^{page 163 note 1} Ф⋯&rho&nu&tau&alpha has been altered by Musgrave into περ⋯ντα, by Orelli into παρ⋯ντα, by Hartung into λ⋯γοντα or προθ⋯ντα. Herwerden obelized, Vitelli suggested &pi&rhoο&sigmaФ⋯&rhoο&nu&tau&alpha ϰ&rho⋯ &lambda⋯&gamaου&sigma⋯.

^{page 164 note 1} This has been left as written, but I understand that the transposition (as also кο&nu⋯&sigma&alpha&sigma⋯ at Hel. 1623) had occurred to Dr. Murray himself.

^{page 165 nota 1} If a reason has to be assigned for everything, it is perfectly possible that, in 710, ην is merely teh echo of ἣν: for these subconscious repetitions or anticipations of a letter, a syllable, or a word, written, read, or floating in the transcriber's mind are found everywhere.

^{page 165 note 2} It may be restored, ó&deltaο⋯ &pi⋯&rho&epsi⋯&rho&gamaο&nu, at Ps.-Liban. Parasiti de domino philosophante querella § 2 (Förster, t. vi. 594, 2):

⋯&lambda&lambda&apos ⋯&pi&epsi⋯&iota&delta⋯ &tau&iota&sigma⋯ &delta&alpha⋯&muω&nu &piο&rho⋯&sigma⋯ &#x1F10&beta⋯&sigma&kappaη&nu⋯ &muο&iota &tau⋯&sigma⋯ &tau&rhoοø⋯&sigma⋯ &kappa&alpha⋯ &pi&alpha&rho&epsi⋯⋯&lambda&epsi⋯&tau⋯ &muο&nu &tau⋯&nu ⋯Φο&rho&mu⋯&nu &tauο⋯ &beta⋯ου&hellip, ⋯&pi⋯ &tau⋯ &kappa03CE&nu&epsi⋯&iota⋯&nu &muο&iota &beta&alpha&delta&iota&sigma&tau⋯ο&nu, &delta&iotaó&tau&iota &tau⋯&nu ⋯&pi⋯ &delta&epsi⋯ῖ&pi&nuο&nu οὐ&delta&epsi⋯⋯&sigma⋯ ⋯&gama&epsi⋯ῖ&tau&alpha&iota.

&piο&nuη&rho⋯&sigma⋯ ⋯&beta⋯&sigma&kappaη&nu&epsi⋯*: &piου &pi&rhoο&rho&sigma&epsi⋯&betaá&sigma&kappaη&nu&epsi⋯ (a nonexistent compound) codd., edd. || τ⋯ν La Ma (a distinct family): τ⋯ν cett, edd. || ⋯π⋯*: περ⋯ codd., edd.

For δα⋯μων πονηρ⋯ς, cf. Invidi se deferenlis or. (Fōrster, vi. 637, 15) &piο&nuη&rhoῷ &sigmaυ&mu&pi⋯&pi&lambda&epsi⋯&gama&mu&alpha&iota &delta&alpha⋯&muο&nu&iota(presumably from comedy), Plut. Crass. 22 &tau⋯&sigma⋯ &sigma&epsi⋯ &delta&alpha⋯&muω&nu &piο&nuη&rho⋯&sigma⋯, ὦ &kappa⋯&kappa&iota&sigma&tau&epsi⋯ ⋯&nu&theta&rhoώ&piω&nu, ἣ&gamma&alpha&gamma&epsi⋯&pi&rho⋯&sigma⋯ ⋯&mu⋯&sigma⋯; Alex. 30 οὐ&delta⋯&nu ἔϰ&epsi⋯&iota&sigma⋯ &alphaἰ&tau&iota&#x1F71&sigma&alpha&sigma&theta&alpha&iota &tau⋯&nu&piο&nuη&rho⋯&nu &delta&alpha⋯&muο&nu&alpha, Procop. t. iii. 118, 7 Haury οὐ&delta⋯&nu ἔϰ&epsi⋯&iota&sigma⋯ &alphaἰ&tau&iota&#x1F71&sigma&alpha&sigma&theta&alpha&iota &tau⋯&nu&piο&nuη&rho⋯&nu &delta&alpha⋯&muο&nu&alpha, Ach. Tat. vii. 2 fin. &piο&nuη&rhoῷ &pi&epsi⋯&rho&iota&pi&epsi⋯&sigma&epsi⋯ῖ&nu &delta&alphaῷ&mu&#03BF&nu&iota—With many Other places.

^{page 167 note 1} Equally excellently and equally in vain Brunck reduced to sanity Track. 781 f. &kappa⋯&muη&sigma⋯&delta⋯ &lambda&epsi⋯&#x3C5&kappa⋯&nu ⋯&kappa&rho&alpha⋯&nu&epsi⋯&iota, &mu⋯&sigmaο&nu &kappa&rho&alpha&tau⋯ &delta&iota&alpha&sigma&pi&alpha&rho⋯&nu&tauο&sigma⋯ &alphaἳ&mu&alpha&tauο&sigma⋯ &theta&rsquo ⋯&muοῥ by writing: &kappa&rho&alpha&tau⋯&sigma⋯ &delta&3x1F72 &lambda&epsi⋯&n#x03B7&kappa &muη&epsi⋯&lambda&3x1F40&nu ⋯&kappa&rho&alpha⋯&nu&epsi⋯&iota &mu⋯&sigmaου,, | &delta&iota&alpha&sigma&pi&aplha&rho⋯&nu&tauο&sigma⋯&alphaἳ&mu&alpha&tauο&sigma⋯ &kappa⋯&mu&#x03B7&mu&sigma⋯ &theta&apos ⋯&muοῥ. That the vulgate is as old as Athenaeus is nothing; for such errors are timeless. The Cairo papyrus of Menander, for instance, begins the metrical argument of the Heros with: ἄ&rho&rho&epsi⋯&nu &tau&epsi⋯&kappaο⋯&sigma&alpha &pi&alpha&rho&theta⋯&nuο&sigma &theta&#x1FC6&lambdaὺ &theta&rsquo ᾃ&mu&alpha (= ἂ&rho&rho&epsi⋯&nu &tau&epsi⋯ &theta⋯&lambda⋯ &theta&rsquo ᾃ&mu&alpha &tau&epsi⋯&kappaο⋯&sigma&alpha &pi&alpha&rho&theta⋯&nuο&sigma Wilamowitz).

^{page 165 note 2} The coniecturae minus probabiles are: ‘1023 ὦ παῖ δι⋯ς Barnes, ὦ τ⋯λας Elmsley, ὡς δαις Bothe, ἄθλιε Fix, οὐ δαε⋯ς Kirchhoff, ὦ τ;⋯αν Dindorf, δαμε⋯ς Kayser 1023 f. σὺ δ⋯…τεκ⋯μενος⋯λ⋯κεις…συγκατεργασ⋯μενος μο⋯ρᾳ vel σ⋯ δ⋯…τεκ⋯μενον πάλιν λυσσ⋯δι συγκατετεργ⋯σαθαι μο⋯ρᾳ Paley 1024 νιν μο⋯ρᾳ Fix, σᾷ μο⋯ρᾳ Kirchhoff, σὺν⋯ρμᾷ F. W. Schmidt, μοι ριπᾷ vel μαϰανᾷ Wecklein (olim), δ⋯ μο⋯ρᾳ J. H. H. Schmidt.’ Underhis text, unquestionably against the grain, he records Wilamowitz's transposition, and for μο⋯ρa, proposes an impossible μωρ⋯ᾳ. Dr. Murray follows Wilamowitz, and quotes Kirchhoff's 
σᾷ› μο⋯ρᾳ, which apparently by a slip he attri butes to Bruhn.

^{page 168 note 1} To these should be added H.F. 1190 in locis ἓκατοκεø⋯λου βαøαῖς ὓδρας, where the problem-child is disowned by its parent. For there is no sign of corruption or omission, if the passage is arranged thus: 1182–3 Aμ. ἓτεκε μ⋯ν 
νιύ οὑμ⋯ςῖνις τ⋯λας, | τεκ⋯μενος δ’ ἓκανε, ø⋯νιον α⋯μα τλ⋯ς. | 1188–90 θη. τ⋯ δρ⋯σας; | Aμ. μαινομ⋯νῳ πιτ⋯λῳ πλαγϰθε⋯σ, | ⋯κατσγκεøλου βαøαῖς ὐδρας. | 1184 ff. θη. εὐøημα øώνει. | Aμ. γουλομ⋯νοισιν⋯παγγ⋯λλῃ. | θη. 꽠 δειν⋯ λ⋯ζας. | Aμ. οἰϰ⋯εθ’ οἰϰ⋯μεθα, πτανο⋯ κτ⋯. Dobree's double transposition (‘Sic pone: I. τ⋯ ø⋯ς; τ⋯ δρ⋯σας—2. πλαγϰθε⋯ς—ὓδρας. 3. ὦδ. λ⋯ζας—πτανο⋯. 4. εὓημα etc.’ Adυυ. ii. 120) is not necessary.

^{page 168 note 2} So the passage is best printed: = δ – – is ‘Hρα θ’ ‘Eρμ⋯ς, after which L² interpolates a wellmeant 
θ'‴. Dindorf's remarks (at 1080) are completely justified: ‘Huius epodi metra aliquot in locis tam admirabilia vel potius nulla sunt ut hanc carminis partem imperfectam videatur poeta, perscriptis quae in metri formara redacturus erat sententiis, si fabulam absolvisset. Eandem rationem illius esse carminis puto quod a V. 1283–1314 legitur eqs.’

^{page 168 note 3} His analysis of which is: δδ ∣ = δ –– an (δ) ∣ ia ∣ ia ∣. Murray gives: γηραι⋯ν ∣ πόδ' ἒλκω, τρομερ⋯ν βάσιν. ίὼ τέκνον ∣ κτέThe lines, clearly, cannot be emended with certainty; though, if it is beyond question that Euripides used the adjective τρομ⋯ς (fr. 868 N. = Anecd. Ox. i. 122,8 τρομ⋯⋯ δράμημα γηραι⋯ ποδός), a simple solution would be: γηραιῷ ∣ ποδι τρομ⋯ν ἒλκω [ποδός] βάσιν ίὼ τέκνον ∣ * κτέ.

^{page 168 note 4} Of Schroeder's two other examples: Tro. 250 is doubted even by himself: δί Αίγαίον at Hyps. 64, 104 may just as easily be scanned ∪ – ∪ –.

^{page 170 note 1} Schenkl, Herwerden, and Wecklein cancel 991–5; but, apart from 992, everything is either Euripides or an imitation perfect to the last detail. If the knife has to be used, it should be only on the sick member. The difference between is simply that between τρεπόμενος and τρεπόμεθα is simply that between τρεπόμ and τρεπόμ—sometimes not so much: for, as Bast pointed out, circumstances arise in which the superscribed symbols become identical. The reader, who thought he saw before him: τί τα⋯τα; δακρύοις ⋯ς τ⋯ θ⋯λυ τρεπόμενος ∣ κτεῖν, ε⋯ δοκεῖ σοι κτέ., is not to be censured for setting matters right to the best of his ability. None the less, he would have done better with:τί τα⋯τα δακρύοις ⋯ς τ⋯ θ⋯λυ τρεπόμεθα* ∣ κτεῖν' ε⋯ δοκεῖ σοι κτέ. The connexion of thought is clear, or would be so on the stage. When Menelaus, in the heat of his oratory, comes to that chilling disyllable νεκρούς, his voice falters a trifle and, to brush away the involuntary tear, he lifts his hand, as he was to it later, if the scholiast does not lie, at the chilling trisyllable ⋯πόδος (Or. 643). Then comes the recollection that he is the son of Atreus, the sacker of Troy, Μενέλαος οὐκ ἂγνωστος ⋯ν πάσῃ χθονί as he announces himself at 504 above, and he rises to the typical bravado: κτεῖν' ε⋯ δοκεῖ σοι by the typical relapse into the better part of valour: μ⋯λλόν γε μέντοι τοῖς πιθο⋯ λόγοις.

^{page 171 note 1} Here ϕρεν⋯ν is a clear emendation by land (the manuscripts have ϕρεν⋯ν, but scholiast renders: ούδαμο⋯ συνέσεως ἦν). In the rest, Wecklein states that L has κοὐδαμο⋯, Murray adding the qualification, ‘teste Pultonio’: for οὖν (VBL), MAV² give ἦν and P nothing. The conjectures based on these discrepancies are negligible: κοὐδαμο⋯ μετ⋯ν ϕρεν⋯ν may be mentioned, but only because its authors were Kirchhoff and Nauck. Wecklein, without the shadow of a reason, deleted the verse.

^{page 171 note 2} Certainly not because here, as it happens, the preceding two sentences were also interrogative. See , Soph. Ai. 1272–81Google Scholar, followed immediately by the triumphant: ἆρ ὑμ⋯ν οὖτος τα⋯τ ἓδρασεν ἓνδικα; (1282).

^{page 171 note 3} So the passage ought to be written from V (⋯ρ' εὐτύχησεν ‘Ελλάς ὠλόμην δ’ έγώ): P—which gives ⋯ρ'—accidentally omits ', and p therefore conjectured â δ' εὐτύχησεν, which has had luck to become the vulgate.

^{page 171 note 4} So Wecklein, rightly, for ἆρ' ⋯ξῄδης: Cobet's ⋯ξῄδησθ was pure carelessness—for a second person, classical or not, destroys the sense.

^{page 172 note 1} An excellent emendation, which appears to have made no converts, is Goram's έν Ταύροις πέσω at I.T. 1010, which ought in all probability to be written: ἂξω δέ σ' ἣνπτερ μ⋯ αὐτ⋯ς έν Ταύροις πέσω πέσω κτέ. (ἂξω δέ σ' a Canter: ἂξω δέ γ' μαὐτ⋯ς Markland: καὐτ⋯ς έν Ταύροις Goram: ⋯νταυθαῖ). And, as I am not likely ever to have a more specious excuse for emending Anach. ep. 2 ca. init., I use the opportunity. Anacharsis is advising Solon not to allow the peculiarities of Scythian garb to warp his judgement upon Scythian wits, and he does so in these words: στ⋯λαι δ⋯ κα⋯ κόσμοις σώματος μ⋯ γενέσθωσαν ⋯μπόδιον όρθ⋯ς κρίσεως ἂλλοι γ⋯ρ ἂλλως κατ⋯ νόμους πατέρων κεκοσμήκασι τ⋯ σώματα. For στ⋯λαι Westermann conjectured στολαι, and so (with κóσμος) Hercher reads. But anyone who, for his sins, has trudged what Democritus would have called the μακρ⋯ ⋯δ⋯ς ⋯πανδόκευτος of the Didot Epistolographi must, I think, concede that the best professional manner calls for a little more local colour: πῖλοι* δ⋯ κα⋯ κόσμοι [ς]* σώματος κτέ. For if Timnes—or Tymnes, ε⋯ τόδ αὐτῷ ϕίλον κεκλημένῷ—told Herodotus truth, Anacharsis was uncle 'Ιδανθύρσου τα⋯ Σκυθέων βασιλέος (iv. 76), and therefore a pilleatus (Iordanes v. 39 ut refert Dio—in the Getica— … qui dicit primum Tarbosteseos, deinde vocaXos pilleatos hos qui inter eos generosi extabant, ex quibus eis et reges et sacerdoles ordinabantur: see , Bud'sDio t. ii. 394, 399Google Scholar, with or. lxxii 3, and compare also , Luc. Scyth. 1Google Scholar οἴκοι δ⋯ οὐ το⋯ βασιλείου γένους ὢν (Toxaris) οὐδ⋯ τ⋯ν πιλο ϕορικ⋯ν, ⋯λλ⋯ Σκυθ⋯ν τ⋯ν πολλ⋯ν κα⋯ δημτκ⋯ν, οἶοί ε⋯σι παρ' αύτοῖς οί ⋯κτάποδες καλούμενοι). By way of abusing the opportunity, I correct also ep. 9 (p. 104, 40): ἲνα τε γ⋯ς χρώματα δια ϕόρους μαστεύον θα⋯μα πεποίηνται. This alone is the authentic tradition, preserved in L: Hercher's text is drawn from interpolated sources and depraved further by himself. Read simply ἲνάς τε γ⋯ς χρώματι* διαϕόρους (gold and silver) κτέ.: cf. , Plut. def. or. 434Google Scholar B in. μόλις οἶον ἶνες ἤ τρίχες ⋯ραιαί διατρεχουσιν ⋯ν τοῖς μετάλλοις. The whole epistle is ruined by Hercher, who alters the soundest readings and leaves the most transparent corruptions untouched—for instance, a couple of sentences later: ⋯ποχέτευσιν ποιεῖσθαι κελεύουσιν οἶς ⋯παινεῖν ἒρως ⋯στίν. Read, of course: ύγιαίνειν*, and compare, if something has to be compared, , Synes. enc. calv.76Google Scholar A ὃστις ύγιαίνειν ⋯ρ ᾷ μιμείσθω τόν ίατρικ⋯ς εύρετήν.

^{page 173 note 1} Philostratus, who was a judge of such things, appears to have agreed with me (Ep. 5 Σκύθης … ⋯π' ⋯κείνου το⋯ βωμο⋯), but not So his editors: at vit. Apoll. ii. 14 Kayser destroyed the pretty: κα⋯ ἒχιδνα ⋯ϕθη ποτ⋯ τοὺς ὂϕεις οὓς ⋯πέτεκε λιχμωμένη κα⋯ θεραπεύουσα έκείνῇ γλώττῃ, by the ridiculous conjecture έκκειμένῇ), which is reprinted in the Loeb Philostratus and even translated by Phillimore. Equally lamentable are the attempts, at Tro. 1188, to displace ὒπνοι' τ ⋯κεῖνοι by such things as ὒπνοι τε κοινοί or ὒπνοι τε κλῖναι— as though the tragedy of life did not chiefly reside in the necessity of saying ‘that’ and not ‘this’.

^{page 173 note 2} If the void has to be acknowledged, it has to be a very little one. A curious instance is the well-known passage: 0.T. 943 f. Io. π⋯ς εἲπας; ἦ τ⋯θνηκε Πόλυβος; Αγ ε⋯ δ⋯ μ⋯ ∣ λέγω γ' έγώ τ⋯ληθές ⋯ξι⋯ θανεῖν. That, and nothing else is the tradition, and it is impossible, and nothing else. Nauck, by one of the clearest and best corrections ever made in the dramatists, restored the hand of Sophocles: Io. π⋯ς εἲπας; ἦ τ⋯θνηκε Πόλυβος; Αγ ε⋯ δ⋯ μ⋯ ∣ λέγω γ' έγώ τ⋯ληθές ⋯ξι⋯ θανεῖν. The reason for the corruption, which Nauck unaccountably failed to notice (‘eine zu εἰ δ⋯ μή beigeschriebene Erklärung λέγω τ⋯ ⋯ληθές mag zur Entstellung des Textes den Anlass geboten haben’), is sun-clear: the eye of some transcriber passed from τέθνηκεν to τέθνηκε, and the relics: Io. π⋯ς εἲπας; ἦ τ⋯θνηκε Πόλυβος; Αγ ε⋯ δ⋯ μ⋯ λέγω γ' έγώ τ⋯ληθές ⋯ξι⋯ θανεῖν were expanded to two trimeters Yet even in the Oxford text the lines appear thus: Io. π⋯ς εἲπας; ἦ τ⋯θνηκε Πόλυβος; Αγ ε⋯ δ⋯ μ⋯ λέγω γ' έγώ τ⋯ληθές ⋯ξι⋯ θανεῖν—the second part of Nauck's proposal (as vital as the first) is not mentioned. That is to say, rather than admit the loss of four words where the loss of some words is undeniable, the unanimous testimony of every manuscript that counts is thrown to the winds by a falsification of εἰ δ⋯ μή λέγω γ' ⋯γὼ to εἰ μή λέγω, and a palmary emendation is degraded to an idle guess which has far less than one chance in ten thousand to be right.

^{page 174 note 1} At Phoen. 878 f. all manuscripts but P (which has ποῖα for όποῖα) offer: ⋯γὼ τί δρ⋯ν ὂποῖα δ οὺ λέγων ἒπη, ∣ εἰς ἒχθος ἦλθον παισί τοίσιν Οίδίπου—the scholia recording a good variant, εἰς ὅχλον ἦλθον. Murray gives, with some Byzantines, τί «οὐ» δρ⋯ν ποῖα δ' κτ⋯. (others had tried τί δρ⋯ν «οὐ»), cites Σ τ⋯ οὺ ⋯π⋯ κοινο⋯ τί οὺ δρ⋯ν γράϕεται δ⋯ κ⋯γὼ τί μ⋯ δρ⋯ν ποῖα δ' (Kirchhoff's version is γράϕεται δ⋯ κα⋯ ⋯γὼ τί μ⋯ δρ⋯ν ποῖα δ, but I have not Schwartz at hand), and concludes: ‘lectio dubia: cf. ad Hel. 56’, where our manuscripts give τί δ⋯τ ἒτι ξ⋯; and those of Aristophanes (Thesm. 868) τί οὖν ἒτι ξ⋯ ‘fortasse recte’. The lection certainly is dubious: ⋯γὼ is in the air, and the negative for δρ⋯ν can only be procured at the cost of a probably illegitimate hiatus. Kirchhoff was so little satisfied that he assumed the loss of a verse between 878 and 879. The assumption is reasonable, though speculative, but again he misplaced his asterisks: for, if 878 and 879 are the first and third verses of a trio, then, whatever the second verse, the necessity for τί οὺ δρ⋯ν in the first is not removed. The lacuna, if it exists, comes where the break is betrayed by the jobbing repair ὂποῖα. No doubt, a number of passable supplements might be invented; to myself Melpomene suggests nothing better than:

⋯γὼ τί δρ⋯ν «⋯πεῖπν έξιώμενος

σπουδῇ ματαίᾳ»;* ποῖα δ' οὺ λέγων ἒπη κτέ.

Again, at , Soph. Phil. 1250Google Scholar ff.

Οδ. στρατόν ' Αχαι⋯ν οὐ ϕοβῇ πράσσων τάδε;

Νε. ξὺν τῷ δικαίῳ τ⋯ν σ⋯ν οὺ ταρβ⋯ ϕόβον

Νε. * * *

Νε. άλλ' οὺδέ τοι σ χειρί πείθομαι τ⋯ δρ⋯ν κτ⋯.

the universally accepted asterisks are due to Hermann, who, like Dobree, required also: τ⋯ν σ⋯ν οὖ ταρβ⋯ στρατόν. The vulgate is defensible at a pinch, but usage speaks in favour of the change (see, for example, , van Leeuwen on Av. 893Google Scholar), and Hermann could have had his στρατόν without the change of a letter or the addition of a star:

Νε. ξὺν τῷ δικαίῳ τ⋯ν σ⋯ν οὺ ταρβ⋯ ϕόβον « στρατόν. Οδ. * * *» ϕόβον.*

^{page 175 note 1} καθαγνίσω Camper, καταψεκ⋯ or κατακλύσω Koechly, στέψω κάτα Bergk, καταπλάσω Rauchenstein, κατασπερ⋯ Zigeler, καταρραν⋯ Gloēl, σ⋯νσποδ⋯ν κατασβέσω or δ⋯ν κατασβέσω δέμας Wecklein, χραν⋯ τ ⋯λαίῳ σ⋯μα σ⋯ν κοίτας ἒσω Goram.

^{page 176 note 1} And, it appears to me, beyond Wecklein's. At any rate, when he made the attempt, he produced only: ἢ δ ⋯κ χιτώνων «νυκτί προσϕερ⋯ν ίδεῖν ∣ ⋯μο⋯ στομωτ⋯ν» π⋯ρ πνέουσα καί ϕόνον κτ⋯. The words themselves have been hammered out on the basis of ἓδοξ τρεῖς νυκτί προσϕερεῖς κόρας(Or. 408), coupled with a hint from ⋯στομωμένη above and the recollection of an emendation by Hermann in an Aeschylean fragment. The underlying theory would seem to be that, while to speak of a Fury breathing fire and slaughter from the shirt(s) is a thing that cannot be done, all proprieties of thought and language are adequately safeguarded, if you add in Johnsonian Greek that the shirt is as black as midnight, Those of the Erinyes were, however, according to the best authorities, only dark-grey.

^{page 176 note 2} I venture an emendation of Xen, . Cyr. vii. 5, 25Google Scholar, which I transcribe with Mr. Marchant's note: οὐδ⋯ν ἆν εἲη θαυμαστ⋯ν ε⋯ κα⋯ ἆκλειστοι αί πύλαι αί το⋯ βασιλείου εῗεν κωμάζει γ⋯ρ ή πόλις π⋯σα τῇδε τῇ νυκτί

‘εἶεν] ὡς ⋯ν CAEGH κωμάζει] κώμῳ δοκεῖ CAEGH: κωμοδοκεῖ DF: κωμ⋯ζει Stephanus: ⋯ν κσμῳ Muretus addito εἰναι Post πᾳσα: εὐωϰεῖται Madvig, unde εὐωϰεῖσθαι γ⋯ρ δοκῖ Schenkl: δειπνεῖ Hug: ὡς ⋯ν omisi quod videtur varia lectio esse: deinde qptissimam coniecturam Sephani accepi.'

The solution does no injustice either to Xenophon or to the reader; but from the data one would suspect that Gobryas said: κώμῳ δ' ἔοικε* ⋯ π⋯λις π⋯σα κτ⋯. At Resp. Ath. i. 14. i n., δοκοῥσι is again impossible: περ⋯ δི τ⋯ν συμμ⋯ϰων, ὃτι (del. Morus) ⋯κπλ⋯οντες συκοøαντοῥσιν ὡς δοκοῥσι κα⋯ μισο⋯σι τοὺς ϰρχστο⋯ς κτ⋯. Mr. Marchant quotes Zurborg's κ α⋯ διᏎκουσι, but, as the confusion between δοκεῖν and δεῖν is ubiquitous, κα⋯ δοῥσι* seems preferable: compare Plut. Per. 16 ⋯ δ⋯ Tηλεκλε⋯δης παραδεδωκ⋯ναι øησ⋯ αὐτῷ τοὺς' Aθηνα⋯ους II⋯λεών τε ø⋯ρους αὐτ⋯ς τε π⋯λεις, τ⋯ς μ⋯ν δεῖν, τ⋯ς δ' ⋯ναλ⋯ειν.

¹When the hinted deed is done the chorus sings (809ff.):

κρε⋯σσων μοι τ⋯ραννο ἔøυς

ἤ δυσγ⋯νει ⋯ν⋯κτων,

ἂ νῥν ⋯ σ ο ρ⋯ ν øα⋯νει

ζιøη⋯ρων ⋯ς ⋯γώνων

ᾃμιλλαν, ε⋯ τ⋯ δ⋯καιον

θεοῖς ἒτ' ⋯ρ⋯σκει.

For the spaced words, unmeaning, unmetrical and me iudice unemended, I propose simply: ⋯σορ⋯σα øα⋯νει* κτ⋯. The sense, at any rate, is clear and good: 'Thou art a better prince for me than those baseborn kings who now, looking thy sword in face, make it aboundingly plain whether or no the good cause is still the cause of God'. With the language there can be no quarrel: for the rather uncommon use of øα⋯νει, compare soph. Trach. 1158 øανεῖος ἔν ⋯ν⋯ρ ⋯μἘμς καλῈ, and for the rest such passages as: ⋯ς κρ⋯νος βλ⋯Ψαντα κα⋯ λ⋯γϰης ⋯κμ⋯ν Suppl. 318, εἰ μὴ κυρ⋯σεις…| øε⋯γοντας ⋯λ⋯ σ⋯ν βλ⋯ποντασ ⋯ς δ⋯ρυ Rhes. 113f., ⋯ς ϰ⋯ρας λε⋯σεις ⋯μ⋯ς; Phoen. 597, κακ⋯ δ' ⋯ς ⋯λκὴν κα⋯ σ⋯δηρον εἰαορ⋯ν Med. 264, et simm.

^{page 180 note 1} So at Tro. 809: ὃθ' 'Eλλ⋯δος ἄγαγε πρ⋯τον ἄνθος ⋯τυζ⋯μενος | πώλων κτ⋯., Where his ⋯τεμβ⋯μενος is the only alteration cited by either Wecklein or Murray. ⋯τυζ⋯μενος, it must be granted, is about the least apposite participle in the language to apply to the mood of Hercules when he had been chated of his horses; but it can conceal nothing more recondite than ⋯τι⋯μενο.* For the genitive with ⋯τιμ⋯ζω cf. Aesch. P.V. 783 μηδ' ⋯τιμ⋯σῃς λ⋯γου (Elmsley: λ⋯γους codd.) Soph. O.C. 49 μ⋯ μ' ⋯τιμ⋯σῃς |τοι⋯νδ' ⋯λ⋯την ὦν σε προστρ⋯πω øρ⋯σαι, Ant. 21 τ⋯øου… |τ⋯ν μ⋯ν προτ⋯σας τ⋯ν δ' ⋯τιμ⋯σας ἒϰει: with ἂτιμο, O.T. 788, El. 1214, Thuc. iii. 58. 5. etc.: with ⋯τ⋯ζω, Ap. Rh. i. 615 οὓνεκ⋯ μιν γερ⋯ων ⋯π⋯ δηρ⋯ν ἂτισσαν. The change is the lightest possible: ⋯τυζομ⋯νω and ⋯τιζομ⋯νω are variants at Pind. Ol. viii. 39, ⋯τ⋯ζει and ⋯τ⋯ζει at Nic. Alex. 193.—On the other hand, at Rhes. 327: ⋯ρθ⋯ς ⋯τ⋯ζεις κἀπ⋯μομøος εἰ ø⋯λοις, the word by itself would be inept, with ⋯ρβ⋯ς it is fatuous. Here Herwerden tried κακ⋯ζεις, which seems to me little better. The sense required is δικα⋯ως ⋯γανακτεῖς—that is to say: ⋯ρθῶς ⋯γ⋯ζεις* κτ⋯. (Hesych. ⋯γ⋯ζειν βαρ⋯ως ø⋯ρειν, E. M. ⋯γ⋯ζει. ⋯γανατεῖ κα⋯ βαρ⋯ως ø⋯ρει). The verb—cf. Aesch. Suppl. 1061 (Where the meaning is not too clear, but cannot be ‘exalt overmuch’, as given by Liddell and scott) and Bekk. Anecd. 336, 6 ⋯γ⋯ζεις ⋯ντ⋯ το⋯ θρασ⋯νεις Σοøοκλ⋯ς—is certainly in keeping with the diction of the play. So, at 835 f.: σὺ τα⋯' ἒδρασας οὐδ⋯ν' ⋯ν δεζα⋯μεθα | οὓθ' οἱ θαν⋯ντες οἄτ' ἂν ον οἱ τετρωμ⋯νοι |ᾳλλον κτ⋯., I Find δεζα⋯μεθα so unnatural that am tempted by the Wardour Steet δι⋯μεθα*.

^{page 180 note 2} The reading is there doubtful but the doubts hardly affect the point at issue. My own version would be something like: μ⋯τηρ δ' ὡσε⋯ πτανοῖς κλαγγ⋯ν | ὃρνισι, οπως ⋯ζ⋯ρζω 'γὼ | μολπ⋯ν οὑ τ⋯ν αὐτ⋯ν 
⋯ϰὰν‴ | οἰανποτ⋯ δ⋯ |σκ⋯πτρῳ IIρι⋯μου den: διερειδομ⋯να || øρυγ⋯ους Wilamowitz: øρ⋯αις.

^{page 182 note 1} Since no spectator at the première of Iphigenia in Tauris had with him an annotated edition of the play, one wonders how many understood the lines (636 ff.): ⋯λλ' εἷμι δ⋯λτοντ' ⋯κ θε⋯ς ⋯νακτ⋯ρων | οἳἳσω τ⋯ μέντοι δυσμν⋯σ μ⋯ μου λ⋯βῃς. |υλ⋯σσετ' αὐτο⋯σ, πρ⋯σπολοι, δεσμ⋯ν ἄτερ. I make no claim beyond comprehensibility for the following:

τ⋯ μ⋯ντϻι δυσμεν⋯ μ⋯ μοὐγκαλῇς,*

(ad servos conversa)

øυλσστ' αὐτο⋯ς, πρ⋯σπολοι, δεσμ⋯ν ἄτερ.

The Kirchhoff, without changing the punctuation, should have conjectured μ⋯ μοι 'γκαλῇς (‘debebat ’γκ⋯λει' remarks Wecklein) is impossible to explain and not easy to believe. But, in any case, if there is any case, if there is anything in the proposal the progress of error (μοὐγκαλῇς = μου λ⋯βῇς implies the crasis as a starting -point.—A probably hopeless passage of the Helen (447 f.) goes:

Με. ἂγγειλον εἲσω δεσπόταισι τοῖσι σοῖς …

Γρ. πικρ⋯ς ἂν οἶμαί γ' ⋯γγελεῖν τοὺς σοὺς λόγους. It once occurred to me that the old woman's line might be a faulty correction of:

πικροὺς ἂρ (Hirsching+Hermann) οἶμαι τοὺς λόγους σοὺς ⋯γγελεῖν (=σούαγγελεῖν)*.

So, at Cyal. 288, Radermacher restored πόρς ἂντρα σούσαϕίγμέους, and at 252 Dr. Murray's πόρς ἂντρα σοὺσαϕίκοντο ξένοι may be the best cure for π⋯ρς ἂντρα τ⋯ ⋯ϕίκοντο ξένοι. At El. 413 κέλευε δ' αὺτ⋯ν τόνδείς' δόμους ⋯ϕιγμένον ∣ ⋯λθεῖν it seems to me that a plausible alternative to his τ⋯νδ' ⋯μο⋯ σαϕιγμένων.*

^{page 182 note 2} Amphitryon obeys, and the passage proceeds (1203 ff.):

ὦ τέκνον,

πάρες ⋯π ⋯μμάτων

πέπλον ⋯πόδικε ῥέθος ⋯ελίω δεῖξον.

βάρος ⋯ντίπαλον δακρύοις συναβμιλλ⋯ται.

ἱκετεύομεν ⋯μϕ⋯ σ⋯ν

γενειάδα κα⋯ γόνυ κα⋯ χέρα προσπίτνων

πολιόν τε δάκρυον ⋯κβαλών.

Here βαρος … συναβμιλλ⋯ται (an unknown compound) is unitelligible, the metre of the next two lines is defective, and ⋯κβαλών should be ⋯κβάλλων. For δακρύοις συναβμιλλ⋯ται, Hermann followed by most, gave δακρύοισιν ⋯μιλλ⋯ται, ‘quod si verum‘, says Dr. Murray (though he prints συναβμιλλ⋯ται, Which passes may own faculty for belief) ‘hiat oratio, supplendaque ex. gr. σοῖσι τάδ' άμέτερα'. Schroeder concurs, so far as the lacuna goes, and he fills it by «σοῖς ⋯μόν». Schenkl, on the other hand, wished to place βάρος … ⋯μιλλ⋯ται after πολι⋯ν ⋯κβάλλων. It would then be clear that the counterpoise to the of Hercules is the tear of Amphitryon, and no omission need be postulated. Unfortunately, the asyndeton is insufferable, and ⋯μιλλ⋯μαι would seriously affect the credibility of the alteration as a whole. But the difficulty can be evaded without the change of a letter:

δάκρυον ⋯κβάλλων

βάρος ⋯ντίπαλον δακρύοις ἳν'* ⋯μιλλ⋯ται.

In what precedes, wilamowitz' ίκετύομεν ⋯μϕ⋯ γενειάδα κα⋯ ∣ γόνυ κα⋯ χέρα σ⋯ν προπίτνων πολιόν τε κτ⋯., While not certain, is unlikely to be bettered.

^{page 184 note 1} At Hel. 1495 ff.: μόλοιτέ ποθ' ἲππιον οἶμα (changed to οἶδμα ∣ δι' αἰθέρος ἱέμενοι ∣ παῖδες Τυνδαρίδαι, there is a marginal note by l: γρ. ἂρμα. If it is an emendation of his own, he intended it to be governed by ἱέμενοι: if, by some improbable chance, it is more, then it is not ‘a bold cognate accusative’ (as even an editor like Pearson assumed it to be, comparing, of all things, ἒστηκε πέτραν), but a nominative or vocative, according to taste, and means ‘a mounted pair’. The nearest approach in Euripides to this sense is at Andr. 276 ἦλθ' ⋯ Μαίας τε κα⋯ Δι⋯ς τόκος ∣ πρίπωλον ἂρμα δαιμόνων ∣ ἂγων τ⋯ καλλιζυγές. Later the affection becomes more common: e.g. Charit. 478, 19 Didot ἂρμα βασίλειον (king and queen), D. Chrys. or. xxxvi. 41 ἂρμα Νισαίων ἲππων, Himer. or. xiv, 10 κύκνοι δ⋯ ἦσαν τ⋯ ἂρμα simm.

^{page 184 note 2} So, at Phoen. 1578 ϕάσλανον εἲσω ∣ σαρκ⋯ς ἒβαψεν, P and the scholia alone preserve ἓβαψεν, M has the honest ἒβαλεν, ABVL substitute ἒπεμψν.

^{page 185 note 1} The interjection might perhaps he restored at I.A. 552, where the chours deprecates too ardent love: ⋯πενέπω νιν ⋯μετέρων ∣ Κύπρι καλλίστα θαλάμων. The trouble too clearly lay in enforcing the veto, and the best chance, it would seem, was to say with a hypodiastole: ⋯πένεπ', ὢ, νιν ⋯μετέρων κτ⋯. The conceit need not be paralleled.