¹ Where in this paper I speak of ‘the manuscripts’, I mean the παράδοσις, die Überlieferung, as it emerges from a careful recensio, leaving out of account accidental or intentional deviations of individual codices or of a group of codices.

^1a Appendix Vergiliana, Oxford, 1966: The Dirae (Lydia), ed. E. J. Kenney.

² There is no such harshness in Naeke's text, for he makes a violent change: multa prius, fuerit quam non mea libera avena.

³ Following Leo, , Gött. gel. Anz. 1898, 59Google Scholar, I have accepted this Renaissance conjecture, without full confidence, but the point is not relevant to our argument.

⁴ Robin Nisbet reminds me of Verg. georg. 2, 460 iustissima tellus, on which Servius comments: proprie, nam si iustus est qui quod acceperit reddit, terra utique iustissima est, quae maiore fenore semina accepta restituit. J. H. Voss's commentary (1800) compares Philemon fr. 105 K. δικαιότατον κτῆμ΄ ἐστὶν ἀνθρώποις ἀγρός κτλ., but there the argument is different. Xenoph. Cyrup. 8, 3, 38 (quoted by Heyne), πάντων δικαιότατον ὄ τι γὰρ λάβοι σπέρμα, καλῶς καὶ δικαίως ἀπεδίδου κτλ., is much more to the point.

⁵ Vollmer, elated by his conjecture sonis nostris felicia rura, jubilates (Bayr. Sitzgsb. 1908, 11. Abh., 51 n.): ‘damit wäre der rätselhafte, sonst nicht genannte senex aus dem Gedichte beseitigt’. τήνελλα καλλίνικε.

⁶ Leo in his unrivalled interpretation of Virgil's two finest eclogues, Hermes XXXVIII (1903), 6(= Ausgew. Kl. Schr. II, 16).

⁷ This relationship and the meaning of Dirae 10 was understood by H. Keil as early as 1849 (see Reitzenstein, R., Drei Vermutungen zur Geschichte der römischen Litteratur, 1894, 35)Google Scholar.

⁸ Cp. Orth, RE VII, 2185.

⁹ I have discussed it Aesch. Ag. p. 655, n. 1 and 683, n. 3. Add Soph. Ant. 40 εἴθ΄ ᾶπτουσα Porson: ἢ θάπτουσα codd.; Cic. Verr. 11, 2, 5 sic illa provincia Cobet: sicilia prouincia codd.

¹⁰ Joss Bade (Iodocus Badius), who called himself Ascensius after his birthplace Aasche near Brussels, was one of those princely scholar-printers of the first half of the sixteenth century.

¹¹ I mention only Vahlen, Opusc. I, 77 f., 11, 279 ff.; Wilamowitz on Eur. Her. 1106; Leo, Anal. Plaut. II, 36 ff. (= Ausgew. Kl. Schr. I, 158 ff.); Lloyd, G. E. R., Polarity and Analogy (1966), 90 ff.Google Scholar

¹² See Thes. l. L. II, 1502, 41 ff.

¹³ Kl. Beitr. I, 89; Horace 358, n. 2.

¹⁴ See my Horace 430, n. 3.

¹⁵ A. C. Pearson on Soph. fr. 620. Madvig, , Lateinische Sprachlehre für Schulen, 3rd ed., 1857Google Scholar, ş 481 n. 1, remarks that ‘man bisweilen im Lateinischen noch auffallender als im Deutschen sagt, jemand thue (selbst) das, was er durch Andere thun lässt’ and quotes two instances from the Verrines. Add, e.g., Cic. Phil. I, 16 (acta Caesaris) quae ille in aes incidit.

¹⁶ ‘Die nur dichterischen nichtkomponierten Part. Praes. ohne Verben wie comans … stellans … sind sicher an gr. hom. (κάρη) κομόωντες erwachsen (Stolz-Leumann 251).

¹⁷ So Rostagni, A. (‘quem honoris causa nomino’), Virgilio minore (1933), 361 f.Google Scholar

¹⁸ See Thulin, C. O., Die etruskische Disciplin III, (1909), 94 ff.Google Scholar

¹⁹ The function of an abrupt nequiquam at the beginning of a line is the same at Catullus 114, 4, where the mood is entirely different.

²⁰ Bailey quotes less useful parallels and omits Dirae 38, which he could have found in Thes. l. L. VI, 1, 1534, 21.

²¹ In passing I mention another case, a particularly disappointing one, where the new edition of the App. Verg. retains the wrong conventional punctuation, Catal. 5, 11. The last section of this fine poem (I am convinced that it was written by Virgil himself) begins ite hinc Camenae vos quoque; ite iam sane dulces Camenae (so Birt, , Erklärung des Catalepton 1910, 77Google Scholar). The usual punctuation, which connects vos quoque ite iam sane, is deaf not only to the tone of this line—Virgil's Muses are reluctant to leave him, so he has to repeat his request more urgently and affectionately—but to the crescendo of emotion in the whole little masterpiece. Its structure is unmistakable; therefore a blank should be put not after 1. 7, but after 5 and 10.

²² Vollmer actually printed exire. The Munich Thesaurus was loyal to the memory of the Munich professor, but Leumann who, Thes. V, 2, 1362, 53, doubtfully lists our passage with some highly technical utterances of the jurists, makes it implicitly clear that here exire is most improbable and Vollmer's gloss, i.e. contingere, downright impossible.

²³ I am using the word in the sense I have defined p. 144, n. 1. Kenney's note, ‘occulet Ω unde coculet ML … occultet S’, baffles me since, according to the indication higher up on the same page, Ω denotes the ‘consensus of MSL’. The reading of S, occultet, represents clearly an attempt to extract some sense, though no metre, from the unintelligible coculet.

^23a Schulze, Wilhelm, K.Z. LXI, 1934, 144Google Scholar (= Kl. Schr., Nachträge, 1966, 861) remarks: ‘Lat. coᾱgulare coᾱgulator coᾱgulum werden in der späteren Sprache [my italics] zu quagulare quagulator quagulum. Diese silbenärmeren Formen leben nicht nur im Romanischen fort’ etc. My wife reminds me of the quactiliari (= coactiliari) in inscriptions painted on the walls of Pompeii, CIL IV, Suppl. 111, 1, 7809 and 7838.

²⁴ Salvatore, A., ‘Hommages à Léon Herrmann’, Latomus 44, 1960, 690Google Scholar, defends cogulet and says: ‘cioè: “efficiat ut coeant,”, ovvero “cogat” secondo I'usus latior che tale verbo ha, ad es. in bell. Hisp. 5, 5.’ He has looked up the Thesaurus, where the conjecture coagulabantur is given without a warning, but not a modern edition of the Bell. Hisp., where coagulabant, the reading of the manuscripts, is retained. The verb is there used in an intransitive sense; see the commentaries of A. Klotz and G. Pascucci.

²⁵ Sometimes such omissions led subsequent copyists to rather stupid supplements. The text of our poem provides some striking instances, 3 dura for rura, 38 purus for eurus, and, above all, 106 (Lydia 3) est vobis. What the poet wrote was in all probability in vobis (Heinsius)—cp. 113 (Lydia 10) in quibus illa pedis nivei vestigia ponet—; Heinsius's emendation makes it unnecessary to assume a lacuna after 105. But whatever the original beginning of 106 may have been, it is obvious that in the arche type of our manuscripts the first letter or the first two letters had vanished or become illegible and that a copyist thought he could repair the loss on his own; in doing so he probably borrowed the verb from 124 (Lydia 21) et vobis nunc est.

²⁶ M is a valuable manuscript (in fact the consensus of two Monacenses), which alone preserves the genuine reading more than once, e.g. 24 ferantur, 41 crebro, 117 (Lydia 14) reclinavit (the notes on this line have dropped out of Mr. Kenney's apparatus), 158 (Lydia 55) cogor, Copa 8 triclya, but it is not free from arbitrary changes, e.g. 7 fata (for avena), 31 in platanis (for impia cum), 116 (Lydia 13) spumantia (for stipendia), 173 (Lydia 70) hoc (for nam). In the last passage the reason for the interpolation is obvious: the idiomatic nam (for its function see e.g. Vahlen, Opusc. I, 101, 215, Ges. Schr. 11, 659; Housman on Lucan 5, 749; Löfstedt, Coniectanea 56, n. 3) was not understood, as it often was not understood by modern readers.

²⁷ He may be forgiven for ascribing to the praetor the function of a iudex (‘praetorum, usitata significatione, ut significet iudicem suum’).

²⁸ See also Reitzenstein, op. cit. (above n. 7) 40, n. 1: ‘die bisher allgemein angenommene Conjectur praetorum ist, weil das Wort in der archaischen Bedeutung sich in dieser Zeit kaum belegen lässt, nicht wahrscheinlich.’

²⁹ There is an additional point in this notion when Ovid trist. 2, 204 (I have taken this passage and Cic. rep. I, 54, quoted above, from Hey, , Thes. l. L. III, 1225, 13 f.Google Scholar) writes to Augustus neve tuus possim civis ab hoste capi, for ‘der Princeps’ is ‘der erste der Bürger’ (Mommsen, Staatsrecht II³, 750).

³⁰ Kenney's apparatus does not, as it does in other cases, suggest an interpretation of tui, but appends to it four conjectures. Reitzenstein's explanation (op. cit. 40 f.) ‘discordia civis ist dem Dichter offenbar ein Begriff, gleich discordia civilis; sie ist sui inimica das heisst eben civibus inimica’, is improbable, for the only natural way is to take tui … civis together. Vollmer tries to explain tui by referring to Leo's discussion, , Nachr. Gött. Ges. 1898, 474Google Scholar (= Ausgew. Kl. Schr. II, 174), of Propertius 1, 22, 5 cum Romana suos egit discordia cives. Leo says ‘die natürliche Fassung des Gedankens ist cum Romanos cives sua discordia egit.’ This seems to me over-subtle; I am inclined to understand the relation between Discordia (with a capital D) and suos cives in the same way as at Dirae 83 the relation between Discordia and tui civis. It may well be, as Reitzenstein surmises, that Propertius had the passage of the Dirae in mind; see also above on 1. 7 and below on 101, where another possibility is considered.

³¹ Hermes 38, 1903, 11, n. 1 (= Ausgew. Kl. Schr. 11, 21, n. 1). I cannot accept Leo's (and many editors') punctuation of Dirae 94, ‘intueor composlongum: manet esse sine illis’.

³² The vexed question in what sense this post is to be taken cannot be discussed here.

³³ For the meaning of the term see Cyril Bailey, Epicurus 235.

³⁴ In 1842 Meineke dedicated his Delectus poetarum anthologiae Graecae ‘Friderico Jacobsio, anthologiae Graecae sospitatori celeberrimo’. Jacobs was not merely a great Greek scholar; for his contributions to the criticism of Petronius see Bücheler's editio maior, p. XXXXIII.

³⁵ I wanted to look up Jacobs's observations, but the Bodleian Library possesses neither his original publication nor his Vermischte Schriften, where it is reprinted.

³⁶ The decisive arguments have been well formulated by Rothstein, , Hermes 23, 1888, 508 f.Google Scholar, and Wochenschr. f. klass. Phil. 1892, 1088, 1118.

³⁷ Hermes 23, 1888, 518. His conclusion was accepted by Reitzenstein, Drei Vermutungen 38.

³⁸ A. Salvatore (see above, p. 149, n. 24) p. 683, puts tua between quotation-marks and makes it part of a speech of the silva.

³⁹ This fifteenth-century correction has been rightly accepted by many editors. The dicens of the manuscripts may have been influenced by the resplendens of the preceding line.

⁴⁰ Ammann, Thes. l. L. VI, 1, 438, 39 and 447, 71 rightly deals with the two passages separately.

⁴¹ Elsperger, Thes. l. L. III, 565, 13; he refers to, but justly sets apart, Claudian 24, 260 casto genuit quas flumine Ladon, where there seems to be a connotation of sexual castitas.

⁴² Anal. Plaut. I, 42 (= Ausgew. Kl. Schr. 1, 116).

⁴³ The impromptu of a great scholar, W. M. Lindsay, should best be forgotten. In a note published in 1918, C.R. 32, 62, he still adhered to Scaliger's belief that the Lydia was the work of Valerius Cato and asserted that it was composed before the Dirae.

⁴⁴ Perhaps it was the interpolator who cut out the end of the sentence which begins at 40, cum tua cyaneo resplendens aethere, silva, but that cannot be proved.

⁴⁵ I have considered the possibility that the poet of the Lydia and the interpolator of the Dirae may have been one and the same person. But that is improbable, for the stylistic standard of the interpolations is considerably lower than even the not very high standard of the Lydia.

⁴⁶ P. 58, 11 Reifferscheid; Vitae Vergilianae ed. Hardie, C. (Oxford 1954)Google Scholar, ‘Vita Donati’, section 17. On the Suetonian origin of the catalogue see Leo's edition of the Culex, p. 18, and his Griechisch-römische Biographie 12.

⁴⁷ The trick has proved highly successful: Petrarch's sonnet 162, ‘Lieti fiori e felici, e ben nate erbe che Madonna pensando premer sòle’, leans heavily on the ‘Virgilian’ Lydia, as Petrarch's commentators have not failed to notice.

⁴⁸ See Pfeiffer on fr. 382.

⁴⁹ Text in J. U. Powell, Collectanea Alexandrina P. 31.

⁵⁰ 9 times; 3 times in the nominative.

⁵¹ Towards the end of Headlam's lengthy note on Herodas 2, 75–6 we read: ‘The name Battarus actually occurs Cato (sic !) Dirae 1.’ Headlam is not the only one to hang on to obsolete literary history. In an article published in 1926 R. Holland, Rhein. Mus. 75, 177, writes: ‘Wiederum ist Battarus der Name eines Rohrflötenspielers [H. believes that that matters] im Kehrreim der Dirae des Valerius Cato.’

⁵² Naeke, Opuscula I, 304 ff., produced a very long and exceedingly learned article on Battarus; the upshot is nil. A connection with the name Battara (see also W. Schulze, Lat. Eigennamen 394, n. 3), which Naeke suggests as a possibility, is unlikely.

⁵³ Cp. e.g. Ar. Av. 369 φεισόμεσθα γάρ τι τῶνδε μᾶλλον ἡμεῖς ἢ λύκων; Lys. 629 οἰσι πιστὸν οὐδὲν εἰ μή περ λύκῳ κεχηνότι, Plato Resp. 3, 415 e εἰ πολέμιος ῶσπερ λύκος ἐπί ποίμνην τις ἴοι. See also Wilamowitz, Euripides Herakles I², 112.

⁵⁴ Serv. Dan. on Aen. 3, 14. Naeke, on Dirae 8, objects to this combination: ‘Lycurgus enim ille Thrax non tam in montes et silvas grassatus est, quam vites: at vites, certe Catonis [or, as we should say, of the poet of the Dirae], non in silvis erant, sed iuxta silvam: inf. V. 42—a not very convincing argument.

⁵⁵ See above, p. 145, n. 7.