2. Hymn 13 consists of 3 lines based on the Hymn to Demeter; 17 is 5 lines most of which occur, differently arranged, in Hymn 33; 25 is a scrap based on Hes. Theog. 94—97. That the short versions are secondary is clear from Hymn 25.4, where an allusion to kings has been inappropriately imported from Hesiod, and from 18.5–9, the expansiveness of which is only suitable as introduction to a longer narration.

3. Cf. for earlier discussions Lenz, pp. 278–86, and now especially Càssola, pp. xvii–xxi. It is a priori highly unlikely that the only such instances of cutting down that occurred are the four we happen to be able to detect. I regard as almost certainly abbreviated (in addition to those already mentioned) the following very short Hymns (of fewer than 15 and usually than 10 lines): 9 (where the details of Artemis’ activities in 3–6 are too specific for a generalizing proem), 12 (which lacks an ending), 21 (which lacks a beginning), 24 (where the detail of line 3 is ridiculously over-emphasized), and 26 (where a developing narrative breaks off pointlessly at 10); these should doubtless take with them 10, 11, 14, 15, 16, 20, 22, and 23, which are similarly tiny. 27–33 (which average 15–20 lines) are harder to judge. Cassola considers them incomplete because mythless (and certainly the main section of 32 ends abruptly at 16); but it is not proven that a Homeric Hymn had to contain a narrative, and about Hestia, for instance (subject of 29), there were virtually (but cf. Hymn 5.24) no stories to tell. Hymn 19, of 48 lines, is very probably complete (despite Cassola's reservations, p. xix), as 7, of 59 lines, certainly is.

4. Janko, R., for instance, Homer, Hesiod and the Hymns (Cambridge, 1982), pp. 140–2, 228–31Google Scholar, ranges the major Hymns from the first half of the seventh century (Delian Apollo) to the end of the sixth (Hermes; but it is unnecessary to see any influence of proto-rhetoric on Hermes, as Dr Doreen Innes kindly advises me); on the lesser Hymns, cf. ibid. p. 276 n. 27. Contrast Cassola, p. lviii: ‘gl'inni rimangono estranei alia dimensione temporale.’ Proems were still sung before epic recitations (whether or not they were still composed) in the fifth century (Pind. Nem.2.1–3); the practice of recitation continued much longer (Càssola, pp. lxi–lxii). As for origins, the genre of the hexameter hymn evidently in some form long antedates the first surviving examples (cf. Horn. Od. 8.499; the evidence of Hesiod, n. 5 below; and the suggestion of West, M. L., Glotta 67 (1989), 135–8Google Scholar, that the Indo-European ‘injunctive’ survived within this genre).

5. Cf. Richardson, pp. 3–4; Càssola, pp. xii–xvi. For hymns introducing epic recitations, cf. Horn., Od. 8Google Scholar.499 (?); Pind, . Nem. 2Google Scholar.1–3; Hymn 3.158–61 (?); 31.18–19, 32.18–20 (the latter two conceivably post-classical); on the other hand, it is clear from Hesiod (see West's notes on Op. 1–10; Theog. 1–115) that hexameter poems of every kind required hymnic proems. The formula μεταβήσομαɩ ἄλλον ἐς ὕμνον in 5.293 and elsewhere reveals nothing about the character of the following song: on the sense of ὕμνος see West on Hes. Op. 657Google Scholar.

6. So e.g. Allen, T. W., Halliday, W. R., Sikes, E. E., The Homeric Hymns (Oxford, 1936), pp. xciv–xcvGoogle Scholar; Wünsch, R., RE s.v. Hymnos, 149, 151Google Scholar. Contrast Richardson and Cássola, as cited in n. 5; Lenz, pp. 278–86.

7. So e.g. Lenz, pp. 10, 17; Herter, H. in Brillante, C., Cantilena, M., Pavese, C. O. (edd.), I poemi epici rapsodici non omerici e la tradizione orale (Padova, 1981), p. 196Google Scholar: unlike epic, the Hymns‘non erano utilizzabili in qualsiasi maniera, ma destinati per determinati occasioni in certe feste, di cui glorificavano le divinità’. That the Hymns were in fact re-utilized is fairly clear from their textual history (cf. Janko [n. 4], pp. 2–3), although Herter might counter that subsequent performances were confined to appropriate festivals.

8. Clay, p. 7, probably goes too far in postulating the symposium/feast as context of performance in the archaic period: contrast Hymn6.19, δòς δ'ἐν ἀуῶνι/νíκην τῷδε, and the Hymn to Delian Apollo (which she seeks, pp. 46–52, to dissociate from a Delian festival). But she may well be right that Hymns, like epic, had once been performed at banquets (cf. Horn. Od. 8.499, and Demodocus’ song of Ares and Aphrodite), and thus that the genre was not occasional in origin. (Hymn 24.4, exceptionally, summons Hestia τóνδ’ ἀνὰ οἶκο — the character of the οἶκς is uncertain.)

9. See the admirable monograph of Lenz, esp. pp. 9–22, from which I borrow the example of epiphanies.

10. See Rudhardt's, J. important study, ‘À propos de l'hymne homèrique à Déméter’, Mus. Helv. 35 (1978), 1–17Google Scholar, also in his Du mythe, de la religion grecque et de la compréhension d'autrui (Geneva, 1981), pp. 227–44Google Scholar; cf. Clay, passim. To a large extent Homeric Hymns stand to Theogonies as do monographs to works of synthesis; for the similarity of the two genres, cf. Hymn 4.57–61 with 427–33; 19.27–8; and the hymn (1–115) which introduces Hesiod's Theogony.

11. Cf. Lenz, pp. 20–21, who refers to 1.10–16; 2.460–9, 483–6; 3.2–13, 186–206; 4.319ff.; 6.14—18; 28. As he observes, these good-humoured and harmonious gatherings are very different from the fraught Olympic scenes often found in Homer.

12. For criticism see Lenz, passim, e.g. pp. 273–4; and especially Smith, P. M., HSCP 85 (1981), 17–58Google Scholar, who shows that even the strongest external evidence for historical Aeneadae (Demetrius of Skepsis ap. Strabo 13.1.52–53) is inconclusive; the Homeric passage ll. 20.300–8 proves only that Aeneadae had an important mythological existence.

13. See Smith's, attractive study, Nursling of Mortality: a Study of the Homeric Hymn to Aphrodite (Frankfurt, Bern, Cirencester, 1981)Google Scholar; also e.g. Segal, C., CW 67 (1973/1974), 205–12Google Scholar; King, H., Arethusa 19 (1986), 15–36Google Scholar; and for qualifications Clay, pp. 186–8. For Smith, p. 5, ‘The underlying concern… is the limitation in time of mortal life’; for King, 17–18, this is only part of the true central theme, ‘the separation of gods from mortal men'. These themes are certainly present; but it seems questionable to declare them central to a Hymn to Aphrodite. Clay argues, p. 166, that the poem shows the last sexual contact between mortals and immortals, and thus the twilight of the heroic age; but this is scarcely in the text (for Zeus’ stated aims see 45–52; 247–54). The post-structuralist reading by Bergren, A. L. T., Classical Antiquity 8 (1989), 1–41CrossRefGoogle Scholar, reaches a strange conclusion, p. 41: ‘the Hymn is unequivocal in this element of its aetiology, its proclamation that any virgin bride might be Aphrodite in disguise, and that no man “seized by eros” for the bride can possibly know the difference.’

14. See Kirk, G. S., Myth (Berkeley, 1970), pp. 240–1Google Scholar; cf. e.g. The Encyclopaedia Britannica, Macropaedia, 1988 ed., s.v. ‘Myth and Mythology’, 710: ‘there is no attempt to justify mythic narratives or render them plausible.’

15. In the fifth century, ‘theogonic’ myths appear, more briefly treated, in two Euripidean choruses, just the ones that have long been suspected of being embolima (Hel. 1301–68; I.T. 1234–83); subsequently they have a precarious survival in lyric hymns (a genre where they must always have been found) such as PMG 934–6, and of course reappear when Callimachus revives the hexameter hymn.

16. See e.g. Kearns, E. D., CR 39 (1989), 61–62Google Scholar.

17. Cf. Wehrli, F., ARW 31 (1934), 77–104Google Scholar; Deichgräber, K., ‘Eleusinische Frommigkeit und homerische Vorstellungswelt im homerischen Demeterhymnos’, Abh. Ak. Wiss. Mainz, 1950, 6, pp. 501–37Google Scholar; Richardson, pp. 74–86.

18. On the near eastern parallels to the motif of Demeter's wrath, and its disastrous consequences, see e.g. Richardson, pp. 258–9; Burkert, W., Structure and History in Greek Mythology and Ritual (Berkeley, 1979)Google Scholar, ch. 6.

19. So Wehrli (n. 17), 80; Deichgräber (n. 17), p. 529: the Hymn-writer ‘homerisierte nicht nur als Techniker des Wortes und des Verses’.

20. See Richardson, pp. 74–86. On Ovid's Greek sources see Montanari, F., ASNP ser. iii. 4 (1974), 109–37Google Scholar; Hinds, S., The Metamorphosis of Persephone (Cambridge, 1987), pp. 51–57Google Scholar.

21. See Richardson, pp. 195–6. Eumolpus is little less mutable.

22. From Clinton, Kevin, ‘The Author of the Homeric Hymn to Demeter’, Opuscula Atheniensia 16 (1986), 43–49Google Scholar, who concludes ‘the author was not from Attica and he was not writing for an Athenian audience. The story of the Rape was no doubt a standard one, sung all over the Greek world. Our author gave it a partly Eleusinian setting; he may have been an initiate, but he was not deeply interested in Eleusis.’ Clinton's expert knowledge of Eleusinian antiquities gives that conclusion great weight. But his argument that the poet was ill-informed about details of Eleusinian cult and organization is much weakened by the chronological gap, probably of at least a century, between the Hymn and the other, chiefly inscriptional, evidence to which he appeals; and the instances of good local knowledge which he acknowledges weigh against him more heavily than he allows. The text presents a counter-case in general and in part (on the site of the rape, and the role of Triptolemus) in specific terms. As for further specific points: Attic poets (e.g. Eur., Suppl. 271Google Scholar; Arist., Ran. 337, 671Google Scholar) and pot-painters (cf. Schwarz, G., Triptolemos [Grazer Beiträge Supp. II, 1987] nn. 58, 61 on p. 39, cf. p. 102Google Scholar) did not treat the name ‘Persephone’ as taboo in an Eleusinian context; Hecate is certainly prominent in Eleusinian iconography (Schwarz, p. 253, index s.v. Hekate), even if cult of her happens not to be attested.

23. Zeus’ involvement is traditional, unless Hes., Theog. 913f.Google Scholar, where the motif also appears, is post-Hesiodic and itself based on the Hymn. It is standard in later accounts.

24. Cf. Lenz, pp. 58–69. Zeus is not, by contrast, explicitly said to have willed the events subsequent to the rape, which led to the final compromise whereby Kore divided her time between upper and lower worlds; but he did approve that compromise (445–7).

25. So, tentatively, Richardson, p. 149.

26. Cf. Richardson, p. 150. For Attic locations see Phanodemus, , FGrH 325 F 27Google Scholar (Attica); Paus. 1.38.5 (Eleusis); Σ Soph. O.C. 1590, 1592 (Colonus). Orpheus: frs. 50–52 Kern, esp. fr. 51 (= Paus. 1.14.3): an Eleusinian setting for the Descent (if not necessarily – cf. below – for the rape itself) is implied by the ‘information’ motif, and by the story of Eubouleus’ pigs. The ‘information’ motif is also found in an Eleusinian context in Tzetzes ad Hes., Op. 32Google Scholar, Σ Vet. Tr. ad Ar. Eq.698, Σ Ael. Arist., Panath. p. 53Google Scholar Dindorf, but Celeus (with whom in Σ Ael. Arist. Demeter co-habits) is here in contrast to the Orphic version father of Triptolemus, and the rape itself is set in Sicily – presumably an air flight must have intervened between rape and descent as in Hymn Orph. 18.12ff. and in the Homeric Hymn (cf. Richardson, p. 159).

27. See Richardson, pp. 162, 165–6; for the iconographical type see e.g. the illustrations to Kerényi, K., Die Mysterien von Eleusis (Zurich, 1962)Google Scholar, passim. The exact extent of aetiological allusions in the Hymn cannot be definitely established (for a review of possibilities cf. Clay, pp. 203–4); this one is certain, as Ovid confirms: hinc Cereris sacris nunc quoquetaeda datur, Fasti 4.494.

28. See Richardson, p. 155, and above n. 22.

29. See Richardson, pp. 211–7.

30. It is sometimes suggested that she adopts Demophon as a substitute for lost Persephone (so e.g. Arthur, M., Arethusa 10 (1977), 22Google Scholar [in a Freudian/feminist reading]; Deal, H. and Rubin, N., QUCC 34 (1980), 8)Google Scholar; Clay proposes, p. 226, that the attempt to confound the line of demarcation between man and god is a first, unsuccessful act of defiance against Zeus. But it is wrong to try to peer into the goddess’ mind: cf. p. 11 above.

31. See Richardson, pp. 81 and 174 (note on 75ff.).

32. Cf. Bremmer, J. N., ‘The old women of ancient Greece’ in Blok, J. and Mason, P. (edd.), Sexual Asymmetry. Studies in Ancient Society (Amsterdam, 1987), pp. 193–215Google Scholar; Silk, M. S., BICS 34 (1987)Google Scholar (= Gredley, B. (ed.), Studies in Greek Drama), 93Google Scholar: ‘for Aristophanes, old women do not attract pathos; old men do.’

33. Paneg. 28.

34. ‘Threptos': Sokolowski, F., Lois sacrees des cites grecques. Supplement (1962)Google Scholar, no. 10.69— 70 (cf. Körte, A., Glotta 25 (1936), 137–9Google Scholar, who identifies him with Triptolemus). Triptolemus: Ov. Fasti 4.550–60; Hyg. Fab. 147 etc. (cf. H. J. Rose ad loc; for the view that the Hyginus tradition derives from Panyassis, see Robertson, N., Hermes 108 (1980), 278 n. 10, with his references)Google Scholar; Σ Nic. Ther. 484c; 5th–c. iconography strongly implies that Triptolemus was already the nursling (cf. Körte, op. cit.; Herter, H., Rh. Mus. 90 (1941), 266Google Scholar; Robertson, M. in Greek Vases in thej. Paul Getty Museum 3, 1986, 86–88Google Scholar; Schwarz, , Triptolemos [n. 22], p. 247)Google Scholar. I know no exactly parallel myths outside Attica, although at Sicyon Demeter nursed the child / Orthopolis (Paus. 2.5.8; and cf. Richardson, pp. 234–8, on Erichthonius and Apollonius Rhodius’ Achilles). The argument does not in fact require that the ‘failed immortalization’ motif be exclusive to Eleusis, but only that it had already been adopted there in the seventh century. Even this, of course, is not strictly demonstrable, given the date of our evidence; if it was in fact taken up later, the argument about the episode's function within the poem can still stand, while that about the relation of poet and cult is reversed (with effects still, though differently, damaging for Clinton's thesis [n. 22 above]: the poet adopts, indeed, a non- Eleusinian motif, but then succeeds, it seems, in causing it to be accepted at Eleusis).

35. 1050—2, Πóτνɩαɩ σεμνὰ τɩθηνοũνταɩ τέλη/θνατοĩσɩν.

36. Cf. Richardson, pp. 231–6.

37. Cf. Arist., Eth. Nic. 1159a 5–7Google Scholar. Ovid's Demeter nicely brings out Metanira's innocence of intention: cui dea ‘dum non es’ dixit ‘scelerata, fuisti’ (Fasti4.557). Cf. Thalmann, W. G., Conventions of Form and Thought in Early Greek Epic (Baltimore, 1984), pp. 92–94Google Scholar.

38. Cf. Rudhardt, J., Mus. Helv. 35 (1978), 11Google Scholar; Gasparro, G. Sfameni, Misteri e culti mistic di Demetra (Rome, 1986), p. 194 (cf. for other views pp. 67–77)Google Scholar; Clay, p. 244 (with further references). The tragic aspect of the story is stressed still more in the variant in which the child immediately dies (Apollod. 1.5.1, ? Orph. fr. 49.100–1). When Triptolemus becomes the nursling, a new balance appears, potentially at least (Sfameni Gasparro, p. 73): in place of immortality, he receives corn, the characteristic food of mortals (Ov. Fasti 4.559–60).

39. See e.g. Ovid, , Fasti 4.559–60Google Scholar; Hyg., Fab. 147Google Scholar.

40. Cf. Richardson, p. 236; but I am much less ready than him to transfer this good cheer in any simple way into the Hymn — though doubtless by an indirect route Demeter's fine promise of 227–30 continues to assert her credentials as a divine ‘nurse’, despite the sad outcome in Demophon's case.

41. Cf. Càssola, pp. 33f. (whence the quotation); Wilamowitz, , Der Glaube der Hellenen, ii. 50 (in the pagination of the edition Darmstadt, , 1959)Google Scholar. Richardson's discussion, pp. 258–60, is more cautious.

42. Cf. n. 18 above.

43. Cf. n. 30 above.

44. Richardson is surely right to see an eschatological reference in these lines (cf. Burkert, W., Gnomon 49 (1977), £445–6)Google Scholar, despite the doubts of Clay, p. 252.

45. On the ‘theogonic’ aspect of this Hymn cf. n. 10 above, and Alderink, L. J., Numen 29 (1982), 1–16CrossRefGoogle Scholar. On the relation of the Demophon incident to Kore's return see the (rather overschematic) paper of Rubin, N. F., Deal, H. M., QUCC 5 (1980), 7–21Google Scholar.

46. Cf. Schwarz (n. 22), passim. The Triptolemus myth first appears on vases around 540, when Eleusinian themes in general first enter vase-painting (cf. Boardman, J., JHS 95 (1975), 7)CrossRefGoogle Scholar. We can only speculate whether it already existed at the date of the Hymn's composition. The stories of the rest of Greece tend to tell of Demeter founding rites, not making a gift of corn; and the myth of Triptolemus may seem to imply an untraditional view of the gradual growth of human culture. Thus it is perhaps a secondary, Attic development. But it can be objected (cf. Richardson, p. 259) that particular myths about the coming of particular gods or skills do not imply a general theory of progress: the gift of corn is as legitimate a subject for traditional myth as is the gift of wine (with which it is in fact sometimes paired on vases: Robertson, M. in Greek Vases in thej. Paul Getty Museum 3, 1986, 71–90Google Scholar; Schwarz, p. 112) or the theft of fire.

47. Cf. Jameson's, M. H. review of Richardson, , Athenaeum 54 (1976), 441–6Google Scholar. He reminds us how loosely associated are myth, rite, and agricultural reality: for in Attica the seed germinates in autumn and thus, in contrast to Kore, is not hidden underground all winter.

48. Richardson, p. 301.

49. Hipp, . Ref. 5Google Scholar.8.39: cf., e.g., Burkert, W., Homo Necans(Berkeley, 1983), p. 290Google Scholar.

50. The distinction between the two grades is already clearly drawn in our first substantial epigraphic text concerning the Mysteries, IG I³ 6 (b) (of c. 460). Further complications (over the ‘Lesser’ Mysteries, and the nature of myesis) need not concern us here.

50. Cf. Burkert, , Ancient Mystery Cults (Cambridge, Mass., 1987) ch. 4, on, particularly, Plutarch fr. 168 SandbachGoogle Scholar.

51. Cf. Deichgräber (n. 17), p. 522.