^{page 188 note 1} I have held the view to be set forth here for many years. A very brief summary can be found in the Proceedings of the Class. Assoc. xlvi (1949), 27Google Scholar (cf. Timpanaro, S., Innsbrucker Anz. v [1952], 202);Google Scholar but an elaboration seems to be called for (Fuchs, H., Mus. Helv. xii [1955], 203, n. 8).Google Scholar

^{page 188 note 2} Hieronymus Columna's edition of Ennius was published after his death by his son Johan nes at Naples ex officina Salviana in 1590. It seems to have escaped the bibliographies that the first two parts of that edition, containing the Annals and most of the minor works, appeared as early as 1585, seen through the press by H. Columna himself. I possess a copy of this earlier edition, from which the first two parts of the complete edition differ only in carrying a slightly amended Elenchus fragmentorum on p. 1, the corrections referring to it on p. 304 being omitted.

^{page 188 note 3} It is accepted, for example, by Vahlen¹ (1854) and ² (1903), L. Mueller (1884), Baehrens (1886), Valmaggi (1900), Diehl (1910), E. M. Steuart (1922), Warmington (1935), Heurgon (1958; see p. 10, n. 3).

^{page 188 note 4} The Annals of Quintus Ennius, Inaugural Lecture (London, 1953), p. 15.Google Scholar

^{page 188 note 5} The epigram envisaged here by Hersilia would dwell on the portentous nature of the victory, e.g. thus: Qui uiduas natas, orbos fecere nepotes Detracts haec generis dant pia dona Ioui.

^{page 190 note 1} S.I.F.C. xxxi (1959), 229 ff.Google Scholar

^{page 190 note 2} 1. 51 ∼ scen. 408; 1. 172 ∼ an. 310.

^{page 190 note 3} Ennius once uses a line with nine mono-syllables, Ann. 431 si luci, si nox, si max, si iam data sit frux, obviously a playful jingle entirely unrelated to our lines, which do not jingle in the least.

^{page 191 note 1} et quis (ecquis Merula; ecqui Timpanaro) se sperat Salmasius: et quis est erat cod., cf. Timpanaro, , Maia iii (1950), 26 ff., who attractively thinks of one of the three pretenders to the kingship in 485, 439, or 385 B.C.Google Scholar

^{page 192 note 1} 162 lines, or a little more than a quarter of all the lines, whole or fragmentary, of the Annals belong to speeches—a very high proportion.

^{page 192 note 2} See particularly the attestation of ann. 5.

^{page 192 note 3} In 1915 Norden, E., Ennius u. Vergilius, p. 62 n.2, complained that it was difficult to obtain a copy of Merula, and my im pression is that Vahlen and L. Mueller relied mainly on Spangenberg's edition. Spangenberg, though on the whole he follows Merula closely, yet asserts his independence: nec tamen Merulam presse secutus sum. Here, he does indeed refer to Livy 8.7 but has discarded Merula's attribution of the line to the speech of the consul and imagines that it was a locus communis opening book 5. Small wonder that later editors were not impressed.Google Scholar

^{page 193 note 1} Livius, , Book IX, Pitt Press, 3rd ed., 1928. I may add in confirmation that, though similar phrases occur at 4. 40. 9 pro uirtute tua fideque, qua una hoc hello res publico stetit and 6. 1. 4 eodem innixa M. Furio principe stetit (urbs), the combination of stare with res Romana is to my knowledge found only in 8. 7. 16 and 9. 16. 19.Google Scholar

^{page 194 note 1} Tusc. i. 89 non cum Latinis decertans pater Decius, cum Etruscis filius, cum Pyrrho nepos se hostium telis obiecissent; fin. 2. 61 nequeporro ex eo (Decio filio) natus cum Pyrrho bellum gerens consul cecidisset in proelio seque e continenti genere tertiam victimam rex publicae praebuisset.

^{page 194 note 2} See Stievano, G., Epigraphica (1951), pp. 33 ff.,Google Scholar and Lévêque, P., Pyrrhos (1957), PP. 395 ff.Google Scholar

^{page 194 note 3} The bracket seems to me strong enough to rule out the suggestion that I, omitted after DIV, was wrongly inserted after LIBV.

^{page 194 note 4} Broughton, , M.R.R. i. 202, n.2.Google Scholar

^{page 194 note 5} Degrassi, A., Inscr. Hal. xiii. 1 (1947), pp. 40 ff., 113 ff.,Google Scholar argues that it was not; Fraccaro, P., Athenaeum, N.s. xxv (1947), 240 ff., thinks it was.Google Scholar

^{page 194 note 6} A.J.A. lx (1956), 243 ff.Google Scholar

^{page 195 note 1} This point is rightly stressed by Stievano, , I.e., p. II.Google Scholar

^{page 195 note 2} On the elimination of unpleasant memories see J.R.S. xliii (1953), 78.Google Scholar

^{page 195 note 3} Pliny, , NH. 28. 13,Google Scholar seems to have in mind an official record when he says: durat immenso exemplo Deciorum patris filiique quo se deuouere carmen, extat Tucciae uestalis incesli deprecatio, eqs. Presumably it is the record on which Livy drew.

^{page 195 note 4} A very different structural analysis is given by Thulin, C.ital. sacr. Poesie u. Prosa (Berlin, 1906), p. 54.Google Scholar

^{page 195 note 5} Of considerable interest is the dicolon uim uictoriam. The alliterative phrase ui uincere is well known; see, e.g., Lucil. 613 uictus ui and the parallels cited there by Marx. But it does not seem to have been observed that the instrumental ablative ui undergoes a strange transformation when, instead of the verb uincere, a nominal derivative is used. Horace, , ep. 1. 10, 37,Google Scholar transforms ui into an adjective: uictor uiolens. No commentary known to me explains that the somewhat surprising expression is the nominal reflection of the set phrase ui uincere. Our formula, on the other hand, coordinates ui with the noun. I suspect this of being an archaizing artificiality rather than genuine archaic diction, but I can offer no proof.

^{page 195 note 6} De dis Rom. Indig. et Nouens., Ind. Lect. Marp., 1892, p. vii. 2: the Nouensiles and Indigetes have changed places, and Tellus seems to have been omitted before Diique Manes.Google Scholar

^{page 196 note 1} Aus altröm. Priesterbüchem (Lund, 1939), p. 48 n.;Google Scholaribid., p. 91, about the prayer as a whole: ‘es ist kein sehr altes Produkt, aber sein Verfasser verstand sich wie auf das Ritual so auf die Formelsprache’.

^{page 196 note 2} A.R.W. viii (1905), Beiheft, pp. 66–81.Google Scholar

^{page 196 note 3} e.g. Livy 1. 24. 8; 9. 5. 3; 21. 45. 8; Paul. Fest. 115. 4 M.: turn me Diespiter… bonis eiciat, uti ego hunc lapidem.

^{page 196 note 4} On dicola in ritual language: Fraenkel, , Plant. im Plant., p. 145,Google Scholar n. 2; lkt.u.Akz., p. 128; Norden, , I.e., pp. 18 ff.Google Scholar

^{page 196 note 5} The post-republican examples (Preuss, S., de bim. dissol…usu, Progr. Edenkoben, 1881, p. 105)Google Scholar are: Sen. dial. 6. 17. 6 qui non incidisset in ilia (Syracusarum mala pestifera) sed prudens sciensque uenisset; Suet. Nero 2 medicumque manumiserit quod sibi prudens ac sciens minus noxium temperasset (uenenum); Apul. apol. 52 tile… ignorans peccat: at tu, miser, prudens et sciens delinquis; Salv. gub. 6. 32 p. et sc; Lact. inst. 2. 3. 3. p. et sc.; 6. 12. 13 p. ac sc. In legal language regularly sciens prudensque: Paul. sent. i. 5. 1; iv. 7. 3; Ulp. dig. 303. 23; 477. 3 (accus.); 560. 34; 721. 39; 743. 17; Gai. 770. 37; so also Querol. p. 11. 15 Ranst. sciens prudensque sacramentorum numquam rupisti fidem?; Div. Fr. Ulp. 482. 18 scientes et prudentes.

^{page 196 note 6} Marcell. 14 and fam. 6. 6. 6 should therefore perhaps be counted as only one occurrence of the phrase. They must, on the other hand, although Caelius in fam. 8. 16. 5 warns Cicero of this same step, count as evidence independent of that passage, because in fam. 6. 6. 6 Cicero culls the phrase from Tragedy.

^{page 197 note 1} Lindsay, W. M., Nonius Marcellus' Dictionary (1901), p. 93.Google Scholar

^{page 197 note 2} Dr. O. Szemerenyi, however, points out to me that narrare is more easily derived from gnarus in the passive than in the active sense. Passive ignarus is attested since Sallust.

^{page 197 note 3} Ennius, I, Les Annales (Cours de Sorbonne, Paris, 1958), p. 71.Google Scholar

^{page 197 note 4} An exact parallel is found in Plaut, . Asin. 562, where, instead of sciens lubens, we have sciens libenter.Google Scholar

^{page 197 note 5} C.Q. xlii (1948), p. 95. After Ennius, of course, we must distinguish between rise and fall. For the rise the pioneer's ruling was hesitatingly accepted, in the fall it was disregarded, and such syllables were either short or a decision was avoided. But Martial is supposed to have written, 5. 69. 3, quid gladium demens Romana stringis in ora, and this prosodic solecism is defended by refer ence to 2. 66. 8, digna specula, where the pheno menon is in the rise and can be paralleled from Catullus (4 exx.), Tib., Gratt., Lucan, Silius, Statius, and Juvenal. The line is obviously corrupt, since Romana ora cannot mean ‘the mouth of Rome’, which is the sense here required.Google Scholar

^{page 198 note 1} Plaut, . Merc. 100; Ovid met. 11. 471.Google Scholar

^{page 198 note 2} ?Plaut. Men. 522 satin ut quemque conspicor ita me ludificant.

^{page 198 note 3} In Hor. sat. 1. 5. 15 absentem ut cantat amicam it is omitted in part of the transmission and probably (so Klingner and others) to be deleted; Heinze retains it but gives it a different interpretation.

^{page 198 note 4} Even the phrase animam de corpore mitto seems to point to a deuotio. Somewhat solemn and circumstantial for, e.g., animam omitto (Plaut. Amph. 240), dimitto (Lucr. 3. 356), fundo (ibid. 1033), it owes its origin, apart from the association by contrast of corpus with anima and the need for a dactyl, to the suggestion of the . Compare particularly in the sciens prudens context Cael. Cie. fam. 8. 16. 5 (above, p. 9), and Varro, , L.L. 5. 148Google Scholarciuem fortissimum eo demitti (the demand of the god, in response to which Curtius hurled himself into the gap in the Forum), further demitto with datives such as neci, Oreo, morti: Virg, . Aen. 2. 85;Google Scholar 398; 5. 691; 9. 527; Hor. c. 1. 28. 11; Val. Flacc.; Stat.; Sil.; al I do not mean to say that de is used in any sense other than separative here. But I do think that the phrase de corpore came to the poet's mind because it dwelt on de-.