Α[ργείους] α[ The supplement, if miswritten apyɩovs—cf. eɩpı in v. 16 and the frequent interchange of — in the Homer papyrus No. CCXXIII "of the same period"-will contain exactly the number of letters required. The remainder I would restore as follows: IO εἰ καὶ ἀπ ̓ ̓Αργείοιο λάχεν γένος Ἡρακλῆος καὶ τούτων φράσσασθε μαχῶν λύσιν· ἶσα δὲ μύθοις Τηλέφου ἶφι δαμέντας· ὅτ ̓ οὐκέτι θωρηχθέντες To this I would add the following commentary:-That the word at the beginning of line 8 means 'if' admits of but little doubt. Merkel in his preface to the minor edition of Apollonius Rhodius, p. v, speaks of the use of for ei, but it seems to me more probable that we have here merely a mistake of the scribe; cf. the similar Homer papyrus No. CCXXIII, E 128, eμev for hμev, and v. 64, ýde for dŋ, though the latter may not be wholly due to phonetic causes. As long as vv. 6-7 are unrestored, it must remain uncertain whether v. 8 is to be connected with them or with vv. 10 ff.; but at present I prefer the latter alternative. For if Telephus is actually (kai) the descendant of Herakles, then the speaker, Astyoche, has a double claim upon Zeus: ov yevernpa | Δαρδάνου ἡμετέροιο καὶ Ἡρακλῆος ἀκούω. For the thought cf. Quintus Smyrnaeus, 10. 40, 319, and 8. 431 ff.: Zeû mátep, ei éteóv ye teĤs ë§ εἰμι γενέθλης τῷ μευ νῦν ἐσάκουσον, which affords also a parallel for the arrangement of the clauses-contrast, e. g., the prayer of the Cyclops, 528 ff. In this connection it may be noted that the author shows a similar unconventionality in his treatment of the unreal conditional sentence. In Homer, in unreal conditions of the past, the apodosis frequently, though not invariably (e. g. 526) precedes; cf. the examples cited GMT., S$435, 440. However, this is not the case in the present unreal condition; cf. GMT., §438, for the examples. This order seems to have impressed the later imitators of Homeric poetry as characteristic, and they have imitated it with great consistency, just as they show a marked tendency to employ Tori as far as possible for pós; cf. La Roche, Wiener Studien, ... XXII 49. So in Apollonius Rhodius the apodosis precedes in I 1298; II 284, 626, 866, 987-995; III 584, 1139; IV 20, 901. The only exception is III 377 ff., which is practically an unreal condition of the present, and so conforms to the Homeric usage. So in the first seven books of Quintus we have this order in I 447, 689, 775; II 507; III 26, 366, 514, 752; IV 301, 329, 563; V 359; VI 503, 542, 570, 644; VII 28, 142, as opposed to but two exceptions: III 444; V 583. Another departure from epic conventions is to be found in the use of the plural λure in a prayer for which I know of no parallel, the citing of Il. 8. 5 by Liddell and Scott being a blunder. Besides, the singular when used in prayer, and the plural in addresses to men, is almost (cf. Quintus, 9. 9) invariably the first word of the speech. For the phrase λαχείν γένος cf. Musaios 30 διοτρεφές αἷμα λαχοῦσα and Quintus 2. 434 Ζηνὸς ὑπερθύμοιο λαχὼν ἀριδείκετον αἷμα. But whether v. 8 be connected with what precedes or what follows, in neither case can v. 9 stand in its present position unless indeed τηλεφον be emended to Τήλεφος, in which case I do not see how a satisfactory close for the line can be obtained. The restitution suggested follows P 426 páxns ámávevőev éóvTes; a possible but less probable ending would have been ἀπάνευθε μένοντα. The occurrence of interpolated lines in the Homeric papyri is not uncommon, and the present instance is no more absurd than, e. g., the insertion after E 83 (O. P., vol. II, p. 101). I would offer the following explanation of its origin. The legend tells how the Greeks were at first successful, while Telephus was absent from the battle, but how he afterwards appeared and swept all before him, until finally he was wounded by Achilles. This crisis may very well have been introduced by some such lines as For the general situation compare the opening of the fourteenth book of the Iliad; for the concrete use of yévos, Apoll. Rhod. 4. 1412, and Quintus 6. 120 Ευρύπυλον κρατεροῦ γένος Ἡρακλῆος. If this line is an interpolation coming from a source of this sort, it follows that the speech before us is embedded in a narrative of the landing of the Greeks in Mysia. And if this is the case, there can 1 Of course I do not mean to insist on the verbal exactness of the first half of this line. Another possibility would be, cf. A 456: ovð laxʼn te tóvos te. hardly be any doubt that the poem was a working over of the material of the Kumpia, bearing a relation to that poem similar to that which the Τὰ μεθ' Όμηρον of Quintus bear to the rest of the epic cycle. A further consequence is that the time of the delivery of this speech must be anterior to the situation in the Iliad. Now, this is in direct opposition to the conclusion which the editors draw from their translation of vv. 1-5. "The situation is therefore posterior to that in the Iliad," and as my restitution of vv. 14-16 turns in part on the same point, it is necessary to inquire into the cogency of this conclusion. The editors evidently can not have based their conclusion upon the unreal condition in v. 2 οὔ κεν ἔτι ζώοντες ἐς Ἴλιον ἦλθον ̓Αχαιοί, for the most that it could have been cited to prove would have been that the Greeks had landed in the Troas-a time ten years before the situation in the Iliad. As a fact, however, it does not prove even that much, for it is merely a case of a type familiar to us all in English-where the speaker, under the stress of emotion, regards as already accomplished that which now seems certain to happen when, had it not been for something, it might have been placed once for all beyond the bounds of possibility. The editors must therefore have drawn their inference from their translation of vv. 4-5: "and Telephus would have slain Achilles, the best warrior among the Argives before he met Hector." "Before he met Hector" in English warrants the conclusion, but πρὶν Εκτορος ἀντίον ἐλθεῖν in Greek does not. This doctrine should at the present time need no proof, as it has long since been distinctly stated; cf. e. g. Foerster apud Sturm, Die Entwickelung der Constructionen mit ПPIN, p. 7: "dass der Infinitiv nach πрív den Begriff einer reinen Handlung bezeichne ohne weitere Angabe, ob eine solche wirklich eingetreten sei oder nicht"; and especially Gildersleeve, A. J. P. 2. 468, n.: “Πρίν is an οὔπω. The 'not yet' may come later, may never come. As I have said of antequam with the subjunctive, the antecedence is necessary, not so the consequence, 'Arédрaσav πрìv κρiðĥvaι, Xen. Hell. 1, 7, 35. They never came to trial." And p. 474: "In Attic it [pi with inf.]... is necessary... when the action does not take place or is not to take place (= ŵσтe μŋ).” After this it seems unnecessary to cite examples, but Apollonius Rhodius, III 374, 660 (πápos), 800, 1395 are all instructive, and Eur. Rhesos 59 ff. (cf. Alc. 362) may be quoted in full: ei yàp φαεννοὶ μὴ ξυνέσχον ἡλίου | λαμπτῆρες, οὐκ ἂν ἔσχον εὐτυχοῦν δόρυ, | πρὶν ναῦς πυρῶσαι καὶ διὰ σκηνῶν μολεῖν | κτείνων ̓Αχαιοὺς τῇδε πολυφόνῳ χερί. So that the only inference that can be drawn is that the situation is anterior to the first meeting of Hector and Achilles and anterior to the landing of the Greeks in Troy, i. e. that the author followed a version of the legend different from that of the source of Tzetzes ; cf. his Τὰ πρὸ ̔Ομήρου, 26ο 17. Since this is the case, no objection can be brought to bear on the restitution of ap]noopa in line 14. At the beginning of this line the editors suggest oude <key>, I prefer, however, ovde <yap>, not only because it makes a better connection with what precedes, but also because the omission of yáp before 'Apyelous is more easily explained. In line 15, powiέartes must, as the syntax shows, be either a misreading or miswriting for powiĝavras. Against [val]μa[ri may be brought objections of both a palaeographical and syntactical nature. In the first place the editors indicate that the lacuna is large enough to hold four letters; but N and A are letters that take a great deal of space, and, in the absence of a facsimile of this fragment, a comparison of the space occupied by NAI in Tov d'idev Aiveías-No. CCXXIII, plate I, 1. 13-which is equal to that sometimes occupied by four letters, will show that these letters may probably be considered as sufficient to fill the gap. The editors give the next two letters as pe, which forces the assumption of a mistake on the part of the scribe, a difficulty that is lessened by his other mistakes-n, apyetot, omission of <yáp> and -and to my mind is outweighed by the fact that the proposed reading fits both metre and sense, if it be admitted that the author would use with the dative as the equivalent of the instrumental. This construction originates in the fact that frequently the same object may be considered either as the instrument or as the place in which an action happens. Hence we find, e. g., both mʊpì káelv and év mʊpì káeiv, the consequence of which is to efface the distinction that originally existed, and to extend the same duality of construction to other verbs where it is no longer logically justified. The beginnings of this encroachment of è with the dative upon the instrumental dative go back to classical times; cf. KühnerGerth3, II 1, p. 464 f.; Lutz, Die Praepositionen bei den attischen Rednern, p. 36; Sobolewski, De praepositionum usu Aristoph., p. 26 f.; that it spread in post-classical times is recognized; cf. Jannaris, Hist. Gr. Gram., §1562, and for Polybius, Krebs, Schanz, = Beiträge, I, p. 71 f.; but the extent of its use in late epic poetry, in the absence of a monograph, is difficult to determine. From Apollonius I have noted: 2. 44 Paidpòs év õμμaoi, rendered by Lehrs alacer oculis; contrast Arist. Knights 550 paidpòs λáμπovтi μετώπῳ; 4. 904 ἐνι χερσὶν ἑαῖς φόρμιγγα τανύσσας = manibus suis citharam intendens. Compare also Musaios 159 Ovμòv épwrоTÓKOLOL napanλáyέas évì μúbois animum amorem-parientibus seducens (in) παραπλάγξας μύθοις verbis; Quintus Ι. 343 ὄβριμον ἐν στέρνοισιν ἀναπνείοντες "Αρηα = pectore spirantes; 1. 400 rà d' év Tooìv ýμáλdvvev = alias pedibus conculcat; 4. 18 Tòv d' év πupì dŋweévra = igni consumptum; contrast Soph. Ο. C. 1319 ἄστυ δῃώσειν πυρί; 4. 225 συνέμαρψεν ἐνὶ στιβαρῇσι Xépeσσ (?) = corripuit robustis manibus; and especially 4. 341 μίξαι ἐν αἵματι χεῖρας ἀτειρέας = ut-tingerent sanguine manus invictas. Furthermore, it may be noted that the MSS read in Eur. El. 1172 ἀλλ ̓ οἶδε μητρὸς νεοφόνοις ἐν αἵμασι | πεφυρμένοι . . . πόδα, and that the Laurentianus reads in Soph. Ο. Τ. 821 λέχη δὲ τοῦ θανόντος ἐν χεροῖν epair xpaivo. Both passages have been corrected, but the errors, if errors they be, are not without significance for the later Greek usage. The above facts may, I think, warrant us in accepting for this poem the construction of pouvioσeiv èv aiμari instead of the usual povíσoei aiμarı, for which compare the examples cited in Liddell and Scott, and Quintus 9. 179. The use of ip in 1. 16 in a periphrasis is without parallel, in Homer or Apollonius, nor do I know of an example from Quintus. The restitution of the next word is suggested by the frequent combination of ipi daμnvai; cf. also Quintus 6. 251 Bin pomáλoso δαμέντα—βίη : ἴς = βίῃ : ίφι. The supplement proposed has exactly the number of letters indicated by the editors and accounts, besides, for the change to the nominative in Owpnx0évres. For it to have been miswritten Tapevtas would have been nothing unusual, as the interchange of and 8 in Egypt is very frequent; cf. No. CCXXIII, passim; Blass, Aussprache, p. 106. T In general style the poem seems to stand much nearer to the level of Quintus than of Apollonius, and I should not for that reason be inclined to place the time of its composition much before that of the writing of the papyrus. A correspondence in mythology with Quintus remains to be noticed. The speaker is Astyoche, the wife of Telephus and daughter or granddaughter of Dardanus; the time is between the wounding of Telephus and the departure of the Greeks; everything would indicate also a time of truce, and hence there |