IV. Textual Tradition

Rhesus is one of the rarer plays in the ‘Euripidean Selection’, joining Alcestis, Troades and Bacchae in the relatively small number of sources in which it survives.158 All or part of the text is transmitted in seven manuscripts worth quoting, which, despite a certain degree of contamination, fall into two distinct classes, OV(Va)(Hn) = Δ and L(P)Q = Λ. Their testimony is supplemented by the Ambrosian fragments Ao and Af, three papyri, the medieval gnomologies gV, gB and gE, the scholia and the probably twelfth-century tragic cento Christus Patiens. None of these sources can be disregarded, although pride of place goes to the manuscripts O, V (with its scholia), L (with Triclinius’ corrections) and Q.

1. Manuscripts

The oldest manuscript of the play proper is O (Laurentianus plut. 31.10),159 written by the industrious scribe Ioannikios, whose activity was initially redated by N. G. Wilson from the later thirteenth or early fourteenth century to the last third or quarter of the twelfth.160 Subsequent research has suggested that he may have practised his craft as early as 1135–40,161 in which case the date of his production (including O) might have to be raised by another decade or two.162

The codex comprises the ‘Selection’, except Troades, with two thirds of Rhesus (1–714) concluding the Euripidean part. Under the siglum K it is also a first-rate witness for the text of Sophocles.163 In Rhesus O stands close to V(Va), although numerous separative errors, individual or shared with L and/or Q,164 preclude any direct relationship. On the other hand, it alone is correct at 179 γ᾽ (om. VΛ), 205 κλωπικοῖς, 236 e9783110342079_i0324.jpg 271 λέγω, 343 εἴργοι, 379 Θρῃκί, 431 (430–1n.) ϕόνος, 505 e9783110342079_i0325.jpg 536–7 ὅδε γ᾽, 548 ϕοινίας, 560 (560–1n.) εἰσπαίσας, 595 λιπόντε, 601 οὔτ᾽ ἄν, 619 κτανόντε, 635 θανεῖν (also in the margin of L) and 654 δ᾽ (om. VΛ).165

V (Vaticanus gr. 909) from ca. 1250 to 1280166 is the only manuscript which contains all ‘Selection’ plays, except Bacchae, with substantial old scholia. Its text wavers between O and Λ,167 and several original leaves are missing. In Rhesus we lack 112–51, 551–630, 792–811 and 941–96,168 which, as in some other places (Hec. 212–56, 712–1068, Or. 1205–1504), can be supplied from the fourteenth-century Va (Palatinus gr. 98),169 a full unannotated copy of V. Together V and Va preserve the truth in almost thirty places, most notably 17 λόχος, 161 μέν, 175 (n.) e9783110342079_i0326.jpg296 ὁδοῦ, 453 ὕστερος, 594 τύχη, 601 Ἀχιλλεύς, 669 κοιμίσαι and 718 Ἀτρειδᾶν. In using Va, however, one must bear in mind that its scribe emended or tried to emend many, especially metrical, errors in the iambics and anapaests of his exemplar, probably by a combination of collation, memory and his own conjectural efforts.170 The text of 941–96, moreover, stems not from V itself (already mutilated when Va was prepared), but from a closely related manuscript,171 which seems to have avoided many trivial mistakes.172

The latest member of the Δ family, Hn (Hauniensis 417) of around 1475,173 is for Rhesus, Alcestis, Andromache and Troades an apograph of Va and to be cited only exceptionally. At 131 (n.) μεταθέμενος, 694 (693–4n.) χέρα, 792 ὀρθὸς e9783110342079_i0333.jpg Ω) and 924 (924b–5n.) κἀτυϕλώσαμεν it is correct against all earlier witnesses, but whether through scribal conjecture or a lucky accident remains difficult to decide in each case.174 In 134 e9783110342079_i0334.jpgOQ: e9783110342079_i0335.jpg<Va>HnL), where V is defective and Va illegible, the (wrong) reading of the latter can only be inferred from Hn.175

The other branch (Λ), which also transmits the ‘alphabetic’ plays, is above all represented by L (Laurentianus plut. 32.2), a vast miscellany of poetic works, written around 1310 for the personal use of Demetrius Triclinius.176 Its Euripidean text, set out like the rest in two columns to be read across the page, shows the characteristics of two or, according to some, three different scribes,177 the second of whom wrote Rhesus, Ion, IT and IA (foll. 119r – 154r) and was identified by Turyn (229–33) as Nikolaos Triklines, presumably a relative (younger brother?) of the famous scholar. Triclinius went over the text three times (Tr1 – Tr3), performing the task of the corrector as well as applying his own conjectures or readings he found in other manuscripts. His understanding of metre (by the standards of his time) and the principle of strophic responsion prompted emendation particularly in the anapaestic and lyric sections, where remedies ‘such as changes of spelling and word order, addition or omission of the article, prepositions, pronouns and particles like γε, τε and even οὐκ …’178 produced some successes, but more often (one suspects) led to further corruption. In Rhesus compare e.g. 715 e9783110342079_i0338.jpg δ᾽ e9783110342079_i0339.jpgεἷρπ᾽ e9783110342079_i0340.jpg(suppl. Tr1) with 559 e9783110342079_i0341.jpg· χρόνιος γὰρ ἄπεστιν, where the same insertion after γάρ (Tr3) was designed to remove a paroemiac, or the restoration of 549, 712, (715) and 716179 with the confusion created at 682, 685 and 702. L and Triclinius offer unique (or nearly unique) good readings also at 9 ϕυλλοστρώτους, 14 ἐνέπειν, 82 e9783110342079_i0342.jpgγ᾽ (+ Π2), 322 ξυμπονοῦσιν, 326 αἱροῦσι, 665 δόκει, 734 ἐσιδών, 748–9 ἆ (four times), 753 κέλσαντ᾽, 772 e9783110342079_i0343.jpg(+ ΣV), 784 θείνοντε, 816 νυν, 825 ἔβριξα (also presupposed by ΣQ), 835 οὐδέν᾽, 927 e9783110342079_i0344.jpgand 942 e9783110342079_i0345.jpg. Many of these mean the absence of minor errors.

In P (Palatinus gr. 287 + Laurentianus Conv. Soppr. 172), a luxurious tragic collection produced about 1315–25,180 the text of Rhesus was, like that of the ‘alphabetic’ plays, copied from L after Triclinius had revised it for the first time.181 The manuscript is thus of limited use, except where L is illegible or Tr2/3 has changed the original (for better or for worse) beyond recovery.182 At 561 P alone may attest the correct διόλωλε, whereas at 636 e9783110342079_i0346.jpgἥκεις … a second hand changed e9783110342079_i0347.jpg(Ω) into ὧπερ by erasing the σ and placing a circumflex slightly to the right of the ω. The learned rubricator Ioannes Catrares added one valid speaker assignation (594, also in VaQ) and perhaps an interesting γράϕεται-variant (92).183

Q (Harleianus 5743) of ca. 1500184 contains Trachiniae, Philoctetes, the end of Alcestis (1029–1163), Rhesus and Troades. Notwithstanding some contamination from the first family, its Euripides up to Tro. 610 ultimately depends on an ancestor of L (or P’s model respectively),185 so that in Rhesus it again helps to reconstruct L, prior to any alterations by Tr1.186 Moreover, several peculiar or shared preservations of the truth make it a valuable witness in its own right.187 The survival of a Triclinian reading in Q suggests that Triclinius found it in L’s exemplar rather than conjecturing it himself.

2. Evidence from Lost Manuscripts

To the medieval history of the Rhesus text belong six further direct sources, which, though less important than the manuscripts and papyri, deserve attention.

The relevant part of the Ambrosianus F 205 inf. (Af)188 consists of thirty-eight paper folios of the thirteenth century, on which Andr. 1–102 and Rh. 856–84 + 985–9 are quoted among Homeric scholia. The poetic text, written continuously like prose, is often difficult, if not impossible, to decipher and offers nothing special, except that it confirms the weakly attested readings and/or conjectures at 881 (879–81n.) e9783110342079_i0348.jpgand 883 (882–4n.) ἀνάγει and avoids some trivial errors of V(Va). The second part (Rh. 985–9) is particularly badly damaged.

A much later Milan manuscript supplements the testimony of VQ for the first argument (Hyp. (a) Rh.). Ambrosianus O 123 sup. (Ao),189 a composite codex of the early sixteenth century, transmits on fol. 32r part of Hyp. (a) Med. (89.25–34 Diggle) and the said Rhesus piece. This stands close to Q, probably descending from the same ancestor of L, after Hyp. (a) Rh. had been added there from a V-type manuscript.190 A far better text of the latter half, however, is given by Π1.

Three Byzantine gnomologies preserve independent manuscript evidence for all or most of the ‘Selection’ plays, collecting passages of sententious or stylistic-rhetorical interest in the original verse order.191 They do not usually offer new readings, but may add welcome support where other sources are scarce (especially in Troades, Bacchae and Rhesus, but also Alcestis and Andromache).192 So the eleventh- or twelfth-century Athous Vatopedi 36 (gV),193 which in forty-seven excerpts covers Rh. 84–961, almost certainly antedates even O (making it the earliest medieval witness for the text of Rhesus) and often sides with it in truth (122 θράσει) or error (106, 161, 327, 583).194 Vaticanus Barberini gr. 4 (gB) of around 1300 has fifty-eight lines from Rh. 7 to 980–2 and the early-fourteenth-century Escorialensis gr. X.1.13 (gE)195 seventy-six lines from Rh. 39–40 to 980–2, each with scattered V-type scholia inherited from their respective exemplars.196 Like gV these are more interesting for the transmission than the constitution of the text, with gB supporting Va only in 962 (μελάγχιμον) and gE supplying the syntactically possible, but unlikely, κυνηγέτης in 325 (325–6n.).

With about 250 passages quoted, the author of Christus Patiens197 placed Rhesus fourth in order of utility after Medea, Bacchae and Hippolytus and before even the triad plays Hecuba and Orestes. On the evidence of the varied textual affiliations, his model preceded any major division, although most borrowings have been deliberately altered to suit their new context.198 Yet a demonstrably ancient (and correct) reading survived for Rh. 52 (n.), and other successful variants for 148, 249, 532–3, 658 (nn.) and perhaps 285 (n.).

3. Papyri

To judge by the small number of papyri (three) that have so far come to light, Rhesus was one of the less popular ‘Selection’ plays already in antiquity. In this respect it again sides with Alcestis (four) and Troades (one), but not apparently Bacchae (nine).

Π1 (PSI XII 1286) of the second century AD199 transmits Hyp. (a) Rh. from ]e9783110342079_i0360.jpg κιν[δ- (63.11 -429.10 Diggle) to the end. After a fragmentary beginning, it gives a far more accurate and complete version of the text than the medieval sources (VQAo), not least because such ‘accessory matter’ tended to be less well protected from corruption. The papyrus was the earliest part found of the so-called Tales from Euripides. Its summary of Rhesus is followed by Rhadamanthys and Scyrii.

Π2 (P. Achmîm 4 = P. Par. Suppl. gr. 1099.2)200 from a fourth- or fifth-century AD papyrus codex presents Rh. 48–96 on either side of a single leaf. It introduces the certainly correct οὔτἂν for οὐκ ἂν (ΩgB) in 60 and supports 52 ἥκεις (above), 63 ἦ (coni. Cobet) and 78 e9783110342079_i0361.jpg αἴθειν (cf. 41–2n.). In addition, the corruptions at 51 (33, 49–5lnn.), 54 (53–5n.) αἴρεσθαι and 59 (59–60a n.) e9783110342079_i0362.jpgare proved ancient.

Π3 (P. Oxy. 4568), dated to the third century AD,201 is a scrap of 3.8 × 5.8 cm, which has the openings of Rh. 839–47. Barely a complete word remains, with an obvious error in 841 e9783110342079_i0363.jpgfor ἦλθον Ω) and a peculiar alteration, perhaps by a second hand, in 846 (n.). Had the end of 847 and/or the beginning of 848 survived, we might have been better equipped with regard to e9783110342079_i0364.jpgσῶν (847–8n.).

4. Scholia and Other Indirect Sources

Our relatively sparse scholia and glosses in V, L and Q202 reveal no otherwise unknown truths. But they bear witness, through their text or interpretation, to the original readings in 177 (n.) e9783110342079_i0365.jpg ἀποινᾶσθαι, 441 (440–2n.) ἐπεζάρει (though cf. below on Cyril / Hesychius) and 685 e9783110342079_i0366.jpgand rightly agree with the minority in 772 μετρῶ, 825 ἔβριξ᾽, 852 ἄν and 881 e9783110342079_i0367.jpg(above). On three occasions they add a verse omitted by their manuscript (7, 781) or comment on it as if it was there (234). At 508 ΣV offers the variant ἢ ἄστεος, intended to remove a clash with Homer’s geography (507b–9a n.).

Among the late-antique and medieval quotations, the Atticist lexicon of Orus (~ Et. Gen. cod. B) has confirmed Musgrave’s μηχαναί in 854 (854b–5n.). Otherwise Cyril / Hesychius supports the paradosis in 118 (n.) e9783110342079_i0368.jpgχνόαι, adds to the evidence for 209 e9783110342079_i0369.jpgand 514 ἀμπείρας and is probably the source of the relevant V- and Q-glosses on 441 (above), whether or not the note originally referred to Rhesus or Phoen. 45.

5. The Aldine

Together with most Euripidean plays, the first printed edition of Rhesus was published by Aldus Manutius in Venice (1503).203 According to a persuasive study by M. Sicherl,204 the exemplar in the non-triadic ‘Selection’ dramas was not P, as had long been presumed, but a lost copy of L and the editor most likely Ioannis Gregoropoulos (Aldus’ corrector-in-chief) rather than his more famous fellow-Cretan, Marcus Musurus.

In general the text of the Aldine, which was to become the vulgate for the following three centuries, is of low quality. But several unique good readings, mostly small corrections, give it a place in the critical apparatus. In Rhesus note 90 τεύχεσιν (-σι Π2Ω), 99 (n.) e9783110342079_i0370.jpg(-ας Ω), 201 (201–2n.) e9783110342079_i0371.jpg, 249 τις (cf. Chr. Pat. 1443: τίς Ω), 369 ἐς (εἰς Ω), 441 e9783110342079_i0372.jpg(παιόνας Ω), 525 (523–5a n.) δέχθαι (δέχεσθαι fere Ω), 906 Οἰνεΐδας (-είδας VL: -ίδας Q) and 909 (908–9n.) e9783110342079_i0373.jpg e9783110342079_i0374.jpgVL: e9783110342079_i0375.jpgQ). Of these the epic e9783110342079_i0376.jpgand e9783110342079_i0377.jpg (in lyrics) are remarkable, but probably not beyond the conjectural ability of the Aldine editor.205 By contrast, ὅποι for e9783110342079_i0378.jpgin 689 may come from the manuscript that was taken as the basis for the print.