¹ I sincerely thank my colleague Martin Bloomer for a critique of an earlier draft of this paper; the three anonymous readers for their incisive criticisms and generous suggestions; and the Editor, Professor Woolf, for his helpful guidance and encouragement. With deep grief I dedicate this article to a teacher who knew and loved the oldest Indo-European languages and poetries very deeply.

² The text is that of Büchner, K., Fragmenta Poetarum Latinorum Epicorum et Lyricorum praeter Ennium et Lucilium (1982 ²)Google Scholar, 40 = Blänsdorf, J., Fragmenta Poetarum Latinorum Epicorum et Lyricorum praeter Ennium et Lucilium (1995 ³), 73Google Scholar, with two exceptions. (1) I read Orc- for Orch- after Courtney, E., The Fragmentary Latin Poets (1993), 47Google Scholar. Cf. ‘Die alte Schreibung und Aussprache ist sicher Orcus. Die Innenaspiration war im 1. Jhdt. v. und im 1. Jhdt. n. Chr. fast allgemein durchgedrungen, wird aber dann wieder aufgegeben’ (PW ² 35.908.45–9 s.v.). The aspiration of the voiceless stops, which Cicero noted as having taken place in some words in his lifetime (e.g. pulc(h)er, triump(h)us, Cart(h)ago, Or. 160), may be due to the r or the l (v. Allen, W. S., Vox Latina (1978 ²), 27CrossRefGoogle Scholar). In an archaizing poem the archaic form may be cautiously preferred. (2) I also read Romai, that is Romāī, after Cole, T., ‘The Saturnian verse’, Yale Classical Studies 21 (1969)Google Scholar, 65 n. 97 in order to produce the usual trisyllable (here, a Bacchius) after the caesura Korschiana (thus the name for the word break that occurs before the final three or four syllables in a Saturnian hemistich of seven or more; for a treatment, cf. Cole, op. cit., 19–21). The form is here a false archaism; cf. n. 9.

³ Most codices have Orcho … thesauro, potentially ‘to Orcus to his treasury’ or ‘to Orcus as a treasure’; for a critique of those possibilities, cf. Courtney, op. cit. (n. 2), 48. The lost codex Buslidianus had orchi … thesauri.

⁴ In view of the subsequent obliti sunt, which relates to a historical event, I take the imperfect flerent also to refer to past time, as commonly in older Latin, perhaps with the common nuance ‘would have been likely to’.

⁵ Leo, F., Der Saturnischen Vers (1905)Google Scholar, 57 n. 1 conjectured oblitae, referring to Camenae; for a defence of the transmitted reading, built on the presence elsewhere of the ‘Verwaisungsmotiv’, cf. Suerbaum, W., Untersuchungen zur Selbstdarstellung älterer römischer Dichter: Livius Andronicus, Naevius, Ennius (1968), 35–6, 304–5Google Scholar.

⁶ Gellius doubtless encountered the epitaph in Varro's de poetis; cf. Dahlmann, H., Studien zu Varros ‘de poetis’ (1962), 65–8Google Scholar. The same letter, 1.24, records also an epigram for Plautus and an epitaph for Pacuvius, and expressly names Varro as the source of the former.

⁷ Marmorale, E. V., Naevius Poeta (1950 ²), 140–1Google Scholar observed that the metre is too regular to be genuinely Naevian. To his arguments may be added the observation of Cole, op. cit. (n. 2), 24 that Livius and Naevius have no octosyllabic second hemistichs as found in 4b.

⁸ Suerbaum, op. cit. (n. 5), 31 n. 96, 36–9 points out that the content of the poem has its parallels in literary, not genuine sepulchral, inscriptions; in this he is followed by Courtney, op. cit. (n. 2), 49, who adds an aesthetic argument, observing, inter alia, that ‘[t]he alliterations in the second cola of the lines are remarkable; indeed they look like sowing with the whole sack, an imitator, as so often, overdoing things’. The authenticity of the epigram had been doubted at least since Ribbeck, O., Geschichte der römischen Dichtung I: Dichtung der Republik (1887), 26Google Scholar; cf. also Leo, F., Geschichte der römischen Literatur (1913), 438 n. 1Google Scholar; op. cit. (n. 5), 57 n. 2. For other earlier opinions, see Marmorale, op. cit. (n. 7), 141–2 n.

⁹ Romāī is strictly a genitive; the archaic locative would have been Romāĭ, with -āĭ scanned as a diphthong. But Romāī loc. is possibly a false archaism, if the poem is the work, as I will argue, not of a speaker of archaic Latin, but of a later period; such a writer might have observed that -ae gen. comes from archaic -āī and wrongly extended the equivalence to the locative, where, furthermore, the ending -ī regularly appears (Corinthī, Carthaginī, rurī, domī).

¹⁰ Here, in using a close formal reading to explore the cultural meaning of an epigrammatic text, I follow the lead of van Sickle, J., ‘The elogia of the Cornelii Scipiones and the origin of epigram at Rome’, Transactions of the American Philological Association 108 no. 1 (Spring 1987), 41–55Google Scholar.

¹¹ For summaries of various views of the structure of the whole prayer, see Guittard, C., Carmen et prophéties à Rome (2007), 206–20CrossRefGoogle Scholar.

¹² Viduus is properly ‘mateless’ and therefore, potentially, ‘unproductive’. That the latter is the meaning underlying viduertas here is suggested by its morphology: the usual abstract noun for viduus is viduitas but viduertas, attested only here, has borrowed its suffixation (-ertat-) from ubertas ‘abundance, fertility’. Viduertas thus has two opposites in the poem: one expressed (vastitudo) and one unexpressed (ubertas); vastitudo is the complement to viduertas along the axis of real vs. potential; ubertas, along the axis of fertile vs. infertile.

¹³ In rural speech, according to Donatus, calamitas was a word for ‘hail’ (‘calamitatem rustici grandinem dicunt’, ad Ter. Eun. 79), which fits nicely with the probable original sense ‘blow’. Calamitas also had wider applications in rustic language, including robigo or ‘rust’, cf. TLL 3.118.61–119.2.

¹⁴ For that sense of intemperies, cf. ‘ex intemperie caeli, raptim mutatione in contrarium facta’ ‘because of the imbalance of the weather, which rapidly shifted from one kind to another’ (Liv. 5.13.4).

¹⁵ I have tried to explore this effect in ‘Binary phrases and the Middle Style as social code: Rhetorica ad Herennium 4.13 and 4.15’, Harvard Studies in Classical Philology 102 (2004), 237–74CrossRefGoogle Scholar.

¹⁶ For the use of phonetics as a linking device in Saturnians, see Goldberg, S., Epic in Republican Rome (1995)CrossRefGoogle Scholar, 65 n. 6, 76.

¹⁷ Euphorion, n. 275 b.c., composed epyllia which were an inspiration to Cornelius Gallus and Catullus; surviving fragments show the Alexandrian interest in aetiology, etymology, glosses, sentiment and amplification of detail.

¹⁸ An accessible summary, with an emphasis on poetic practice, is O'Hara, J., True Names (1996), 7–56Google Scholar. Ancient Latin etymologies are gathered by Maltby, R., A Lexicon of Ancient Latin Etymologies (1991)Google Scholar.

¹⁹ For example, genus divum humanos (with permissible hiatus at dipody break), immortalis humanos.

²⁰ Before Christian writers the words appear in close proximity chiefly in philosophical contexts, e.g. ‘exspectare immortalis mortalia membra’ ‘that immortal [beings] await mortal limbs’ (Lucr. 3.778), ‘homo ad immortalium cognitionem nimis mortalis est’ ‘Man is too bound to death to understand deathless things’ (Sen., Dial. 8.5.7), ‘quid enim immortale manus mortales fecerunt?’ ‘What immortal thing have mortal hands ever made?’ (Sen., Dial. 11.1.1).

²¹ That is, ‘mortal’ is meaningful only or chiefly in opposition to ‘immortal’ and vice versa.

²² Euripides, Hipp. 1396 (Artemis nearly cries for Hippolytus); Callimachus h. 6.17 (Demeter cries for Persephone); Verg., Aen. 1.228, 8.380 (Venus nearly cries for Aeneas), 10.464–5 (Hercules cries for Pallas; on this, see Feeney, D., The Gods in Epic (1991), 156CrossRefGoogle Scholar); Ovid, Am. 3.9.1 (Eos cries for Memnon; Thetis, for Achilles, cf. Aen. 8.383–4), Am. 3.9.46 (Venus perhaps cries for Tibullus), Met. 2.621–2 (Apollo almost cries for Coronis), Met. 13.689 (the nymphs cry for the Coronides Menippe and Metioche). The examples are gathered by F. Bömer, Die Fasten (1957), vol. 2 ad 4.251 and Die Metamorphosen (1969), vol. 5 ad 10.45.

²³ For other derivatives, see de Vaan, M., Etymological Dictionary of Latin and the other Italic Languages (2002)Google Scholar, s.v. Languages often distinguish ‘speak’, which emphasizes the action, from ‘say’, which emphasizes the result of the action (Fr. parler and dire, Ger. sprechen vs. sagen; the difference underlies the organizational scheme of Buck, C. D., ‘Words of speaking and saying in the Indo-European languages: first paper’, American Journal of Philology 36 no. 1 (1915), 1–18CrossRefGoogle Scholar). The outcomes of *bheH₂- have both senses in Latin: īnfāns ‘the one who cannot speak’; facundus ‘fluent’; but fāma ‘*thing said’, not ‘power of speech’; praefārī ‘say [a particular thing] first, make prefatory remarks’. In the sense ‘speak’ fārī was replaced by loquor and in the sense ‘say’ replaced or encroached upon, as the case may be, by dīcere (the latter probably an Italic development, since Oscan and Umbrian have the same word).

²⁴ fānum < *fasnom < *dhH₁sno-, with the ø-grade of the root appearing in fēriae ‘religious festival’ (*dheH₁s-io-), Oscan fíísnú |fsnɔ| ‘fanum’ (as if L. fēsna), Gk. θɛός ‘god’. Cf. de Vaan, op. cit. (n. 23) s.vv.

²⁵ cf. ‘dies fasti, per quos praetoribus omnia verba sine piaculo licet fari’ (Varro, L.L. 6.29), ‘fasti, in quibus ius fatur, id est dicitur’ (Isid., Orig. 6.18.1), cf. ‘fasti dies sunt in quibus ius fatur, id est dicitur, ut nefasti quibus non dicitur’, Isid., Nat. 1.4. For other etymological connections of fārī and related words, cf. Maltby, op. cit. (n. 18), s.vv. and the discussion of Bettini, M., ‘Weighty words, suspect speech: Fari in Roman culture’, Arethusa 41.2 (Spring 2008), 313–75CrossRefGoogle Scholar. All the etymologies are gathered in TLL 6.1.1029.35–75.

²⁶ e.g. ās, assis; mās, măris; vās, vāsis.

²⁷ cf. n. 31.

²⁸ Serv. gloss: iusti atque iniusti. The choice of gerund(ive)s has the additional benefit of implying morphologically what fas expresses lexically, namely necessity. The OLD and L&S list fandus as a separate lemma (i.e. unconnected to fas); Ernout, A. and Meillet, A., Dictionnaire étymologique de la langue latine (1985 ⁴)Google Scholar, ad loc. recognize the connection, as does the TLL (6.1.1032.7–19, cf. velut pro genetivo huius vocis [= fas] ponitur fandi, 6.1.287.83 and ff.). For comparable examples, cf. ‘omnia fanda nefanda malo permixta furore’ ‘everything holy and unholy roiled in wicked madness’ (Cat. 64.405), where the equivalence of fanda and nefanda to fas and nefas is suggested by close parallels in which the latter pair appears: ‘fas nefasque | confusura ruit’ ‘She rushes off, soon to mingle right and wrong’ (Ov., Met. 6.585–6); ‘respersae fando nefandoque sanguine arae’ ‘altars sprinkled with holy blood and unholy’ (Liv. 10.41.3); and ‘ad fas nefasque miscendum coorti sunt’ ‘They have arisen with the aim of mixing right and wrong’ (Sen., de Ira 2.9.2).

²⁹ The original *-to- participle, with correct ø-grade, survives in făteor and fătuus; fātum and fātus show a generalized full grade. A line of Ennius with fātum was probably meant as a figura etymologica: ‘neque me Apollo fatis fandis dementem invitam ciet me’ ‘and Apollo drives me, willingly mad, declaring dooms’ (61W). Cf. Verg., Aen. 1.261–2 (‘fabor enim … | longius et volvens fatorum arcana movebo’); fātum is the technical name for Sibylline prophecies.

³⁰ The TLL has it backwards: ‘vis prisca verbi elucet inde quod hic illic significat “loqui cum gravitate” “prophetari” … , quamquam non desunt loci ubi nihil significet nisi facultatem naturalem’ (6.1.1029.75–8). Cicero, de Orat. 3.153 lists effari as one of several antiquated or poetic words, along with tempestas = tempus, proles, suboles, nuncupare, rebar, opinabar; cf. also Quint., I.O. 8.3.27 ‘quaedam [sc. vetera] et necessario interim sumuntur, ut fari’. For a treatment of the interrelated cultural associations of fari and fas, see Bettini, op. cit. (n. 25).

³¹ This connection of fās to fārī could be correct. The original meaning of fās might have been ‘saying, utterance’, from *bheH₂os > PItal *faos > L fās. From the fixed phrase ‘[that] is the saying, [thus] it is said’ (fās est) — ‘Also sprach [der Gott]’, as it were — arose the familiar meanings of ‘lawful, divinely ordained’, etc. In the oblique cases *bheH₂-os would have made (to take the genitive as illustrative) *bheH₂-es-es > PItal *faases > L. ^xfāris, like *gen-os gen-es-os > gen-us gen-er-is.

³² So suggest searches in the Brepols LLT-A database for -Vre near -Vre(?t, m, s, ??s) (where V represents a, e, or i), and the appropriate searches for posse, esse, velle, malle, nolle.

³³ ‘If that weren't the case, I would be pleading with you to do whatever ill you could do to him’ (‘nam ni ita esset, tecum orarem ut ei quod posses mali | facere faceres’, Pl., Bacch. 554–5), ‘If I were at liberty to accuse, I would instead be accusing others that would enhance my position’ (‘nam si mihi liberet accusare, accusarem alios potius ex quibus possem crescere’, Cic., Rosc. Am. 83). Comparable is ‘unde sciret si quid sit scire nesciret’ (Aug., trin. 10.1.3). Echoes without the same syntax are ‘nisi se sentire sentiret’ (Aug., lib. arb. 2.4) and perhaps ‘hoc uno maestum, quod adire nequires’ (Sil. Ital., Pun. 13.455).

³⁴ Suerbaum, op. cit. (n. 5), 36 notes the logical flaw: ‘Die einfache Verknüpfung der beiden Gedanken “die Musen müßten eigentlich um Naevius weinen” und “die Römer … haben nach Naevius’ Tod nicht mehr Latein gesprochen” durch “und so” (oder schlimmer noch durch “deshalb”) ist unverständlich. Ein Zusammenhang und eine Begründung für das itaque ergibt sich erst dann, wenn man den Zwischendanken einschiebt “er war nämlich ein Dichter, von dem allein man mit Recht sagen konnte, er versteh es, Latein zu sprechen”. Auf diese Weise wird die Fortsetzung des Gedankens sinnvoll: “und so kommt es, daß man sich selbst in Rom nach seinem Tod nich mehr darauf verstanden hat, (wirkliches, gutes) Latein zu sprechen”.'

³⁵ Perhaps this contact between the divine and mortal words, implied by a phonetic figure, opens up a detail of the first line? A word break usually follows the fourth syllable of the first half of a Saturnian line, as in 1a. Here the last syllable before that word break is the same as the last syllable before the caesura, -īs: inmortalis | mortalis ||. Rhyme occurs at that position in Saturnians but, in attested examples, at any rate, only with co-referential lexemes (‘argenteo | polubro || aureo eclutro’ ‘[in] a silver basin [and] golden pitcher’, Liv. Andr., Od. 6W; ‘ferunt pulcras | creterras || aureas lepistas’, ‘They carry beautiful bowls, golden goblets’, Naev., Pun. 12W; ‘patrem suum | supremum || optumum appellat’, ‘She calls on her father, the highest [and] mightiest’, 16W; ‘onerariae | onustae || stabant in flustris’, ‘Freight-ships freighted stood in the calms’, 41W). In the Naevius epitaph, by contrast, the lexemes are from two different arguments, subject accusative (inmortalis) and object accusative (mortalis) — but that is not entirely clear until the second hemistich. Until 1b sets them in their place, for an instant — paradoxically, given their meanings — the forms cling together.

³⁶ For other such roots, cf. PIE *mm- > murmurāre, μορμύρω ‘roar and boil’; λαλέω ‘prate’, λαλαγέω ‘prate, babble, chirrup’; καυχάομαι ‘speak loud, boast’ (*χαυ-χαυ-); pīp(i)āre, pīpīre ‘chirp, cheep’; susurrus ‘humming, whispering’; balbus, balbutīre ‘stammer, stutter’, if from *b-b- (cf. de Vaan, op. cit. (n. 23), s.vv.).

³⁷ The instances of nēnia are gathered and sorted by Heller, J., ‘Nenia “παίγνιον”’, Transactions of the American Philological Association 74 (1943), 215–68CrossRefGoogle Scholar, who argues for an original sense ‘plaything’. Habinek, T., The World of Roman Song (2005), 234–8Google Scholar, following Arnobius, argues for an original meaning ‘end of the intestines’ (implying a semantic chain like ‘end of the intestines’ > ‘final sacrificial bit’ > ‘final funeral song’ > ‘other kinds of rhythmic speech’).

³⁸ If, in the speech of the composer, ī from *ei was still pronounced as (its probable intermediate stage before becoming ī) and ae had, even partially, smoothed to ē, as in rural dialects and eventually Vulgar Latin, then the line would be dominated by e-sounds: flērent dvē Camēnē Nēvium poētam. On the intermediate vowel  < *ei, see Sihler, A., New Comparative Grammar of Greek and Latin (1995)CrossRefGoogle Scholar, §57.2.b. For an example, cf. Lebro (CIL I² 381) = Līberō. That vowel still survived in rural dialects at the time of the Social War, if Cicero is any guide: ‘And that's why it seems to me that our Cotta — whose broad pronunciation you sometimes imitate, Sulpicius, getting rid of i’s and saying full e's — is more like farm workers (“reapers”) than the orators of old' (‘qua re Cotta noster, cuius tu illa lata, Sulpici, non numquam imitaris, ut Iota litteram tollas et E plenissimum dicas, non mihi oratores antiquos, sed messores videtur imitari’, de Orat. 3.46). As for -ae-, that it could be pronounced -ē- in casual urban speech in the late Republic is suggested by the pedagogi of Rhet. Her. 4.14, a deliberately somewhat colloquial passage which also has oriculas for auriculas, aures.

³⁹ Nonius Marcellus, 145, 24M: ‘nenia ineptum et inconditum carmen, quod a conducta muliere quae praefica diceretur, is, quibus propinqui non essent, mortuis exhiberetur’ citing Varro, de Vita Populi Romani 4: ‘ibi a muliere quae optuma uoce esset perquam laudari dein neniam cantari solitam ad tibias et fides’.

⁴⁰ The combination is nowhere else in archaic poetry (comedy included) and virtually absent even in prose (Cic., Tusc. 2.26; Liv. 1.54.5, 9.45.11).

⁴¹ Famously on a Scipio epitaph: ‘Annos gnatus XX is l[oc]eis m[an]datus’ ‘Twenty years old handed over to the places [below]’ (CIL I² 11 l. 5, with Mommsen's restoration loceis = locis infernis; Ritschl has [div]eis ‘to the gods’). There is a pun with mandatus in another sense in the following line: ‘ne quairatis honore | quei minus sit mandatus’ ‘Seek ye not, why no honour was entrusted to him’ (with honore[m] in prolepsis).

⁴² ‘θησαυρός is a strongbox, where the sanctuary's money was kept. … Excavated θησαυροί are heavy stone monuments, usually provided with locks. An example with the word θησαυρός inscribed upon it was found in Athens, in its second use, during the demolition of a house. … It is dated on epigraphical grounds to the early 4th century b.c., and was used for the collection of a money donation (ἀπαρχή) connected, in this case, with a wedding (προτέλɛια γάμος). It is not a portable box (κιβωτός), but a heavy, stone monument that weighs 1,472 kg and locks with a key. We might therefore conclude that the difference between a θησαυρός and a κιβωτός is chiefly one of size and weight’, Adrianou, D., ‘Late Classical and Hellenistic furniture and furnishings in the epigraphical record’, Hesperia 75 (2006), 561–84CrossRefGoogle Scholar. That sense is recognized for this passage by E. Forcellini, Totius Latinitatis Lexicon (1858–), p. 725 col. A s.v., under the heading ‘latiore sensu est locus, in quo aliquid servandum reponitur, apotheca, conditorium, ripostiglio [closet]’.

⁴³ cf. supra n. 2.

⁴⁴ ‘Scripsere alii rem | vorsibus quos olim Fauni vatesque canebant’ ‘Others have treated the topic in the metre the Fauni and seers once used to sing’ (Ann. 231–33W).

⁴⁵ For general discussion, see Skutsch, O., ‘Enniana, I’, CQ 38 no. 3/4 (1944), 79–86CrossRefGoogle Scholar.

⁴⁶ ‘Insece Musa manu Romanorum induperator | quod quisque in bello gessit cum rege Philippo’ ‘Queath, Muse, what each commander of the Romans achieved by force in the war with King Philip’ (Ann. 322–3W) is, as argues Hinds, S., Allusion and Intertext: Dynamics of Appropriation in Roman Poetry (1998), 58–9Google Scholar, a correction of the opening of Andronicus' Odusseia: ‘virum mihi, Camena, insece versutum’ ‘Queath me, Camena, the versatile man’. For the numerous disputes regarding another line with Musae, ‘Musas quas memorant nosces nos esse <Camenas>’ ‘Thou wilt know that we Camenae are those whom they call Muses’ (Enn. var. 43W), cf. Suerbaum, op. cit. (n. 5), 347–9; Hinds, loc. cit.

⁴⁷ ‘For general reasons it seems unlikely that before Ennius’ time the Romans were altogether unfamiliar with the name of Musae. Evidence, however, to show that they knew it is not available. Nevertheless, even if the name Musae was known, in invoking their patronage for his poem Ennius is making a departure from the habits of earlier poetry, of which he is both conscious and proud. In the beginning of the seventh book he speaks with scorn of his predecessor Naevius: “scripsere alii rem | versibus quos olim Fauni vatesque canebant | cum neque Musarum scopulos …” Now the reign of the Muses begins, and the grave virus of the Saturnian line is driven out by the munditiae of the hexameter’ (Skutsch, op. cit. (n. 45), 79).

⁴⁸ Suerbaum, op. cit. (n. 5), 304–8 and Courtney, op. cit. (n. 2), 50 note some of the precedents repeated here.

⁴⁹ καὶ δ' αὐταὶ στοναχɛῦντι σὺν ɛὐϕόρμιγγι Λυκɛίῳ | ἔρρηξαν Μοῦσαι δάκρυα Πιɛρίδɛς | μυρόμɛναι τὸν ἀοιδόν (7.10.5–7).

⁵⁰ Μοῦσαι δ' ἐκλαύσαντο (7.412.5).

⁵¹ ‘There is a hint [in thesauro] at the traditional etymologies of Dis as diues and Πλούτων as πλούσιος’, Courtney, op. cit. (n. 2), 48. Dīs for dīves is not a folk etymology but the genuine article. Dīs was back-formed as a nominative to the genitive dītis, contracted regularly from dīvitis. The Dies- of Diespiter is formally unrelated, coming from Italic *dijē-, one stem of the PIE *diēus seen in Iūpiter (cf. de Vaan s.v.). Dispiter for Diespiter is attested only in Paul. ex Fest. Πλούτων is also borrowed directly as Plūtō, but not before Cicero.

⁵² So Jocelyn, H. D., The Tragedies of Ennius (1967), 331Google Scholar. The connection was made explictly by Ennius: ‘Pluto Latine est Dis pater, alii Orcum vocant’ (var. 31f. W).

⁵³ The exchange is sometimes taken as evidence of a feud between the Metelli and Naevius but is unlikely to represent a historical incident accurately. Cf. Goldberg, op. cit. (n. 16), 33–6; Gruen, E. S., Studies in Greek Culture and Roman Policy (1996), 96–106Google Scholar.

⁵⁴ cf. Goldberg, op. cit. (n. 16), 35 and n. 16.

⁵⁵ For discussion and recent bibliography, see Kruschwitz, P., Carmina Saturnia epigraphica: Einleitung, Text und Kommentar zu den Saturnischen Versinschriften (2002), 108–15Google Scholar.

⁵⁶ Thus ‘honc oino ploirume cosentiont R[omai] | duonoro optumo fuise viro | Luciom Scipione’ (L. Cornelius Scipio cos. 259; CIL I² 8, 9 = Bücheler CLE 6), ‘consol censor aidilis quei fuit apud vos’ (L. Cornelius Scipio Barbatus, cos. 298; CIL I² 6, 7 = Bücheler CLE 6), ‘hunc unum plurimae consentiunt gentes | populi primarium fuisse virum’ (A. Atilius Calatinus, cos. 258, 254; FPL 7).

⁵⁷ ‘Est ut viro vir latius ordinet | arbusta sulcis, hic generosior | descendat in campum petitor, | moribus hic meliorque fama | contendat, illi turba clientium | sit maior: aequa lege Necessitas | sortitur insignis et imos, | omne capax movet urna nomen’ (3.1.9–16).

⁵⁸ Goldberg, op. cit. (n. 16), 163.

⁵⁹ Liv. 26.19.5, Gellius 6.1.6, de vir. illustr. 49.3.

⁶⁰ On the Scipio legend generally and the likelihood of his temple visits, cf. Scullard, H. H., Scipio Africanus: Soldier and Politician (1970), 18–23Google Scholar.

⁶¹ Livy appreciated Scipio's attempt to shape a public persona: in his analysis, not only was Scipio remarkable because of the virtues he really did possess, but he also carefully contrived to put them on display (‘arte … quadam ab iuventa in ostentationem earum [sc. virtutum] compositus’, 26.19.3). Livy attributes the credence given to legends of Scipio's divine birth to this habit of self-display (26.19.6–7).

⁶² The incident is much discussed; see the references in Walbank, F. W., A Historical Commentary on Polybius, vol. 2 (1967), 196Google Scholar; Scullard, op. cit. (n. 60), 19.

⁶³ There is perhaps a much earlier parallel in Numa's alleged contacts with Egeria, but it is hard to know how the gesture would have been parsed; Livy's report is sceptical: ‘qui [sc. Numa] cum descendere ad animos sine aliquo commento miraculi non posset, simulat sibi cum dea Egeria congressus nocturnos esse; eius se monitu, quae acceptissima diis essent, sacra instituere, sacerdotes suos cuique deorum praeficere’ (Liv. 1.19.5). Cf. ‘quo quia se persaepe Numa sine arbitris velut ad congressum deae inferebat, Camenis eum lucum sacravit, quod earum ibi concilia cum coniuge sua Egeria essent’ (1.21.3); Cic., Leg. 1.4.

⁶⁴ Goldberg, op. cit. (n. 16); Hinds, op. cit. (n. 46), 52–98.

⁶⁵ For a discussion of the aesthetics of the Saturnian, Goldberg, op. cit. (n. 16), 58–82.

⁶⁶ van Sickle, op. cit. (n. 10).

⁶⁷ cf. also Hinds, op. cit. (n. 46), 60: ‘Is it really self-evident that it marks an advance in Hellenizing innovation to transliterate a Greek goddess into the Roman alphabet rather than to seek an Italian cultural analogue to render her in her new context?’

⁶⁸ cf. Waszink, J. H., ‘Camena’, C&M 17 (1956), 139–48Google Scholar. The Camenae, properly, were goddesses of a spring, meadow and grove below the Caelian hill (cf. Suerbaum, op. cit. (n. 5), 32–3; 303–4 with lit.). One of them, Egeria, was the nymph who was said to have inspired King Numa (Liv. 1.19.5, 1.21.3).

⁶⁹ cf. e.g. Goldberg, op. cit. (n. 16), 162–3.

⁷⁰ Skutsch, op. cit. (n. 45), 79.

⁷¹ On the invention of grave epigrams, see Dahlmann, op. cit. (n. 6), 65–100, who compiles evidence for the practice of concluding the vita of a literary figure with such an epigram, some of which are plainly literary fictions.