 Kinkel); but it is not possible to demonstrate that this eighth-century sanctuary mentioned by Eumelus (on his date, see most recently Huxley, , Greek Epic Poetry 62Google Scholar; it seems reasonable to suppose that Eumelus had a real sanctuary in mind) was in Corinth, for the legend may have changed by the time of Euripides. The sanctuary of Hera Acraea in Corinth appears to have been on the roof of the spring Glauke (above, n. 241), but that is of little help for the date of the sanctuary, since the spring has no clear date (cf. Corinth i, pt. 6, 222); in any case, it is possible that the sanctuary existed before the construction of the spring.

²⁵⁶ The cemetery (ADelt xvii (1961/1962) B, 52 ff.Google Scholar; cf. also BCH lxxxvi (1962) 686 ff.Google Scholar) contained graves of the MG II period; cf. Coldstream 95 (Ay. Theodoroi). The pots that were buried in the graves can naturally provide no information about the political control of Crommyon at that time.

²⁵⁷ Hammond's argument (above, p. 199) that it was only in the late eighth century at the earliest that commercial and strategic considerations could have made control of the Perachora peninsula vital for Corinth can here be given a different application. The late eighth century certainly seems the earliest possible period for the Corinthians to have considered such matters; but since the Heraeum was Corinthian from c. 800 at the latest, the only available time for a Corinthian conquest of the peninsula is that of the foundation of the city.

²⁵⁸ Professor Huxley has drawn my attention to a passage in Strabo which, like Plut., Qu.Gr. 17Google Scholar, is concerned with a Megarian claim to territory outside the Megarid of the Classical period. At 394 Strabo discusses the Megarian claim to Salamis, and remarks that the Athenians, in pressing their claim to the island, inserted the line

after that which refers to Ajax in the Homeric Catalogue of Ships (Il. ii. 557Google Scholar); the Megarians, however, substituted the following couplet:

This Megarian interpolation was clearly primarily designed to lay claim to Salamis, by associating Ajax not only with Salamis itself but also with places in the historical Megarid; but it has also been argued that the lines included a Megarian claim to land which had been taken from them not by Athenians, but by Corinthians. Of the four places named in the interpolation, Polichne is impossible to identify, for there is no other reference to a place of this name in the area (cf. Kirsten, , RE xxi (1952) 1371 f.Google Scholar); Nisaia is clearly the Megarian port Nisaea, while Tripodes is to be identified with Tripodiskos (Strabo 394; cf. above, n. 212). The interest centres on Aigeiroussa, which has been placed by some scholars in the Perachora peninsula; if that is correct, the interpolation will have been partly intended as a Megarian claim to the territory Megara lost very early in her history to Corinth, and Aigeiroussa will have been on the same footing as the areas occupied by the Heraeis and the Piraeis. The argument, however, depends upon an improbable identification of Aigeiroussa. Steph. Byz. s.v. Αἰγείρουσσα explains that the place was in the Megarid, and refers to Strabo, doubtless meaning this passage (394); he also states that an alternative name was Αἴγειρος, and refers for this form of the name to Theopompus (F.Gr.Hist. 115 F 241). It is likely that a third reference to the place is to be found in Plutarch, , Qu.Gr. 59Google Scholar, where it is explained why the Megarians were called ‘wagon-rollers’: some travellers on the way from the Peloponnese to Delphi camped by a lake at a place called Αἴγειροι; the Megarians attacked them there and rolled them and their wagons into the lake. The names Aigeiroussa in the Megarian interpolation, Aigeiros in Theopompus, and Aigeiroi in Plutarch all probably refer to the same place, since they are all in or near the Megarid (cf. Hammond, , BSA xlix (1954) 120Google Scholar); but the only lake in this area that still exists today is Lake Vouliagmeni, on the south side of the Perachora peninsula (see FIG. 15). It has therefore been proposed that the lake in Plutarch's story should be identified with Vouliagmeni; it will then follow that Aigeiroussa, which Plutarch shows must have been close to the lake, was situated in the Perachora peninsula, and therefore in the part of the early Megarid that was lost to Corinth (cf. Curtius, , RhMus iv (1846) 200 ff.Google Scholar; Hirschfeld, , RE i (1893) 951Google ScholarHalliday, , Plutarch: Greek Questions 219 f.Google Scholar; Meyer, E., RE xv (1931) 169Google Scholar). There are, however, grave difficulties in this reasoning, for Lake Vouliagmeni can never have been on the road from the Peloponnese to Delphi, and Plutarch's pilgrims will therefore have been well out of their way if they camped beside that lake (cf. Hammond, op. cit. 120 n. 38); and although it is true that Vouliagmeni is the only lake in the area at present, it is more than likely that at least one other existed in antiquity, a little way north of east from Pagae at the modern Megalo Vathichori (Hammond, op. cit. 108, 118); we may therefore conclude that Aigeiroussa was to be found close by this lake, and that it was therefore in territory that Corinth never won from Megara (cf. Hammond, op. cit. 119 f.). It is perhaps worth noting that another mention of a lake in this area—the in Aeschylus, Ag. 302—is equally difficult to explain satisfactorily if it is identified with Vouliagmeni, as has sometimes been proposed (for a full discussion, see Fraenkel, , Aeschylus: Agamemnon ii. 160 f.Google Scholar); Hammond (op. cit. 118) has suggested that this lake too is to be identified with the one that probably existed in antiquity at Megalo Vathichori (for a different, but less persuasive, interpretation of the passage in Aeschylus, cf. Beattie, , CR iv (1954) 78 f.Google Scholar; Denniston, and Page, , Aeschylus: Agamemnon 95Google Scholar).

²⁵⁹ This view of course conflicts with the evidence of Thucydides (iv. 42. 2) that the Dorians who founded Corinth had to contend with Aeolians who were living on the site; the Piraeis, as Megarians, were presumably Dorian. But the evidence of Thucydides is suspect, since it is likely to be dependent on the fertile imagination of Eumelus, that falsifier of much early Corinthian history (on Eumelus, see Dunbabin, , JHS lxviii (1948) 66 ff.Google Scholar; Huxley, , Greek Epic Poetry 60 ff.Google Scholar).