¹ Some such division mark was required; for it could never have occurred to Catullus that his text might be presented in the form in which we see it, for instance in the Oxford text, with IUVENES and VIRGINES above the initial lines of the stanzas and quasi-stanzas.

² Most editors (but not Ellis and Lafaye) assume the loss of a line after 41. I would not deny the possibility of such a loss, but the only argument in favour of it which I have found mentioned is the petitio principii of a consistent and mechanical symmetry. I therefore agree with Wilamowitz, , Hellenist. Dichtung II, 278Google Scholar (although I do not share his view on the function of the refrain), in believing that nothing is missing in this stanza.

³ cf. Sabbadini, R., Le scoperte dei codici latini e greci I (1905), 1 fGoogle Scholar. The jurist and poet Lovato Lovati (1241–1309; see the sketch in Sabbadini, , o.c. II [1914], 105 f.Google Scholar), the ‘decano dei preumanisti dovani’ (Billanovich, Giuseppe, I primi umanisti le tradizioni dei classici latini, Fribourg, 1953, 21Google Scholar), already possessed a copy of Catullus, as will be demonstrated by Guido Billanovich; in the mean-time see the hint given by Giuseppe Billanovich, o.c. 41.

⁴ The correct view has often been stated. I mention the point to put the reader on his guard against the confusion in Ellis's commentary.

⁵ cf. Rehm, P-W VIII, 1255 f., Wilamowitz, , Hellenist. Dichtung II, 179, n. 1Google Scholar.

⁶ Neither Kroll in his short app. crit. nor Schuster (1949) in his extensive one mention V's reading, parent.

⁷ The passage in question is Sappho fr. 95 Bergk (120 Diehl, 104 Lobel and Page), Ἔσπερε πάντα φέρων ὄσα φαίνολις ἐσκέδασ᾿ Αὔως κτλ. In the following line Lobel could perhaps have spared the first of his daggers, for φέρεις ὄιν, φέρεις αἷγα might be accounted for in the manner suggested by Wilamowitz, , ‘Textgeschichte der griech. Lyriker,’ Abhdlgn. Gött. Ges. d. Wiss., Phil.-hist. Kl. N.F., Bd. 4, 3, 1900, 72Google Scholar. But the end of the fragment is a puzzle. What puzzles me most is whether ἀποφέρεις τῇ μητρί can mean you take away from the mother’ as well as ‘you bring back to …’. So we cannot even be quite certain that the fragment comes from an epithalamium.

⁸ The view that custodia denotes the ‘tres viri nocturni’ seems to have been held by the earlier commentators; see e.g. Valpy's note and the rejection of that view in Friedrich's commentary. That is at any rate more sensible than Baehrens's reference to the ‘sensus amatorius’ (reproduced, along with Friedrich's note, by Lenchantin) or Friedrich's wild suggestion (‘der Stern wacht selbst’) or the classification of the Thesaurus IV, 1558, 42Google Scholar, under the heading ‘in re militari’. Kroll says ‘die Wache wird ausgestellt’ and leaves it at that.

⁹ Etym. M. 117, 20; cf. Liddell and Scott, 9th ed., p. 2111.

¹⁰ Wilamowitz, , Hell. Dichtg. II, 285Google Scholar, endeavoured to change the character of this stanza. Accepting Bonnet's punctuation (mentioned in Riese's commentary, who, as against it, compares ll. 176 ff., in Ellis's Oxford text, and by Lafaye, who adopts it), he removed the comma after novi and placed it at the end of the preceding line (31), after voca. Kroll in his second edition (see also ‘Nachträge’, p. 297), Lenchantin, and Schuster followed suit. It looks then as if the text in this rearranged form were henceforth to become the vulgate. I reject the new punctuation and what it implies. To say nothing of its artificiality, it disrupts the connexion between ‘mentem amore revinciens’ and the rest of thestanza, ‘ut tenax … errans.’ According to the traditional punctuation, which I accept, mentem means the mind of the bride, and it is, of course, the bride, symbolized by the ivy, of whom it is said that her mind is being bound fast in love; it is also implied that she is embracing, clinging to him, the male, the tree (tenax … implicat). If we take together ‘coniugis novi mentem revinciens’, we make the final clause with its typical imagery all but meaningless.

¹¹ It is not without interest to compare the pattern of ll. 42–4 with that of 64, 146–8; the sublimitas, to which the repetition of certain words is subservient, is equal in both passages.

¹² But et is retained by e.g. Is. Vossius, Ellis, Riese, G. Giri (De locis qui sunt … corrupti in Catulli carminibus, Turin, 1894, 268Google Scholar), Merrill, Lafaye, Lenchantin. How enduring the influence of a convention can be may be seen from T. Heyse, who prints Et tu but translates ‘Doch du …’.

¹³ Sufficient examples of simple αίνοι concluded by this formula of application are quoted Rhein. Mus. 73, 1920, 366 fGoogle Scholar. Here I add a few passages from early poetry in which the same formula is used in addressing a person who is to apply to himself the lesson of a preceding general maxim. Homer, , Il. 9, 513Google Scholar, Hesiod, Erga 27, Theognis 99 (in the last two cases the transition is made by σὺ δέ, where δέ has its ‘continuative’ force), Pind., Ol. 12, 13 ffGoogle Scholar. (where καἱ σὺ is replaced by καἱ σὺ … τιμά). At Bacchyl. 13 (12), 66, καἱ σὺ. after the general maxim introduces not an advice but a report. Finally a Hellenistic example: Theocr. 29, 1 f. The famous line Virg., Ecl. 10, 69, ‘omnia vincit Amor; et nos cedamus Amori,’ is a perfect instance of the pattern with which we are concerned.

¹⁴ This view is still found in Kroll's introductory note (‘nichts, das nicht griechischem Brauch entspräche; da auch die geographischen Andeutungen nach Griechenland weisen [about this error see p. 3 above], so ist klar dass C. ein griechisches Original überträgt’). However, in the appendix to his re-edition (1929), p. 297, he mentions Wilamowitz's correction. I need hardly add that Kroll's commentary as a whole is very valuable.

¹⁵ G. Friedrich (p. 281 of his commentary) says ‘wir bemerken in dem Gedicht echt römische Züge’, but all that he adduces to prove it is the topic of ‘vitis ulmo coniuncta marito’, which will be discussed presently. A much sounder scholar, Arthur Wheeler, L., professed the same conviction when in his book ‘Catullus and the Traditions of Ancient Poetry’, Sather Classical Lectures IX, 1934, 190Google Scholar, he said that in this poem allowance should be made ‘for a possible admixture of Roman elements’, but he no more than Friedrich came to grips with the problem.

¹⁶ Lucian, Symp. 8, αἱ γυναῖκες ὅλον τὸν κλιντῆρα ἐκεῖνον ἐπέλαβον, οὐκ ὀλίγαι οὖσαι, καὶ ἐν αὐταῖς ἡ νύμφη πάνυ ἀκριβῶς ἐγκεκαλυμμένη ὑπὸ τῶν γυναικῶν περιεχομένη. For further evidence see e.g. K. F. Hermann-Blümner, Griech. Privatalterthümer 271 f., Daremberg-Saglio III, 1650, P-W VIII, 2129 f.

¹⁷ Riese in his introductory note: ‘als die (vom Mahl, an detn sie teilgenommen [Lukian conv. 8]; aufgebrochene) Braut wieder aus ihrem Gemach ‘erscheint …’.

¹⁸ Robert, C., Hermes 35, 1900, 659Google Scholar.

¹⁹ Not, however, Isaac Vossius, of whom Riese, l.c., says that he was ‘der erste, der das Gedicht als aus Sappho übersetzt ansah’. Vossius, , whose annotated edition (London, 1684)Google Scholar I have before me, did nothing of the kind; he merely adduced several fragments of Sappho to illustrate Catullus.

²⁰ Kroll regarded Robert's argumentation as conclusive. Wheeler, o.c. 274, n. 70, is sceptical, but instead of going into the matter (he could have found here a striking instance of ‘the Roman elements’) he contents himself with saying that Robert's and Kroll's ‘assertion is not certain’.

²¹ Pomponius, , Dig. 23, 2, 5Google Scholar, ‘deductione enim opus esse in mariti [non in uxoris] domum, quasi in domicilium matrimonii.’

²² The conservative obstinacy of Ellis, who at l. 77 clings to the final stroke of the MS reading adest (despite 95 f., 110, 120, and 122), is now repeated in Schuster's Teubner edition of 1949.

^22a If it were the case that ‘Hesperus … abstulit unam’ depends on Ἔσπερε πάντα φέρων κτλ, we should have to conclude that Catullus made the important change from the timeless ϕέρεις to the past abstulit.

²³ In general cf. V. Hehn, Kulturpflanzen und Hausthiere, 6th ed., 73, Hermann-Blümner, Griech. Privataltert. 108.

²⁴ That the wedding of the vine points to Italy rather than to Greece has often been said, cf. e.g. Schmidt's, B. edition, Proleg. LXXV f.Google Scholar, Magnus, H., Bursians Jahresb. 97, 1899, 217Google Scholar, Friedrich's commentary, p. 281. The petitio principii in Kroll's note on l.49 (‘das mehrfache Vorkommen bei römischen Dichtern wie Hor. Epod. 2, 9, adulta vitium propagine altas maritat populos weist auf hellenistische Vorlagen’) is of a piece with his view on the absence of the bride.

²⁵ Ars. P. 325 ff., ‘Romani pueri longis rationibus assem discunt in partis centum diducere,’ etc.

²⁶ P-W IX (1916), 132Google Scholar (‘Hymenaios’). His statement that Catullus LXII is ‘bedeutende und einheitliche Poesie’ compares favourably with E. Norden's verdict (Römische Literatur, 4th ed., 1952, 37), ‘eine blosse Studie: Sappho in moderne Technik umgesetzt, wahrscheinlich schon nach Vorgang eines hellenistischen Dichters.’

²⁷ For the importance of Wilamowitz's studies on Catullus, see JRS 38, 1948, 32Google Scholar.

²⁸ See especially p. 280, ‘Wir lernen also, dass Catull keineswegs ein sapphisches Gedicht übersetzt, sondern aus ihren Hochzeitsliedern hier und da hernimmt was ihm gefällt, daneben auch von Kallimachos.’

²⁹ So also Wheeler, o.c., 187 and 215.