Davies, Malcolm, Sophocles, Trachiniae. Introduction and Commentary. Oxford: Clarendon Press, 1991. Pp. xl + 296. ISBN 0-19-814899-2.
Reviewed by Robert L. Fowler, University of Waterloo.
Malcolm Davies' astonishing production continues apace. In the same year as the first volume of Poetarum Melicorum Graecorum Fragmenta appeared, and only three years after his edition of the epic fragments, we have this introduction and commentary on the Trachiniae. As explained in the preface, the book grew out of a seminar jointly given with T.C.W. Stinton, whose death in 1985 prevented his collaboration in the final product; but although Stinton's contribution must have been large in the beginning, the final work may be regarded in all essentials as Davies' (who frequently disagrees with Stinton's published notes). The text is that of Lloyd-Jones' and Wilson's new Oxford Classical Text. The introduction is especially to be commended for its discussion of the previous state of the myth; Davies' resources here are unequalled. The commentary is learned, acute, judicious, and subtle. Its special strengths are, first, a grasp of the secondary literature such as few scholars can boast (the most impressive citation was perhaps "for phrases like AI)SXRO\N KAI\ LE/GEIN / LO/GW| cf. Vahlen's edition of Aristotle's Poetics, 3rd edn. [Berlin, 1885], Mantissa Adnotationis Grammaticae, p. 212"). Secondly, an admirable sensitivity to nuance and style. Reflexes of idiom, figures of speech, rhetorical structures, niceties of tragic diction and imagery are constantly picked out for illustration and analysis. Thirdly, as expected, textual criticism; Davies' philology is faultless and his judgment sane. Although Lloyd-Jones twice read drafts of the commentary, and Davies was privy to the new OCT and Sophoclea in advance of publication (one can imagine many spirited exchanges), he has retained his independence from his formidable Oxford colleagues; I count at least 39 places where he disagrees with the text of the OCT, and more where he has not committed himself strongly to one version or another. He makes no conjectures on the text; but this is good. At this stage of the study of Sophocles one's suspicion should increase in direct proportion to the number of conjectures an editor makes.
No commentary has everything. In his review of P.E. Easterling's commentary (CR 34  7-9; correct the reference on p. vi), Davies found her wanting in textual criticism, but strong in literary appreciation; her articles are important for the latter, and he cites them often and approvingly. Davies cannot be faulted for his Philologie, but I found his attention to literary matters slight compared with Easterling. He is not insensitive to these, and I think he has struck the right balance on pp. xix-xx between the play's many subordinate themes and its main theme; but he is not inclined to discuss them overmuch. (An exception is formed by the well-balanced treatment of the play's ending at pp. xx-xxii. For an important addendum which may clinch a reference to apotheosis, see below on 1199.) Throughout the commentary, passing reference is made to matters of thematic importance as they crop up in the text; but his attention is mainly engaged elsewhere. A choral ode might receive one sentence, if that, on its role in the play, but a page on the metre. It may be that Davies was content to leave the job to Easterling, but in other matters he has not behaved as if he were an addendum to her, and indeed that would be a strange procedure for a commentator to adopt without warning. Perhaps the preface means to warn us in this way -- it speaks of "much in the way of both commission and omission that demands explanation" and "room for another commentary, one which would (for instance) devote a larger space to consideration of textual problems" -- but a clearer statement is possible. The intended audience of Easterling's commentary in Davies' opinion was the proverbial sixth-formers and University students; he explains that although his commentary originated in work with the second category (mostly graduate students, I should think), he hopes to interest colleagues as well (who in Oxford seminars talk mostly about textual problems). Whether deliberate or not, the marked preponderance of textual over literary matters will impede the usefulness of the commentary for some users. Davies' brief remarks on interpretation are nonetheless worth consulting; his references to further discussions are made with discernment. In sum, this book is an important contribution to the understanding of Greek literature, which every classical scholar should own.
Now for the details. P. xviii n. 4. The play is "undatable"; but if, as Jennifer March (following others) has argued with some plausibility (The Creative Poet. Studies on the Treatment of Myths in Greek Poetry [BICS Suppl. 49, 1987] 62f.), Bacchylides 16 copies Sophocles, we have a terminus ante quem (Bacchylides' latest datable poem is from 452, and there is some reason for thinking he died shortly thereafter). Davies (xxxii) in my view underestimates the force of her arguments. xxv. "Heracles' marriage to Deianeira is brought forward in time and made to precede the murder of Iphitus and the consequent servitude so that Heracles' long absence from home creates a great deal of suspense." Firstly, if that were Sophocles' only aim, one may ask why he has made the marriage precede not only the death of Iphitus, but most of the labours (27ff.). Secondly, the innovation is not Sophocles'. Pherecydes (FGrHist 3F82), who is earlier, treated the story. According to him, Herakles asked for Iole not as a bride for himself, but for his son Hyllus, child of Deianeira and presumably of marriageable age. Jebb (p. xxv of his commentary) already made the point. Sophocles was happy enough with this chronology, for in addition to the suspense mentioned by Davies, it allows Hyllus to be old enough to fulfil his crucial role in the play as a mediator between the two halves. It allows the portrait of Deianeira as a long-suffering wife, and creates the desired gap in age between her and the chorus (see below). Perhaps Pherecydes got this novelty from Peisander; it was not Panyassis (see fr. 17 Davies). xxviii n. 26. "Friedländer ... observes that the whole tradition of Iphitus' murder contains an odd contradiction: if, as Apollod. 2.6.1 has it, Iphitus alone of the Eurytids spoke out for giving Iole to Heracles, why should he be treated so villainously? If, on the other hand, he was as ill-disposed to Heracles as his brothers and father, why should he lodge with him in this foolhardy manner?" (sc. at Tiryns, when looking for his lost horses; see Trach. 270ff.). Apollodorus is sensible of the difficulty and so has Herakles go mad to kill him. Embarrassment is pronounced in fr. 82b of Pherecydes: "Although Polyidos the prophet told him not to go to Tiryns on his search, since it would not be profitable, it is said that he ignored him and went anyway" -- pigheadedness seems to be the only motive. The tale displays a need to have Iphitus killed at Tiryns regardless of narrative plausibility; in such cases one smells an aition. It is useless to ask how a figure of Tirynthian cult comports with an Oichalian homonym, or to analyse the myth along such realistic lines. Similarly one would not complain that Demeter in the Homeric Hymn unrealistically forgets her absent daughter throughout the episode at Celeus' house. Apollodorus' notice that Iphitus alone of the family supported Herakles' request looks like an innovation of someone trying to explain this very difficulty, why Iphitus would have expected friendly lodgings with Herakles; it may be left out of any account. Ll. 16f. On the motif "may I die before I see this" implied by Deianeira here see G. Vagnone, "Aspetti formulari in Stesicoro, Pap. Lille 76 abc: il desiderio di morte," QUCC 12 (1982) 35-42, and the same scholar's article in QUCC 30 (1988) 25-40. 25 (cf. 465). For misfortune brought on by beauty Easterling quotes Eur. Hel. 27, but the lines really calling for quotation are Hel. 304-5. 57-8. Roscher's conjecture (KAKW=S for KALW=S) deserves a mention. The difficulty is plain in Easterling's comment, apparently made without knowledge of the conjecture: "The Nurse does not mention the possibility that Heracles may be faring badly"; this possibility is what Deianeira has just said she fears, and what the nurse's suggestion is meant to address. 117. With Davies I feel little reluctance to adopt Reiske's change of TRE/FEI to STRE/FEI. The contortions of those trying to defend the paradosis are sufficient indication that it is indefensible. Stinton, JHS 96 (1976) 129 f. = Collected Papers on Greek Tragedy (Oxford, 1990) 209ff. showed clearly that a contrast is implied between AU)/CEI and what precedes; see then what Burton writes (The Chorus in Sophocles' Tragedies [Oxford 1980] 47): "It may perhaps be possible to force such a contrast by stressing the physical sense of AU)/CEI, 'raises him up', and taking TRE/FEI of a less turbulent motion, 'supports'. We may then imagine ourselves on the shore watching a swimmer now in a trough, now on the crest, as the waves follow one another in unending succession. At one moment he may be almost invisible, at another he seems to grow in size as he rides the sea." How can being almost invisible in a trough be a form of "support"? Burton has translated STRE/FEI. Stinton, responding to Easterling's comment that the up-and-down motion of the waves reflects the ups and downs of Herakles' fortunes, writes (n. 43): "But BA/NT' E)PIO/NTA T' surely refers not to the 'regular up-and-down movement of the sea', and so to the rising and falling fortunes of the swimmer, but to the perpetual succession of waves. The notion of 'up-and-down' in this stanza depends entirely on AU)/CEI, which I argue may be corrupt; though of course the cyclic alternation of good and bad fortune is the central theme in the antistrophe and epode." "Though of course" all but gives the game away; the notion of alternation has been present in the ode in one form or another from the first line. In the epode the theme is generalized in the familiar maxim "no human is perennially fortunate." With this common theme strong in the ode, with POLU/PONON following, with one of the verbs clearly having positive connotations, and a contrast guaranteed by the particle, the notion of alternating good and bad is hard to avoid. Furthermore, the GA/R in 112, which introduces the stanza and shows how it fits into the argument of the ode, is difficult with TRE/FEI; for the chorus is then made to sing "Deianeira has been worried sick about her husband. For he is now nourished, now made great, by constant waves of troubles in life's sea. But some god always saves him," which is, to say the least, awkward. Substitute "buffeted" for "nourished" and it works. (I find that A.S. McDevitt, Eranos 81  9, makes the same point.) Incidentally, remove the comma after POLU/PONON in the OCT. 200. On uncut sacred land cite also R. Parker, Miasma (Oxford 1983) 160ff. The dramatic importance of mentioning Oita here (which Easterling on 1191 calls "a distant presence since early in the play") also needs to be stressed. 205-6. The OCT writes DO/MOS in 205 (Burges' emendation); in this case O( MELLO/NUMFOS goes with it and refers to the coming reunion of Herakles and Deianeira. Davies finds this strained, and prefers to retain the manuscripts' DO/MOIS and write A( MELLO/NUMFOS with Erfurdt, so that the reference is a collective one to the maidens of the chorus. Davies makes the point in his review of Easterling that literary criticism often cannot proceed independently of textual criticism; it seems to me that this is a case where the latter needs the former. Why do we have a chorus of maidens, and not a group of women like Deianeira herself? Because if the chorus consisted of bored, middle-aged housewives, the play would be by Aristophanes, and they would have many blunt suggestions for Deianeira in her difficulty. The fears of Deianeira, for all that she makes a great show of her experience, are not much different from those of the new bride; her fortune depends for good or ill entirely on her man, over whom she has no control. She must be old enough for her fear of the younger Iole's charms to be realistic, but the gap in age between her and the chorus actually serves to stress her similarity to them, not her difference. The first stasimon, which recounts the long-ago marriage of Deianeira as if it were present reality, and ends by describing her as a PO/RTIS E)RH/MA, is merely the most powerful textual indicator of this relationship; an equally powerful indicator lies in the plot itself, for the instrument of Herakles' death springs from the time of Deianeira's marriage, "when I was yet a child" (557, cf. 585 [del. Wunder] of Iole), a time to which reference is repeatedly made from the prologue on. The dominant note of the play up to this point has been Deianeira's longing for a happy reunion with her spouse. Hence DO/MOS O( MELLO/NUMFOS is right. It really is remarkable that Sophocles has chosen to treat the death of the greatest hero in this fashion. He was very likely responsible for the softer portrait of the "man-slayer" Deianeira we find in this play; traditionally she was probably a bloodthirsty murderess like Clytemnestra. The commentators are silent on the truly pathetic nature of 459f.: "Why is it so terrible that I should know [about Iole]? Hasn't Herakles already had more women than any other man?" Marriage for women in Sophocles' Athens was a matter of social status, financial security, of union between OI)=KOI; but most of all if was a matter of fearful apprehension and helpless dependence on men, many of whom had mistresses. Deianeira's memorable sympathy for the captive women is part of this; note how she dwells on their loss of social standing, referring to their fathers at 300ff., 311, 316, 377. It is not hard to see how Sophocles, contemplating the jealousy of Deianeira and the death of Herakles in the light of his own day, produced what he did. This play weighs heavily in the balance when making a case for Athenian men, even if many other things weigh heavily against them. 225-496. "On strictly realistic grounds it is odd to find Deianeira asking Lichas whether her husband is still living (233) when the messenger has already told her he is (182) and the chorus has just performed an astrophic lyric inspired by joy at that news." It seems perfectly realistic to me that she should want to hear the news again from the original source. An actor would in any case deliver the lines in eager tones suggesting that (s)he fully expects a positive answer. 313. Commentators have been rightly baffled by this line so discussion is desirable. 362ff. The attribution of conjectures in the note does not match the attribution in the apparatus of the OCT. 372-3. "E)CELE/GXEIN: on the meaning see Sophoclea, p. 87" is the extent of the comment. The note illustrates an occasional tendency to refer the reader too quickly to some authority, in the form "on x see y," without telling us what exactly "x" is; in the present instance it would cost little effort to convey Sophoclea's conclusion about the meaning, and thus serve the reader's convenience. 444. Davies accepts deletion on the grounds that Iole's "feelings for Heracles are quite beside the point, dramatically irrelevant to plot and theme alike in the austere and selective world of Greek tragedy"; but Sophocles has already made much of Deianeira's feeling for Iole, and the lot of women is a very strong theme which this play has austerely selected. He continues: "Stinton ... has a sensitive defence of the line, which he seeks to justify by reference to 461-2 below. But the argument that the subject of E)NTAKEI/H there 'is most naturally Iole' is a circular one which will only convince those who already believe in the authenticity of 444." The argument is not based only on 444, for the grammar at 461f. makes it very hard to take anything but Iole as the subject. On those lines the authors of Sophoclea write (p. 160): "Deianeira does not care what Iole feels about Heracles, but she does care what Heracles feels about Iole," echoing Lloyd-Jones in CR 33 (1983) 172: "The subject is surely Heracles. Deianeira cares not whether Iole loves Heracles but whether Heracles loves Iole," which Davies quotes with approval. With the greatest respect, I should like to invite either of these gentlemen's wives to fall in love with me, move into their homes, and see whether their thoughts are not at least partly taken up with my feelings in the matter. Of course the subject is Iole. The point is, a normal woman would be apt to despise Iole even if she did not return her husband's affections; even more if she did. Deianeira will not hate her in either case. That she naively assumes that Iole does respond to her husband's attentions is also part of her characterization. 449-50. The antithesis has a Sophistic ring to it (two mutually exclusive and all-encompassing opposites set out in ME/N and DE/ clauses). For Sophistic influence on Sophoclean tragedy see e.g., O. Navarre, Essai sur la rhétorique grecque avant Aristote (Paris 1900) 102ff.; K. Reich, Der Einfluss der griechischen Poesie auf Gorgias (Würzburg 1909); J.H. Finley, Jr., HSCP 50 (1939) 50ff. 470. On present vs. aorist imperatives see now M. van der Weiden, The Dithyrambs of Pindar (Amsterdam 1991) 191f. 512. We are referred here to Davies' forthcoming commentary on Stesichorus for the attributes of Herakles in art; in the meantime one may consult R. Vollkommer, Herakles in the Art of Classical Greece (Oxford 1988) and M.J. Venit, Hesperia 58 (1989) 109ff. (not to mention LIMC). 531ff. "The apparent discrepancy between Deianeira's speech here and at 436ff." is apparent indeed. One may wonder at what point she conceives her plot, and whether she is really so ingenuous in her speech to Lichas as she seems (cf. R.P. Winnington-Ingram, Sophocles: An Interpretation [Cambridge 1980] 77ff.), but in what she says about Iole there is no discrepancy at all. She says in both places quite clearly that she accepts Herakles' dalliance with other women. But what she will not tolerate is that woman's presence in the house (545). Nothing she said in her first speech commits her to accepting such an insult. D.A. Hester's article in Antichthon 14 (1980) 1ff. is well cited by Davies. 539ff. For U(PAGKA/LISMA Davies refers us to A.A. Long, Language and Thought in Sophocles (London 1968) 119-20, but the reader is entitled to find a reference to Ant. 650 in the commentary (YUXRO\N PARAGKA/LISMA, which Gildersleeve -- reference lost -- used to translate irreverently as "a frigid hugging-piece"). 554. "Stinton defends the paradosis ... as referring to the pain Deianeira feels at using deceit: the phrase is taken as an oxymoron... Lloyd-Jones however (Sophoclea, p. 162), more convincingly interpets the phrase in the light of the common tragic metaphor A)/KOS TOMAI=ON (Aesch. Cho. 539) vel sim... Such a remedy by excision is naturally painful but can also be conceived as bringing release... So here Deianeira's remedy for Heracles' affliction can thus be painful, but will also bring release." That Lloyd-Jones actually supports Stinton is not clear from this note; nor can it be right, if this is what Davies means, that Deianeira expects Herakles to suffer. 590-1. The other examples cited of W(S = W(/STE in Sophocles all have an infinitive, which strikes me as easier. 596-7. What Deianeira means by "shame" here is worth comment, as pertinent to her intentions and character in the earlier speech. Cf. Hester, loc. cit. 7. 628. Davies disapproves the OCT text on the grounds that it places "an excessive and inappropriate stress on the reactions of Lichas. What we are interested in is Deianeira and Iole, not the herald." The stress in "you know yourself how well I received her" is quite natural; she means that, as Lichas is fully aware of the situation, he can report it to Herakles with conviction. The stress is the same in 624f.: "you know full well how things stand here at home" (E)CEPI/STASAI; note the preverb). 629 also follows better (Lichas replies, "so that my heart leaps with joy"); Lichas explained earlier that his whole reason for lying was his fear of Deianeira's reaction (479ff.). 661ff. We expect a reference to the robe in these lines; PARFA/SEI gives it obliquely, but Haupt's conjecture FA/ROUS (reported by the OCT) is worth mentioning as having at least diagnostic value. 777. For ... W(S "ut primum" see S. Timpanaro, "Ut vidi, ut perii," Contributi di filologia e di storia della lingua latina (Rome 1978) 219-87. 792. TO\N *OI)NE/WS GA/MON to mean "marriage with Oeneus' daughter" is sufficiently odd to call for comment. 810. The irony is rich here; the son of the house, appealing repeatedly to QE/MIS, publicly curses his mother, whose actions were motivated only by her concern for the OI)=KOS. 930-1. "Hel. 301-2 ... a passage deleted by Hartung (prob. Kannicht)"; contra dixit Fowler, HSCP 91 (1987) 10ff. 940. "Pearson's AI)TI/AI <'M>BA/LOI was one of his three Sophoclean conjectures that Housman found 'evidently true'"; the OCT does not even see fit to put it in the apparatus. Let that hearten anyone who is afraid of living lions. 943ff. On E)FH/MEROS see also R.L. Fowler, The Nature of Early Greek Lyric (Toronto 1987) 114 n. 80, and independently R. Descat, "Idéologie et communication dans la poésie grecque archaïque," QUCC 9 (1981) 7-27 at 20 n. 44. 1055. On XLWRO/S see M.E. Irwin, Colour Terms in Greek Poetry (Toronto 1974) 31-78. 1058. For the negatives cite line 145 of this play. 1085-86. The anapaestic outburst "may be regarded as two tripodies or three metra" (West, quoted here by Davies); but if metra, the second will have the shape - - - u u, whose extreme rarity is remarked on 1275. 1089. Dawe, Studies on the Text of Sophocles III (Leiden 1978) 95, is cogent on this line. E)CW/RMHKEN is a feeble metaphor after DAI/NUTAI and A)/NQHKEN. This is not enough to warrant ousting it, but when we notice that two classes of manuscripts have a gloss, and L a scholion A)NEKA/XLASEN, suspicion is confirmed; that is no gloss for E)CW/RMHKEN. Dawe hunts about in medical writers for a suitable word, and hits upon E)CW/GHKEN, which somehow fails to convince. West's E)CW/RGHKEN, while palaeographically easier and supported by good parallels for the simplex form of the verb at least, also assumes that the answer ought to look like E)CW/RMHKEN; but E)CW/RMHKEN to me has the appearance of a gloss like A)NEKA/XLASEN, so that the origil need look nothing like it. A better tactic might be to discover what company A)NAKAXLA/ZEIN keeps elsewhere. TLG turns up John Chrysostom In Psalmos 55.354.4 TOIOU=TON O( QUMO\S, LOGISMO\N OU)K E)/XWN: OU(/TW ZEI= TO\ PA/QOS KAI\ A)NAKAXLA/ZEI. PW=S OU)=N A)\N SBESQEI/H TO\ NO/SHMA; ZEI=N indeed turn up in medical contexts, and E)KZEI=N is used of a disease breaking out at Arist. Probl. 861b10. Hesychius E 3566 uses E)KZE/SAI to gloss E)CANQH=SAI, showing that the two verbs could have gone naturally together in some context, perhaps a medical one like this. At Aesch. Sept. 709 the same verb is used of curses. It appears to be transitive there like E)CANAZE/SEI at PV 370, but since the verb in both simplex and compound forms, and other verbs of similar meaning, display the capacity to be either transitive or intransitive, the absence of an intransitive parallel from tragedy does not prevent me from suggesting that E)CANE/ZES' is what Sophocles wrote here. 1109. Bruhn and Diggle are cited here for examples of this use of E)K (e.g. Eur. Hcld. 148 E)C A)MHXA/NWN "in a desperate situation"); their examples are all from tragedy, except that Diggle has an instance of a similar use of A)PO/ in Thucydides. In Thucydides I hve noted ... W(S E)K TW=N U(PARXO/NTWN at 7.76 and 8.1.3 ("as best he could in the circumstances"). There may be other prose examples. 1139. E)/NDON is quite specific and hearkens back to the main point of Deianeira's grief: the concubine is to be brought into the house. 1199. Herakles enjoins his son in the strongest possible terms to keep silent at the pyre. Easterling writes "like Heracles himself in the past (cf. 1074)" (i.e., it would be unheroic to weep); Davies writes "A paradoxical forbidding (cf. OC 1751ff.) of the traditional ritual mourning and lamentation expected for a great hero. In fact if H is not really going to die ... this detail will have an extra point." Neither explanation fully accounts for the fearsome curse with which Herakles threatens his son should he fail to carry out the command of silence. Now Socrates expostulates at the end of the Phaedo: "Really, gentlemen, what a fuss you make. It was not least because of this that I sent the women away, to keep them from striking such a false note. I have heard too that one should die in silence," E)N EU)FHMI/A|. M. Alexiou, The Ritual Lament in Greek Tradition (Cambridge 1974), refers to the Platonic passage on p. 5, explaining that to weep before the dying person is actually dead is a bad omen; she cites also Il. 6.500, where Andromache and the women unwisely mourn Hektor though he is yet alive -- a "prophetic and sinister" act (Kirk ad loc.). Of the tragic passages Alexiou mentions, Eur. Alc. 526 seems most apposite; Herakles, who gathers from Admetus that Alkestis is moribund, warns him not to mourn prematurely. Of course, she is dead; but the plot will see her resurrection, and our knowledge of it renders the scene, and this line when understood in this light, ironic. On pp. 26-7 and 38 of her book Alexiou records instances from medieval and modern Greece of the same prohibition. The point is that the soul's progress to its appointed place, whatever that might be, will be impeded. Socrates, who has just spent the whole dialogue arguing that the soul is immortal, and has at this point begun his journey, plainly refers to an established custom; Herakles mentions it, then, not because it is unusual, but because it is particularly important in his case. If the thing is not done properly, his soul will not join the gods. Herakles need have no precise knowledge of this fate; it is enough that he have an inkling. Socrates has no certain knowledge either, no more than any other Greek who adhered to the custom; one merely felt that to avoid bad omens would ease the death itself and prevent whatever was coming from being any worse than it had to be. Realism is in any case easily sacrificed to Sophocles' desire, which was the same as Plato's, to point an allusion without being explicit; Sophocles makes his character sense this need because he is sensible of it. Understood in this manner, the curse now makes excellent sense: "If you do not [remain silent], though I be in the nether world, the full force of my curse shall await you forever." Failure to keep EU)FHMI/A will, in effect, send him in the wrong direction. If the injunction is to be understood in this perspective, then the hunt for certain allusions to the apotheosis of Herakles at the end of the play is over. 1269. It could be commented that when E)FORA=N is used of the gods looking upon mortal deeds it often carries with it the connotation that they will be or ought to be morally outraged; older examples include Od. 13.214, Archil. fr. 177.2 West, Solon fr. 13.17 West. For the indignant overtone here (they look on these events, yet take no action) cf. especially El. 825. 1275. E)P' also seems to be preferred at Sophoclea 178 in spite of what is printed in the OCT.
Of the few misprints I mention only: p. vi n. 1, for "967ff." read "893ff."; p. xv, add to the chart the elongated breve as the symbol for a short syllable at verse-end. Commentary: on 547ff., put the comma after A)/NQOS; on 629, read KE/AR and QUMO/S. Finally, attention should be drawn to the Addenda which incorporate many important observations of James Diggle, and also to unpublished conjectures of Wilamowitz which Davies has been able to report from the margins of Fraenkel's copy of Radermacher's Trachiniae left to the Ashmolean Library.