^{page 82 note 1} Agamemnon, p. 512, from ‘Kolon und Satz’, NGG (1932), 202Google Scholar = Kl. Beitr. 1, 78Google Scholar.

^{page 83 note 1} Λ 550 makes rather against meaning ‘spoiled’, but if it meant ‘filched for himself’ we should certainly expect the middle here, as always with this sense. The subject need not be, as some have assumed, the of the previous fragment, out for his own gain; all that we can tell from the introductory remark of Aristotle and Plutarch is that he would have made a worse job of reforming the constitution than Solon did. He may then be the immoderate reformer who, unlike Solon, would have thrown the state into chaos by his drastic measures, and so deprived it of its best element. Cf. Plutarch's paraphrase of part of the Phocus poem in Vit. Sol. 15, with its ending: … needs little adjustment to fit the metre, and there is no reason to doubt that , matching in fr. 25. 8, is Solon's own word. Nor is an economic interpretation ruled out, cf. in fr. 23. 20 D: the immoderate reformer would have ruined the economy and deprived the state of its wealth. Arthur Platt (J.Phil. XXIV (1896), 256Google Scholar) condemned Bergk's order, despite his assertion (PLG ⁴ 11, 54Google Scholar) ‘solet passim Solo verborum traiectione satis licenter uti’. He points out that Bergk's other examples (frr. 1. 43–5 and 23. 5 D) are very much easier. And if Bergk is right about Solon, Solon's practice is not evidence for tragedy. See JHS XCVI (1976)Google Scholar.

^{page 83 note 2} H. Amman, in his study of word-order in Homer, (Untersuchungen zur homerisclien Wortfolge (1922), esp. pp. 43–5Google Scholar), observes that a phrase added after the verb often stands as a loosely detached complement, further defining the sense. His argument is carried too far, but it certainly holds for the examples quoted. Cf. Kirk, G. S., YCS xx (1966), 105 ffGoogle Scholar.

^{page 84 note 1} I have counted BABA as fully interlaced no less than ABAB; but since BAB is normal and ABA is not, BABA is perhaps less remarkable.

^{page 84 note 2} Many of the examples of hyperbaton given by Schöne, H. in ‘Umstrittene Wortstellung des Griechischen’ (Hermes L (1925), 144–73Google Scholar) involve parts of or (pp. 152, 156, 163-4).

^{page 84 note 3} His fondness for it is parodied by Aristophanes, , Ach. 449, 460, 465, 468Google Scholar; cf. 1137, 1140 (Lamachus).

^{page 84 note 4} I.e. separated from its object by the main verb, or part of the main clause; it is then generally at the end of the sentence, as in A. Suppl. 726 below. The so-called pleonastic also often ends the sentence, e.g. Ar. Nub. 1105Google Scholar, but as ex hypothesi it governs nothing which has not already a construction, it is never strictly in hyperbaton. Pleonastic occurs in Aeschylus at Ag. 1596 and PV 55, where it should not be altered to (‘take and smite them about his arms“, with νιν = as in S. El. 436, 624 and probably A. fr. 17. 14M, E. Andr. 19; for pleonastic coming first, cf. Bacch. 1064, D. 18. 118 al. ). Such examples as Ar. Av. 56Google Scholar, ‘You get a stone and strike (the rock) with it’ (cit. Jebb on S. OC 475; cf. Ar. Thesm. 212Google Scholar, fr. 269), might lead us to take Ag. 1127 the same way, ‘she takes a black-horned device and strikes him with it’; but , as Fraenkel shows (Ag. pp. 512 f.), must be taken with , which therefore has Agamemnon as its object.

^{page 85 note 1} In Lys. 1186 is in hyperbaton, but there is no interlacing of main and participial clause, since is common to both.

^{page 85 note 2} I agree with Johansen, Friis (General reflection in tragic rhesis, p. 87Google Scholar n. 111) against Kannicht ad loc. that the required ellipse is impossible, and that something is missing after 713, viz. ‘will some day have luck’; and with both against Dale, that the topos in play is simply ‘the future is unpredictable’, rather than ‘success is the reward of toil’, which not only needs a supplement but makes the whole train of thought more complex than it need be.

^{page 86 note 1} The article in is idiomatic, cf. Ar. Eccl. 781Google Scholar and probably Pax 1334, and , of the dead, in Ran. 1462, fr. 488. 14 (possibly = , but it was not quite these benefits that Menelaus strove for). ‘Faring well’ (Pearson), ‘having fared well’ (Paley), do not do justice to the aspect of

^{page 87 note 1} The line is still puzzling, however, since here really means ‘give me up, abandon me’, the standard sense of the active, rather than ‘let go of’, the middle sense. Perhaps ἐμέ is governed by both verbs after all; cf. McDowell on Ar. Vesp. 416Google Scholar.

^{page 87 note 2} These examples, with Phil. 771–2, Med. 473–4, Rhes. 790–1 above and Callim. fr. 75. 10–11 below, I owe to Dr James Diggle.

^{page 88 note 1} Cf. the postponement of γάρ, which is frequent in New Comedy and presumably colloquial (see Wilamowitz, , Schiedsgericht, p. 156Google Scholar; Handley on Dysc. 66–8), while in serious poetry it is a stylistic device for special effect, e.g. PV 29 .

^{page 88 note 2} In fact only IT l.c. gives ABAB rather than BABA; cf. p. 84 n. 1 above.

^{page 88 note 3} would suit either weapon: ‘black-horned’ because fatal and sharp-pointed (or possibly ‘ horned’, of an axe's shape). For the sinister tone of black, cf. S. Aj. 231Google ScholarTrach. 856 , and E. HF 780 , with Wilamowitz's note.

^{page 89 note 1} In Suppl. 462, the king asks the leader of the Danaids: ; ‘What are those girdle-contrivances for?’ – not ‘implements’. At ibid. 956 as Tucker points out (cl. Pers. 142 ), ‘denotes the shrewdness of construction which makes them impregnable’. In Pers. 722 it is applied (plur.) to Xerxes' bridge of boats over the Hellespont, and so probably at ibid. 113 (but my argument holds even if this phrase refers to the Persian fleet, as Coxon among others supposes (CQ viii (1958), 45–6Google Scholar)). In fr. 139 N (231M) the word is in fact applied to a weapon: ‘So is the Libyan fable famed abroad: the eagle, pierced by the bow-sped shaft, looked at the feathered contrivance and said, “Thus, not by others, but by our own plumage are we slain”.‘ Here it would make sense to translate as ‘implement’ or ‘weapon’; but the word also suggests the cleverness of the device which made the eagle's feathers his own undoing. Finally the phrase (Sept. 132)suggests that Poseidon's trident is an ingenious gadget cleverly contrived for its purpose, viz. catching fish; though there too the translation ‘implement’ would make sense.

^{page 89 note 2} Fraenkel, , Agamemnon, p. 511Google Scholar.

^{page 90 note 1} Riv. di Fil. Cl. CI (1973), 41fGoogle Scholar.

^{page 90 note 2} S. Rizzo, art. cit., cl. Smith, O., Class, et Med. XXVI (1965), 10–72Google Scholar (esp. 12). Much the same point was made by Hermann: ‘quoniam tauri et vaccae appellatione usa erat, transfert cornua ad id de quo proprie dici non potuerunt’.

^{page 90 note 3} Hades king of the underworld and Thanatos fetcher of the dead are of course two distinct persons; in Hesiod Thanatos and Hypnos are the children of Night, and in paintings they generally appear together, sometimes as boys. But in the literature of the classical period no sharp distinctions can be drawn between the two functions (see Heinemann, K., Thanatos (1913), pp. 29 ff.Google Scholar; Lesky, A., Alkestis, der Mythus und das Drama (Sb. Wien CCIII. 2(1925), 62–5Google Scholar, and art. ‘Thanatos’ in RE, esp. col. 1255–7). Thanatos is not a poetical invention, as Robert and Wilamowitz held, but an original figure of folktale, like Charon (Heinemann, op. cit. 50 ff.; Lesky, loc. cit.); though the representations in art are mostly dominated by the poetic conception (but see p. 92 n. 1 below).

^{page 90 note 4} The Homeric is hardly personified, though it doubtless lies behind the other epithets, and behind the dark skin of Thanatos (or Hypnos, cf. Hes., Th. 256Google Scholar) in some of the paintings.

^{page 91 note 1} For Hades, v. Roscher, 1, 1796; Farnell, , Cults of the Greek states, I, 105Google Scholar; the hair sometimes white, ibid. III, 286, Pl. xxxii; for Thanatos, cf. the lecythoi illustrated by Heinemann, op. cit. Pl. 7, 8, 9. (See also p. 92 n. 1 below.)

^{page 91 note 2} Usener, K., Götternamen (1895 (1948 ² )), p. 221Google Scholar.

^{page 91 note 3} So LSJ; fanciful interpretations in the scholia.

^{page 91 note 4} See H. Frisk, Etymologisches Wörterbuch, s.v. .

^{page 91 note 5} (i) , attributed by Σ Townl. to Aristotle (perhaps wrongly, Lehrs, v., Aristarchi Studia ³, p. 42 n.Google Scholar). (ii) , attributed by Eustathius to Aristotle (851, 52). (iii) , a method of dressing the hair in horn-shaped spirals, like the at Athens (cf. Th. 1. 6, Σ Ar. Nub. 980Google Scholar, Pollux 2. 30–1). This is the explanation favoured by the ancient commentators and by Plutarch, , de Soil. Anim. 976 FGoogle Scholar, who quotes Archilochus (fr. 59 D = 117 W), (cf. Σ Townl. ad ω 81). (iv) a view known to Plutarch and the commentators, but rejected by most of them, probably because Aristarchus denied that Homer so used the word (Apoll. Lex. s.v. ). Herodian, however, seems to have allowed the use in Homer (2. 75, 23–4 Lentz, cf.Σ Townl., Σ d, e¹, e² Erbse ad Λ 385), and Pollux (loc. cit.) keeps an open mind: (cf. ω 81, μ 253). See Pearson on S. fr. 875, Leaf on Λ 385, cl. Helbig, W., Das Homerische Epos aus den Denkmälern erläutert ² (1887), pp. 241 ff.Google Scholar; Bremer, W., Die Haartracht des Mannes in archaischgriechischer Zeit (1911), pp. 49 f.Google Scholar, and art. ‘Haartracht’ in RE; Marinatos, S., ‘Haartracht’, Archaeologia Homerica 1 (1967), 1312–13Google Scholar. On the meaning of see Hauser, ,Jahres. Oest. Arch. Inst. XI (1910), Beibl. 87Google Scholar; Fink, J., Die Haartrachten der Griechen (1935), pp. 44 ff.Google Scholar, cf. ibid. 12 n. 12. Several types of horn-like coiffure are illustrated on vases.

^{page 92 note 1} There may also be a further point in speaking of horns in this context; for on two lecythoi in the British Museum Thanatos has unkempt, spiky hair, giving him a demonic appearance which perhaps better represents the popular conception than do most of the paintings (ARV² 851.272, 1228. 12; best pictures in Buschor, , Grab eines attischen Mädchens, p. 5Google Scholar; Robertson, C. M., Greek painting (1959), p. 150Google Scholar). This demonic appearance is found also in Boreas, who is sometimes shown with his hair standing on end (e.g. ARV² 496 [Oreithyia painter] = FR ii, Abb. 68; cf. Robertson, op. cit. p. 148). It may have influenced the Etruscan type of Charun with hair erect in the Orcus tomb (Weege, F., Etruskische Malerei (1921), pl. 60 = Abb. 53 on p. 50Google Scholar; Charun begins to figure prominently on Etruscan vases at much the same time, cf. Beazley, J. D., Etruscan vase-painting (1947), pp. 8 f.Google Scholar).

^{page 92 note 2} Other kenning-prone objects are men or heroes, animals and important things (especially in ritual); see Schadewaldt, , Monolog u. Selbstgespräch, p. 59Google Scholar, and in general Ingrid Waern, , Kennings in pre-Christian Greek poetry (1950)Google Scholar. The link between epithets of gods used as names and such kennings as (sea) is noticed by Wilamowitz, , GV, p. 409 n.Google Scholar, cf. Pindaros, p. 441.

^{page 92 note 3} Cf. Roscher, 1, 1782 fin., Farnell, , Cults, III, 218Google Scholar. itself a kenning, ‘the gatherer’, by etymology; v. Frisk, s.v.

^{page 92 note 4} The text is not certain, as the paradosis has , which does not make sense or correspond, having more syllables than the strophe; but it is much more likely that (or ) is intrusive, as names often are, than that the strophe, impeccable in sense and metre, is defective.

^{page 93 note 1} Kapsomenos, S. G., Sophokles Trachinierinnen und ihr Vorbild (1963), esp. pp. 86 ffGoogle Scholar.

^{page 93 note 2} Wakefield's interpretation. means a bird-net at Ar. Av. 194, 528Google Scholar, Callim., Aet. fr. 75. 37Google Scholar, and might appear somewhat less suitable than Aeschylus' trawl. But it is strongly supported by Trach. 1052 , which corresponds exactly to Ag. 1580Google Scholar.

^{page 93 note 3} Cf. Ag. 1580, Trach. 1052, quoted in n. 2. The formal counterpart of Clytemnestra in Sophocles is of course the innocent Deianeira, though she may not have been innocent in the earliest version of the legend (cf. Plut, . Mor. 811Google Scholar D, and see Schwinge, E.-R., Die Stellung der Trachinierinnen im Werk des Sophokles (1962), p. 25 n. 3CrossRefGoogle Scholar).

^{page 93 note 4} A comparable though slightly different example is Ag. 239 refers to the saffron robe of Iphigeneia; it cannot mean her blood, as some have thought, cl. 1121 . But neither is the echo in 1121 a meaningless accident. suggests the blood soon to be shed, and this gives the phrase an added dimension. (Cf. Lebeck, Anne, The Oresteia (1971), pp. 85, 191Google Scholar; for the interpretation of the line see CQ xxv (1975)Google Scholar-).