Main Text


The present commentary is based on and keyed to James Diggle's excellent Oxford Classical Texts edition; no personal inspection of manuscripts has been made. This section will therefore contain only a bare minimum of information on Rhesus' manuscript tradition, along with essential bibliography on the matter.

The earlier stages of the manuscript tradition of Euripides' plays are extremely hard to reconstruct with any confidence; readers who wish to explore the question will find an excellent starting-point in Barrett's or Zuntz's essays on the main stages of the transmission.265 Here we need only state that the Euripidean manuscripts now extant are generally considered to represent a tradition ultimately deriving from the Alexandrian edition by Aristophanes of Byzantium. A selection of ten plays (Alc., Andr., Med., Hipp., Hec., Tr., Ph., Or., Ba., and Rh.) seems to have gradually developed and established itself, perhaps in the third century of our era.266 Some time 900 a scholar or scholars produced an edition, in the new minuscule script, of the ten-play selection; equipped with variant readings, glosses, and commentary, this edition made good use of earlier scholarship, and was destined to become the ancestor of all extant medieval manuscripts of Euripides.267

pg lxxviAs we have seen, Rh. was one of the selected plays, and so manuscript evidence for it is relatively ample. Here is a list of the manuscripts used in Diggle's OCT; most of them are regularly referred to in my commentary:


Laurentianus 31. 10 (c.1175). It contains only ll. 1–714.


Vaticanus graecus 909 (c.1250 × 80).268 It contains only ll. 1–111, 152–550, 631–791, and 812–940.


Palatinus graecus 98 (14th c.); an apograph of V, used mainly to supplement its missing text.


Hauniensis 417 (c.1475); an apograph of Va, now scarcely cited, although often used in nineteenth-century editions.


Laurentianus 32. 2 (c.1300 × 1320).


Palatinus graecus 287 (early 14th c.). An apograph of L incorporating Triclinius' first set of corrections in that codex (Tr1; see below).


Harleianus 5743 (late 15th/early 16th c.). Its scribe has been identified by Wilson 1966: 337 as Gian Francesco Burana of Verona (b. 1474).


Ambrosianus O 123 sup. (early 16th c.). It offers only Hypothesis (a) Diggle.


Ambrosianus F 205 inf. (13th c.). It contains only ll. 856–84, 985–9 ad.269

Papyrus evidence for Rh. is limited to three items: (i) PSI XII 1286 (2nd c. ad), which offers part of Hypothesis (a) Diggle (see Hyp. a n.); (ii) P.Achm. 4 = P.Par. inv. BN, Suppl. gr. 1099. 2 (no. 427 Mertens–Pack3; ad 4th/5th c.), containing (parts of) ll. 48–96 (see 41–3 n.); and (iii) P.Oxy. 4568 (ad 3rd c.), containing traces of ll. 839–47.270 Portions of Rh. are also transmitted in Byzantine gnomologia:271 Athous Vatopedii 36 (gV, 12th c.), Vaticanus Barberini gr. 4 (gB, c.1300), and Scorialensis gr. X. 1. 13 (gE, early 14th c.); however, their usefulness for the establishment of the text is limited (cf. 161–2, 206, 327–8, 411–12, 482 nn.).272 Finally, an important source of often correct readings (cf. e.g. 65–7, 90–4, 285–6, 388–9, 443 [a remarkable pg lxxviiinstance], 506–7, 532–3, 656–9, 938–40 nn.) is the so-called Christus Patiens, a Christian pastiche drama sewn together largely from Euripidean passages.273

It has long been known, on palaeographical grounds, that VVaO on the one hand and L(P)Q on the other belong to two distinct families (Diggle's Δ‎ and Λ‎ respectively).274 L is a scholar's copy written in Thessalonica by Nicolaus Triclinius, and corrected at three successive stages (commonly abbreviated as Tr1, Tr2, Tr3) by his more famous brother Demetrius Triclinius.275 Its apograph P (on whose precise relationship with L see next paragraph) helps us recover L's reading where the latter is illegible. A late manuscript of rather small value owing to its serious corruptions, Q, derives from the same ancestor as L. Their only difference is that, whereas L was copied directly from that putative ancestor, Q seems to derive from it only 'through intermediaries and after the prefatory matter, and also a few places in the text of the play had been altered on the model of a manuscript of the V type';276 moreover, Q contains none of Triclinius' alterations of L. In the Δ‎ family, V, the work of a scholar (perhaps in Planudes' milieu) rather than a mere scribe, has significant similarities with the Λ‎ family, although not affiliated with it, and is sometimes the only manuscript to preserve the correct reading (e.g. Rh. 161, 168, 238, 248, 323, 453, 501, 601 [Va], 669; cf. 607, 851, 879–81).277 As for O, it often offers better readings than V (e.g. Rh. 138, 177, 505, 560, 595, 619, 643, 687, 705; cf. 253–7, 496, 704–5), although it suffers from the uitium Byzantinum, i.e. the tendency to reshuffle the words in order to make the line paroxytone, in conformity with Byzantine standards of versification (cf. 170, 218, 220, 426, 433, 503, 506, 606, 618, 635, 636).

The relationship between L and P continues to be a vexed question: 'Is one of them a direct copy of the other, or do they derive, independently, from some common ancestor, or has still another relation between them to be acknowledged?'278 Alexander Turyn, the scholar who laid the foundations for the serious study of Euripidean manuscript tradition, believed that L and P derived from a common ancestor rather than P being a copy of L as earlier pg lxxviiischolars (Vitelli, Wecklein) had thought.279 However, Zuntz, on the basis of a thorough comparison of readings and metrical marginalia in L and P, showed that in the 'alphabetic' plays—namely, El., Hel., Hcld., Herc., Su., IA, IT, Ion, Cyc., i.e. plays surviving independently of the selection of ten plays mentioned above—P was copied from L after Triclinius had applied on L his first set of corrections (Tr1).280 (We now know that Turyn's notion of a common ancestor for LP had also been exploded by W. S. Barrett 2007: 422–5 in a review published only posthumously.) This also holds for Rh., although it is not an 'alphabetic' play: P seems to have copied both the text and the prefatory material (Hypotheseis etc.) directly from L after the latter had been revised by Tr1, since later alterations made by Triclinius on L (Tr2, Tr3) do not recur in P.281 Responding to K. Matthiessen's objection that the case in favour of P's being a copy of L was unconvincing as far as Rh. is concerned,282 Diggle 1994: 508–13 undertook a careful re-examination of the evidence, which confirmed that (i) P cannot be independent of L, and (ii) 'P has all of the corrections (restoring readings not otherwise attested) which can be attributed with certainty to Tr1.'283 Despite the exact and careful work of Zuntz and Diggle, opposing voices, according to which L and P are twin copies made directly from a common model, do not seem to be dying away.284


265 W. S. Barrett (ed.), Euripides: Hippolytos (Oxford, 1964), 45–90; see also id. 2007: 420–31; Zuntz 1965: 249–88. For excellent, concise accounts see also Lesky 1983: 201–8 and, more recently, D. Kovacs in J. Gregory (ed.), A Companion to Greek Tragedy (Malden, 2005), 379–93, esp. 383–7.

266 The selection may reflect certain cultural tendencies, such as a more or less established repertoire of plays considered classic (possibly as early as the 4th c. bc), or may be due to the exigencies of the school curriculum, in which Christian supremacy would have allowed only limited room for pagan plays; for the former factor see Easterling 1997; for the latter see Zuntz 1965: 255–6. At any rate, Wilamowitz's theory (Einleitung in die griechische Tragödie (Göttingen, 1895), 216) that the selection was compiled on the initiative of a single person for use in schools is no longer considered an adequate explanation.

267 See Zuntz 1965: 261–75.

268 However, Wilson 1966: 342 argues for a somewhat earlier era, namely 1204 × 61.

269 For a list of the MSS see Diggle 1994: 508 with n. 2; for descriptions see Jouan, pp. lxxi–lxxiii; on the character of L and P in particular see Turyn 1957: 222–64; Zuntz 1965: 126–40; on Q see Turyn 1957: 288–98; Zuntz 1965: 144–51; Wilson 1966: 337; on V see n. 277 below. For bibliography on OVLPVa see J. Diggle, The Textual Tradition of Euripides' Orestes (Oxford, 1991), 6 (OV), 8 (LP), 10 (Va).

270 On the papyri cf. also Jouan, pp. lxxiii f.

271 Cf. n. 242 above.

272 On the manuscript tradition of some Euripidean gnomologia (including Barb. and Scor.) see A. Garzya, BPEC2 16 (1968), 77–84 at 77–81 = Storia e interpretazione di texti bizantini (London, 1974), no. XX(vi), pp. 77–81. For collations of these three gnomologia see: on Ath. Vat. G. A. Longman, CQ2 9 (1959), 129–41 (esp. 141 on Rh.); on Barber. K. Matthiessen, Hermes, 93 (1965), 148–58 (esp. 153 on Rh.); on Scor. id., Hermes, 94 (1966) 399–410 (esp. 408 on Rh.); cf. Jouan, p. lxxiv n. 162.

273 Cf. p. lxix above, and see further Jouan, pp. lxxv f. with bibliography.

274 On colometric errors (i.e. errors in the identification and separation of cola, or metrical units) shared by VVaO but not by L(P)Q and vice versa see Pace 2001: 10; cf. Delle Donne 2004/5: 192–3. On the status of P and Q in particular see next paragraph.

275 The first to see that L's corrector was Demetrius Triclinius was Turyn 1957: 224, 242–54. On its ancestry see Zuntz 1965: 180–92; on Triclinius' work on the MS see again Zuntz 1965: 16–125 (the classic disquisition); for a concise statement cf. Jouan, p. lxxi.

276 Zuntz 1965: 145–51; the quotation is from p. 149. Cf. Wilson 1966: 337.

277 See Di Benedetto 1965: 18, 23–51 for a full account of V's character and position with regard to the rest of the manuscript tradition. On its similarities with LP see again Di Benedetto 1965: 25–8. On instances in which V alone preserves the correct reading see Di Benedetto 1965: 34–5, who also points out (p. 32–4) that V's quality has sometimes been overestimated.

278 Quotation from Zuntz 1965: 1.

279 A. Turyn, The Byzantine Manuscript Tradition of the Tragedies of Euripides (Urbana, 1957), 222–306, esp. 264–88. For a critique and refutation of Turyn's theory see H. Lloyd-Jones, Gnomon, 30 (1958), 503–10 at 505–8; cf. P. G. Mason, JHS 82 (1962), 163–4; Barrett (n. 265) p. 429 and esp. id. 2007: 420–31.

280 See Zuntz 1965: 1–15, 21–2, 51–62; cf. Diggle 1994: 512; on the general reliability of Zuntz's identification of three different inks used by Triclinius at the three successive stages of correction see Wilson 1966: 335–6. Di Benedetto 1965: 114–15 had also seen that the differences between L and P are no less numerous, or significant, than their similarities. Cf., however, the caveat offered by O. L. Smith, Mnemosyne4, 35 (1982), 326–31: P seems to contain a set of metrical scholia which, although usually considered to be late, may well reflect Triclinius' later work (i.e. even later than Tr3).

281 Cf. also P. G. Mason, Mnemosyne4, 11 (1958), 123–7.

282 K. Matthiessen, Studien zur Textüberlieferung der Hekabe des Euripides (Heidelberg, 1974), 16 n. 14.

283 Quotation from Diggle 1994: 513.

284 See e.g. A. Garzya in J. L. Heller and J. K. Newman (eds.), Serta Turyniana (Urbana, 1974), 275–90, severely criticized for errors by Diggle 1994: 298–304; A. Tuilier, Étude comparée du texte et des scholies d'Euripide (Paris, 1972), 122–35 (esp. 132), castigated for its many shortcomings by Diggle 1994: 137–42. More recently, S. Martinelli Tempesta, RIL 125 (1991), 227–68, after giving a useful account of the status quaestionis (227–32), provided the first-ever full collation of L and P in Rh., and concluded that the two MSS were twins because, among other things, of cases in which P seemed to take no account of Tr1 (258 n. 95, 260–3). However, Tempesta often resorts to special pleading, and Diggle 1994: 517 argued that he often mistook Tr1 for Tr2 or Tr3. Rather surprisingly, Jouan (p. lxxiii) follows Zanetto (pp. xviii f.) in not even attempting to distinguish between Triclinius' three different sets of corrections. Recently, Pace 2001: 14 argued that a study of colometric variants supported Turyn's thesis that, at least for Rh., L and P are derived from a common ancestor; however, as Pace herself is aware, the evidence of colometry cannot be decisive; on the significance of colometric evidence cf. also Delle Donne 2004/5: 191–2.

logo-footer Copyright © 2018. All rights reserved.
Access is brought to you by Log out