¹ For a useful collection of essays on this topic, cf. Most, G. (ed.), Collecting Fragments/Fragmente sammeln (Göttingen, 1997).Google Scholar

² A particularly noteworthy example is provided by Ucciardello, G., ‘Nuovi frammenti di oratori attici nel Vat. gr. 7’, in Miscellanea Bibliothecae Apostolicae Vaticanae XIV (Vatican City, 2007), 431–82.Google Scholar

³ Diller, A., The Textual Tradition of Strabo's Geography (Amsterdam, 1975), 50Google Scholar. Diller had already published one of these excerpts in ‘A new source on the Spartan ephebia’, AJPh 62 (1941), 499–501.Google Scholar

⁴ The restoration is dated to the fourteenth century at the latest by Allen, T.W., ‘MSS of Strabo at Paris and Eton’, CQ 9 (1915), 15–26CrossRefGoogle Scholar, at 21. Against the attribution of the restoration to Maximus Planudes or, at any rate, to the Planudean milieu (Sbordone, Lasserre), cf. Diller (n. 3), 47 n. 5.

⁵ Cf. Diller (n. 3), 50 for the quotation, and 50–1 for a description of some of this material. For a recent concise description of the manuscript, cf. Radt, S.L., Strabons Geographika, Band 1, Prolegomena. Buch I–IV: Text und Übersetzung (Göttingen, 2002), vii–viii.Google Scholar

⁶ In the manuscript they are followed by a text (in the same hand) of a completely different kind (lines 6–13, not mentioned by Diller), which almost exactly coincides with the first chapter of the work περὶ ϲχημάτων, printed (after Iriarte and Walz) in the third volume of Spengel, L.'s Rhetores Graeci (Leipzig, 1856)Google Scholar, 105 (but the title and the author's name are absent in the excerpt) to 106.14 (corresponding to about a third of the whole treatise as transmitted), attributed to a Polybius of Sardis. On this author, long thought to be of uncertain date (cf. Wendel, C., RE 21.2 [Stuttgart, 1952], 1580–1Google Scholar, s.v. ‘Polybios. 14’), but recently identified with a Polybius mentioned in some 2nd-century inscriptions from Sardis, cf. Jones, C.P., ‘Polybius of Sardis’, CPh 91 (1996), 247–53.Google Scholar

⁷ The beginning of the first three lines has suffered extensive damage from the humidity and is extremely difficult to read. My transcription is based both on an earlier autopsy of the original and, later, on a photographic reproduction. I am grateful to Tiziano Dorandi and Christian Förstel, who double-checked some dubious point for me on the original, using a Wood's lamp, in June 2005.

⁸ This division with the double lambda at the beginning of the line is not problematic in a text of this date: cf. Crönert, G., Memoria Graeca Herculanensis cum titulorum Aegypti papyrorum codicum denique testimoniis comparata proposuit G.C. (Leipzig, 1903), 11–18Google Scholar (particularly at 16, where this treatment of geminate consonants in late manuscripts is defined ‘nec iam vitium sed aetatis consuetudo’).

⁹ For the reading of the uncertain, very damaged words following Ἕλλησιν and preceding γραϕόμ(ενον?) I am indebted to Tiziano Dorandi and Christian Förstel.

¹⁰ The context renders the completion of the ending of the abbreviated word uncertain: the accent, though, is preserved in the manuscript.

¹¹ The last superscript letter has a circumflex accent above it and the context suggest that an iota must have been meant.

¹² Antimachus' quotation, differently from the other quotations in this and in the following excerpt, is marked out by six downward facing superscript brackets.

¹³ The word is abbreviated with a superscript ending: the last letter looks more like an upsilon than as a sigma; the preceding one is fairly tallish round letter. Perhaps ου was meant.

¹⁴ Slater, W. J. (ed.), Aristophanis Byzantii Fragmenta (Berlin and New York, 1986), 72–7.CrossRefGoogle Scholar

¹⁵ Nauck's correction is supported by the reading of the parallel sources edited under Aelius Dionysius μ 6 and 7 Erbse, also based on Aristophanes' lexicographical work: cf. Slater (n. 14), 82–3.

¹⁶ Slater (n. 14), 75.

¹⁷ One of these words may lurk in the damaged portion of the text at the beginning of line 2, but I have not been able to make sense of what I can make out of the writing.

¹⁸ Cf. also Etym. Magn. 84. 24–6, ἀμμά· ἡ τροϕὸc καὶ ἡ μήτηρ, κατὰ ὑποκόριcμα. καὶ ἣ Ῥέα δὲ λέγεται καὶ ἀμμὰc λέγεται καὶ ἀμμία, where the wording makes it just possible, but far from necessary, to understand that ἀμμία too was used as an epithet of Rhea. It is more plausible that a stronger punctuation is required after ἀμμάc.

¹⁹ Cf. Schmidt, V., Sprachliche Untersuchungen zu Herondas (Berlin, 1968), 19–20Google Scholar, Di Gregorio, L., ‘La figura di Metriche nel primo mimiambo di Eronda’, in Studia classica Iohanni Tarditi oblata (Milan, 1995), 675–94, at 686.Google Scholar

²⁰ In his text, as transmitted by Athenaeus, Didymus attributes to Polemon the opinion that the word among the Laconians meant also δεῖπνον, as it did among the other Dorians. This contradicts Polemon's own wording, as transmitted by Athenaeus a few paragraphs earlier.

²¹ Cf. Nafissi, M., La nascita del kosmos (Naples, 1991), 181–4.Google Scholar

²² Its occurrences in literary texts and documentary papyri are all very late (late antique or Byzantine), with the possible exception of Aristophanes of Byzantium, fr. 7 Slater, where the word is used as an explanation of ϲτεγανόμιον in Paris. Suppl. Gr. 1164 (fourteenth century). It is far from certain, however, that the wording may go back to Aristophanes himself. It is already attested, anyway, in two second-century inscriptions found at Panamara (IStratonikeia 17 and 270, cf. Pantel, P. Schmitt, La cité au banquet [Rome, 1992], 314–15Google Scholar).

²³ In this case, clearly, we are dealing with the instrumental meaning of the suffix in –τήριον (see, for example, θελκτήριον) rather than with the local one implied in all the other attested occurrences of the word (I owe this point to a remark of CQ's anonymous referee).

²⁴ We should not take it for granted that Alcman would have been considered by ancient grammarians to be an example of ‘Laconian’ dialect (see A.C. Cassio, ‘Alcman's text, spoken Laconian, and Greek study of Greek dialects’, in I. Hajnal (ed.), Die altgriechischen Dialekte: Wesen und Werden (Innsbruck, 2007), 29–45: I am grateful to L. Prauscello for drawing my attention to this publication), but he certainly did provide evidence for the use of a word in Laconian, as implied by his quotation in the context of the polemics of Didymus against Polemon at Ath. 4.140c–d.

²⁵ The search was updated on the online database on 08/07/2011.

²⁶ Cf. Diod. Sic. 2.53.7, and sch. Nic. Ther. 533: ῥάδικεϲ κυρίωϲ μὲν οἱ ἀπὸ ϕοινίκων κλάδοι, νῦν δὲ καταχρηϲτικῶϲ κέϲρηται τῆι λέξει. It is never interpreted as ‘root’ in any of its occurrences in Nicander (our verse and Alex. 57, 331, 528 s. v. l., Ther. 533). On Latin radix and Greek ῥάδιξ, cf. A. Ernout, ‘Senex et les formations en -K- du latin’, in Philologica <1> (Paris, 1946), 133–63, at 150–2.

²⁷ Cf. W. Morel, ‘Iologica’, Philologus 83 n. F. 37 (1928), 345–89, at 354–5.

²⁸ Some of Nicander's manuscripts offer the same wrong μὴν as the addition supra lineam of our excerpt (γε μὴν p μὴν V, according to Schneider's apparatus: not mentioned in J.-M. Jacques' apparatus, Nicandre. Oeuvres. Tome 2, Les Thériaques. Fragments iologiques antérieurs à Nicandre [Paris, 2002]).

²⁹ Cf. sch. c ad v. 523: ̓Ιόλαοc (Ἀπολλᾶc Preller, perhaps rightly: cf. Jacoby ad 266 III A Komm. [Leiden, 1943], 200; second half of the third century b.c.e.) δὲ ἐν τῶι περὶ Πελοποννηϲιακῶν πόλεων τὸ πήγανον ὑπὸ Πελοποννηϲίων ῥυτὴν καλεῖϲθαί ϕηϲιν. It is defined more specifically as Spartan in Eutecnius' paraphrase.

³⁰ Three occurrences are in Aretaeus, who also has ϕάρ= far, on which more below.

³¹ Cf. W. Haas, Die Fragmente des Grammatiker Tyrannion und Diokles (Berlin and New York, 1977), 98 and 176–7, E. Rawson, Intellectual Life in the Late Roman Republic (London, 1985), 69.

³² Cf. E. Gabba, ‘Il latino come dialetto greco’, in Miscellanea di studi alessandrini in memoria di A. Rostagni (Turin, 1963), 188–94, M. Dubuisson, ‘Le latin est-il une langue barbare?’, Ktema 9 (1984), 55–68, D. Briquel, Les Pélasges en Italie: Recherches sur l'histoire de la légende (Rome, 1984), 441–57, J. Werner, ‘Περὶ τῆς ῾Ρωμαϊκῆς διαλέκτου ὅτι ἐστὶν ἐκ τῆς ῾Ελληνικῆς’, in E.G. Schmidt, Griechenland und Rom. Vergleichende Untersuchungen zu Entwicklungstendenzen und – höhepunkten der antiken Geschichte, Kunst und Literatur (Tbilisi, Erlangen and Jena, 1996), 323–33, B. Stevens, ‘Aeolism: Latin as a dialect of Greek’, CJ 102 (2006/7), 115–44, with previous bibliography, P. Ascheri, ‘The Greek origins of the Romans and the Roman origins of Homer in the Homeric scholia and POxy. 3710’, in F. Montanari and L. Pagani (edd.), From Scholars to Scholia. Chapters in the History of Ancient Greek Scholarship (Berlin, 2011), 65–86. For a list of the grammarians who dealt, in various ways, with the issue, cf. the third chapter of B. Rochette, Le latin dans le monde grec. Recherches sur la diffusion de la langue et des lettres latines dans les provinces hellénophones de l'Empire romain (Brussels, 1997).

³³ Cf. Ch. Theodoridis, Die Fragmente des Grammatikers Philoxenos (Berlin and New York, 1976) 12 and frr. 311–29: R. Giomini, ‘Il grammatico Filosseno e la derivazione del Latino dall'Eolico’, PP 8 (1953), 365–80; J.L. Heller, ‘Nepos “σκορπιστής” and Philoxenus’, TAPhA 93 (1962), 61–89. A collection of his fragments, assembled by F. Razzetti, is available online at the website of the Lessico dei Grammatici Greci Antichi (www.aristarchus.unige.it/lgga).

³⁴ Cf. frr. 324, 325, 328, 329 and the fr. dub. 681, where a more common etymological system is used.

³⁵ Cf. Giomini (n. 33), 373–6, Heller (n. 33), 88–9. Cf. e.g. also the case of fr. 312, where the Greek word ἁρμόϲ is discussed but no Latin term is actually adduced. The comparison between a Greek word connected to the same root, ἅρμα, and Latin arma actually occurs in other lexicographical texts (where Philoxenus is not quoted, and which are not included in Theodoridis' edition) such as Choeroboscus' Epimerismi in Psalmos (whence it also reached the Etymologicum Gudianum and other later compilations), which draws also on an epitome of Orus who, in his turn, extensively used Philoxenus (on the relation between these works, cf. Theodoridis [n. 33], 72–3). It is very unlikely that a discussion of the word ἁρμόϲ might have occurred in a work on Latin without its being compared to a Latin word, and it seems to me more than likely that the texts mentioned above also derive ultimately from Philoxenus (on other passages related to this in the Homeric scholia see now also Ascheri [n. 32], 73–4). For the comparison between ̔Ρωμαῖοι and Ἕλληνεc (as in the first section of our excerpt), cf. fr. 328 Theodoridis ( ̔Ρωμαῖοι, ὥϲπερ καὶ Ἕλληνεc …).

³⁶ The Index auctorum in Theodoridis (n. 33), 408–410 (which includes, however, also authors quoted by later sources along with Philoxenus) lists, among others, Alcman (3 quotations), Antimachus (5 quotations), Callimachus (19 quotations), Epicharmus (5 quotations), and the Theriaca of Nicander (3 quotations).

³⁷ Philoxenus mostly uses general labels such as Aeolic, Dorian and Ionian. A problematic case is fr. 567 Theodoridis where Philoxenus quotes the use of ἦδοc for ὄξοc as peculiar to the Cyrenaeans while according to Ath. 2.67c it was an Attic usage, and according to sch. Pl. Phd. 72b (50, p. 29, in D. Cufalo, Scholia graeca in Platonem I. Scholia ad dialogos tetralogiarum I–VII continens [Rome, 2007]) a Chalcidian one (part of the same scholium, though without the digression on the dialectal use of ἦδοc, occurs also in Ps.-Didymus, who attributes it to Boethus: it is *F3 in A.R. Dyck, ‘Notes on Platonic lexicography in antiquity’, HSPh 89 (1985), at 78–9). Hsch. γ 30 has γᾶδοc· … ἄλλοι ὄξοc (a Laconian form), and α 1173 attributes ἇδοc to the Syracusan Deinolochus. The Syracusan glossa πᾶϲ (fr. 266 Theodoridis) may theoretically derive from Epicharmus, Deinolochus or Sophron. For the references to the various dialects in ancient grammarians, cf. Cassio, A.C., ‘Parlate locali, dialetti delle stirpi e fonti letterarie nei grammatici greci’, in Crespo, E., Ramon, J.L. Garcia, Striano, A. (edd.), Dialectologica Graeca (Madrid, 1993), 73–90Google Scholar and Cassio (n. 24).

³⁸ Cf. Roller, D.W., The World of Juba II and Kleopatra Selene: Royal Scholarship on Rome's African Frontier (New York and London, 2003), 170–3Google Scholar; his work seems to have dealt mainly with ‘the origin of Latin historical and cultic terms, and perhaps details of material culture, with a particular emphasis on the similarities between Latin and Greek’ (172).

³⁹ The Greco-Latin Hermeneumata included in G. Goetz's Corpus Glossariorum Latinorum, for example, which do not make use of such rare poetic texts, have a different focus and belong to an altogether different tradition (on which, see the diverging views of Dionisotti, A.C., ‘From Ausonius' schooldays? A schoolbook and its relatives’, JRS 72 [1982], 83–125Google Scholar and Kramer, J., Glossaria bilinguia altera [C. Gloss. Biling. II], Archiv für Papyrusforschung Beiheft 8 [Munich and Leipzig, 2001])CrossRefGoogle Scholar. For more recent contributions on these and similar texts, cf. Ferri, R., ‘Textual and linguistic notes on the Hermeneumata Celtis and the Corpus Glossariorum’, CQ NS 60 (2010), 238–42CrossRefGoogle Scholar; Dickey, E. and Ferri, R., ‘A new edition of the Latin-Greek glossary on P.Sorb. inv. 2069 (verso)’, ZPE 175 (2010), 177–87Google Scholar, Dickey, E., ‘The creation of Latin teaching materials in antiquity: a re-interpretation of P.Sorb. inv. 2069’, ZPE 175 (2010), 188–208Google Scholar (all with further bibliography).

⁴⁰ For a general overview on ancient reflections on the comparison between the two languages, cf. Schöpsdau, K., ‘Vergleiche zwischen Lateinisch und Griechisch in der antiken Sprachwissenschaft’, in Müller, C.W., Sier, K. and Werner, J. (edd.), Zum Umgang mit fremden Sprachen in der griechisch-römischen Antike (Stuttgart, 1992), 115–36.Google Scholar

⁴¹ Cf. e.g. the contribution on some aspects of this problem in Bianconi, D., ‘Le traduzioni in greco dei testi latini’, in Cavallo, G. (ed.), Lo spazio letterario del medioevo, 3. Le culture circostanti, 1. La cultura bizantina (Rome, 2004), 519–68Google Scholar (and, more particularly, at 554–64 on the Planudean period).

⁴² Another case difficult to date, in which more than one layer of philological efforts may be stratified, is the discussion of the word κηκάϲ found in Etym. Genuin. s.v., where Callimachus (fr. 656 Pf., fr. 132 Massimilla) and Nicander (Alex. 185, but introduced as ἐν Θηριακοῖϲ) are quoted, and where a parallel with Latin caecus is explicitly drawn: cf. Massimilla, G., ‘I primi due libri degli Αἴτια di Callimaco nell'Etymologicum Genuinum’, SIFC 83 (1990), 180–91, at 189.Google Scholar

⁴³ An argument actually against Philoxenus could be that he seems to have argued that Latin was basically an Aeolic dialect (cf. fr. 323 Theodoridis), while our source, in the second section, makes use of one word defined as ‘Laconian’. In the preserved fragments attributed to this work accepted in Theodoridis' edition, however, Philoxenus never quotes Aeolic forms (nor, for that matter, any other dialectally marked forms). For a further, possibly weightier argument against this attribution, see below, n. 48.

⁴⁴ When medical authors of the Imperial and later periods, such as Aretaeus, Soranus, Aetius and Alexander of Tralles use the Greek word ϕάρ in the sense of ‘spelt’, this should obviously be understood as a Latin loanword. Particularly interesting, for our case, is Dioscorides, Materia medica 2.96 (p. 173 Wellmann), where the following equivalence is provided χόνδροc· οἱ δὲ δίκοκκον, οἱ δὲ ζέαν καλοῦϲιν, Ῥωμαῖοι ϕάρρεμ. Galen, De compositione medicamentorum secundum locos 13.257.3 Kühn, has the genitive ϕάρρουϲ. The Latin loanword is also used by Dion. Hal. Ant. Rom. 2.25, in order to explain the technical term confarreatio.

⁴⁵ Cf., in particular, Lobeck, C.A., Aglaophamus, sive de theologiae mysticae graecorum causis libri tres, 2 (Königsberg, 1829)Google Scholar, 866 n. [u]; Paralipomena grammaticae graecae, Pars prior (Leipzig, 1837)Google Scholar, ‘De nominibus monosyllabis’, 74; Pathologiae sermonis graeci prolegomena (Leipzig, 1843), 62Google Scholar; ΡΗΜΑΤΙΚΟΝ sive verborum graecorum et nominum verbalium technologia, (Königsberg, 1846), 304Google Scholar; Pathologiae graeci sermonis elementa, 2 (Königsberg, 1862)Google Scholar, (Diss. VIII, de apocope), 291; Kuhn, H., ‘Ablaut, a und Altertumskunde’, Zeitschrift für vergleichende Sprachforschung 71 (1954), 129–61, at 145Google Scholar; Frisk, H., Griechisches étymologisches Wörterbuch 2 (Heidelberg, 1970)Google Scholar, 1011 s.v. ϕῆρον; Chantraine, P., Dictionnaire etymologique de la langue grecque. Histoire des mots (Paris, 1968–80)Google Scholar, Tome 4.2 (Paris, 1980), ed. J. Taillardat, 1196–7.

⁴⁶ Cf. Wilson, W.W., ‘The partheneion of Alkman (fragment 23 Bergk)’, AJPh 33 (1912), 57–67Google Scholar, at 62–3; Pavese, C.O., Il grande partenio di Alcmane (Amsterdam, 1992), 77–8.Google Scholar

⁴⁷ Comparison with forms attested in Italic, Slavic and Germanic languages suggests a derivation of Latin far from *bhars-. In Greek this would not, as a rule, produce a form such as either ϕάροc or ϕᾶροc, as the sequence *V̆rsV remains unchanged in most dialects, while in Attic (and occasionally elsewhere) it evolves into V̆rrV. It produces V̄rV only in the forms of the sigmatic aorist and in few other dubious cases. Cf. Lejeune, M., Phonétique historique du Mycénien et du Grec ancien (Paris, 1972), 124–6.Google Scholar

⁴⁸ This kind of derivation, based on the subtraction of the initial phi, is compatible with Philoxenus' method, and could tell against the attribution to him of the different etymology of our excerpt. As a matter of fact, Philoxenus seems to have derived the similar words ϕαρόϲ and ἀϕαρόϲ (as predecessors of the adjective ἀϕαυρόϲ) from the monosyllabic root ϕῶ, through various stages of changes and additions (fr. 55 Theodoridis) and fr. 617 Theodoridis suggests a link between, among others, ϕέρω and ϕάροc.

⁴⁹ On the two alternative interpretations see Hollis, A., Callimachus Hecale (Oxford, 2009 ²), 294–5.Google Scholar

⁵⁰ The first edition is that of Dindorf, W., in Grammatici Graeci 1 (Leipzig, 1823), 36–7Google Scholar for this passage. The edition of the whole work was greatly improved by Lehrs, K., Herodiani Scripta tria emendatioria (Königsberg, 1848)Google Scholar, 138, § 36.31–37.2 for this passage. Cf. also Lentz, A., Herodiani technici reliquiae, in Grammatici Graeci 3.2.2 (Leipzig, 1870), 942.14–16.Google Scholar

⁵¹ The reference is to what is now line 61 of the Louvre partheneion: see below, p. 646.

⁵² The reading of the last word (εὕρηται) was confirmed by Egenolff, P., ‘In Herodianum technicum’, RhM 35 (1880), 98–104Google Scholar, at 103, while Wyss, B., Antimachi Colophonii Reliquiae (Berlin, 1936)Google Scholar still presents it as a conjecture of Stoll for the transmitted εἴρηται (which actually was a mistaken reading present in the editions of Dindorf, Lehrs and Lentz). Wyss prints the last two words attributed to Antimachus between cruces, while Matthews accepts the correction of the transmitted χατεύουcιν in χατέουcιν (attributed to Lehrs, but already present in Dübner's edition, first published in 1840).

⁵³ Cf. Wyss (n. 52): xxxv ‘[f]oedissimam illam κατὰ τέταρτον τροχαῖον τομήν sibi non indulsisse Antimachum sponte intellegitur’. This reconstruction is defended by Deroy, L., ‘Les mésaventures d'un archaïsme grec: ΦΑΡΟΣ “terre cultivée, champ”’, AntCl 54 (1985), 49–55Google Scholar, who does not even realize it would entail a major metrical anomaly. Deroy is tentatively followed by Matthews, V.J., Antimachus of Colophon. Text and Commentary. Mnemosyne Suppl. 155 (Leiden, New York and Cologne, 1996)Google Scholar, 358 (commentary), who does not address its metrical implications in a satisfactory way.

⁵⁴ Apart from Bergk, who reconstructs the sequence as part of a pentameter: ἀεὶ ϕάρεοc <γῆν> χατέουcιν ἔχων. On Deroy and Matthews, see above (n. 53).

⁵⁵ Here is a (probably exhaustive) list of the corrections proposed: Ἐχίων Dübner [‘Κάββαλεν ὅπλ᾿ ἐπ᾿ ἀγροῖϲι ϕάρεοc χατέουcιν Ἐχίων, collato Ov. Met. 3.127 sua jecit humi monitu Tritonidos arma (sc. Spartus Echion)’], ἑκόντεc Lehrs (followed by Lentz), χατέουcαν Ἐχῖνον Lobeck, ἐχετλῶν O. Schneider, <γῆν> χατέουcιν ἔχων Bergk, ἔχοντεc Matthews (as an alternative to keeping ἔχων with Deroy).

⁵⁶ Conceivably, perhaps, even the transmitted ἔχων could be kept, if a masculine noun such as ϲτίχοc was meant.

⁵⁷ Commentarium III ad Arist. Soph. el. 180a9, 76. It is a twelfth- or thirteenth-century text: cf. S. Ebbesen, Commentators and Commentaries on Aristotle's Sophistici Elenchi: A Study of Post-Aristotelian Ancient and Medieval Writings on Fallacies. Corpus Latinum Commentariorum in Aristotelem Graecorum 7, vol. 2, Greek Texts and Fragments of ‘Alexander's’ Commentary (Leiden, 1981)Google Scholar, 231. On Commentarium III, id. vol. 1, The Greek Tradition (Leiden, 1981), 290–301.Google Scholar

⁵⁸ According to Lehrs ‘[v]identur fuisse qui vellent scribi ϕάρευc’ (i.e. instead of ϕάρεοc), but this would involve a long scansion of the first syllable, against Herodian's explicit statement that in this part of the text he is dealing with the meanings of the word with a short alpha. It is more likely that the issue regards the fact that the quotation was syntactically incomplete.

⁵⁹ ‘[U]t Byzantini alicuius interpolationem secl. Maas’.

⁶⁰ Cf. 3.2.414, 8 and 545, 6; in Cath. pros. 3.2.63, 22 (cf. 3.1.292, 25): οὕτωc ἔχει ἡ τῶν ἀντιγράϕων παράδοcιc, τῶν ἀντιγράϕων is Lehrs' conjecture for the transmitted τοῦ Ἀντιϕάνουc.

⁶¹ The metre could be restored also by deleting the preposition ἐν, but this would hardly improve the text from other points of view.

⁶² I have no explanation for the omicron superscript on the ending in the excerpt.

⁶³ The form is apparently attested in Antimachus fr. 116.6 Matthews.

⁶⁴ Matthews accepts the hiatus here. In other cases (listed by Matthews, loc. cit.) the hiatus can be explained on ‘etymological’ grounds.

⁶⁵ Cf. Matthews (n. 53), 61.

⁶⁶ Cf. Campbell, M., ‘Hiatus in Apollonius Rhodius’, in Fantuzzi, M. and Pretagostini, R. (edd.), Struttura e storia dell'esametro greco (Rome, 1995), 193–220Google Scholar, at 203–4. Two lacunose quotations in the Cairo commentary (fr. 104.2–3 Matthews) have suggested even the possibility of hiatus after a short vowel (in which case ἐν πενταπόληϊ ἀεί would theoretically be a further possible alternative). The Cairo text, however, is too damaged in that section for us to be able to draw any firm conclusion from it. For a masculine caesura followed by spondaic word and without bucolic diaeresis, cf. fr. 22.3 Matthews.

⁶⁷ Contra Deroy (n. 53), who thinks that in both passages the word meant ‘cloak’, and Matthews (n. 53), 356 according to whom ‘it is surely more natural to assume that Alcman is cited for one meaning [i.e. ‘plough’] and Antimachus for the other [i.e. ‘cloak’]'. Matthews's main argument is that the quotation from Antimachus is introduced by the words ἀλλὰ καὶ παρ᾽ Ἀντιμάχωι, and that this would imply that ‘Herodian is moving from ϕάροϲ meaning “plough” in Alcman to the meaning “robe” in Antimachus’. This has no weight at all, as in most of the other occurrences of the sequence in Herodian it introduces a new source for the same usage (cf. 3.2 p. 629.10; 3.2 p. 748.17; 3.2 p. 812.16 in Lentz's edition). This is attested in other grammarians too: cf. Apollonius Dyscolus, De pronominibus p. 66.8, Pollux 10.39.5 and, to quote just an example from a different corpus, sch. Aesch. Sept. 101: ἀλλὰ καὶ παρ᾽ Εὐριπίδηι (introducing a second example of a dochmiac sequence).

⁶⁸ The Louvre papyrus, dated to the first century b.c.e., is well anterior to Herodian. For a reproduction and a transcription of the relevant portion of the papyrus, cf. Turner, E.G., Greek Manuscripts of the Ancient World, ed. Parsons, P.J.. BICS Suppl. 48 (London, 1987 ²), 44–5Google Scholar, with the addenda and corrigenda at 148. There is no point, in this context, in quoting the huge and largely inconclusive bibliography on the modern interpretation of the word in Alcman (a recent, but far from exhaustive discussion of the word is that of Priestley, J.M., ‘The ϕαροs of Alcman's Partheneion 1’, Mnemosyne 60 [2007], 175–95Google Scholar).

⁶⁹ A further one is mentioned by a single source (sch. Pind. Pae. 4.13) for Ceos, otherwise usually known as a Tetrapolis.

⁷⁰ C.A.G. Schellenberg, Antimachi Colophonii Reliquiae (Halle, 1786), 80–1, fr. xxxiv. After advancing the hypothesis that the fragment might have dealt with the story of Coronis and Ischys, Schellenberg adds, at 81: ‘an potius de Thessalis, Dotio relicto, in Cnidum insulam olim profectis … Haec quidem coniectura, si ulla in talibus hariolatio esse potest, melior videatur priore, cum illorum amorum tristis exitus omni fugae valde adversetur’. Wyss and Matthews ad loc. attribute the idea to the later edition of H.W. Stoll (Antimachi Colophonii Reliquias explanavit Henr. Gui. Stoll [Dillenburg, 1845], 75 ‘[f]ortasse de Thessalis, qui Dotio relicto in Cnidum insulam et Rhodum et Symen (Muell. Orchom. p. 195.) profecti sunt, dicuntur’).

⁷¹ Piccirilli, L., ΜΕΓΑΡΙΚΑ. Testimonianze e frammenti (Pisa, 1975), 43–6Google Scholar thinks that he could be as late as the third or second century b.c.e.

⁷² The source is Ath. 6.262e. The story is followed by that of the shipwreck of Phorbas and his sister Parthenia, and of their arrival at Ialysus, but no explanation of the content of the curses is given. For an interpretation of the aition of Phorbas (but with no discussion of the Curse Islands), cf. Robertson, N., ‘The ritual background of the Erysichthon story’, AJPh 105 (1984), 369–408.Google Scholar

⁷³ The source is Steph. Byz. α 374 Billerbeck. We should perhaps read Ἀραιαί with Casaubon. Berkel's objection, as reported by Billerbeck, that Dieuchidas ‘den Namen periphrastisch mit dem Ethnikon τῶν Ἀραιῶν νήϲων gebildet habe’, does not correspond to what Athenaeus actually transmits.

⁷⁴ When the curses were first uttered the group of cities would have formed a Hexapolis rather than a Pentapolis. It is the ritual practice, which involved the Pentapolis, that must have caused this anachronism in the aition. FGrH 444 F 2 might provide a clue of the possible cultic background: Ἡ τῶν Δωριέων πεντάπολιϲ, Λίνδοϲ, Ἰάλυϲοϲ, Κάμειροϲ, Κῶϲ, Κνίδοϲ. Ἄγεται δὲ κοινῇ ὑπὸ τῶν Δωριέων ἀγὼν ἐν Τριοπίῳ Νύμϕαιϲ, Ἀπόλλωνι, Ποϲειδῶνι. Καλεῖται δὲ Δώριοϲ ὁ ἀγὼν, ὡϲ Ἀριϲτείδηϲ ϕηϲί.

⁷⁵ Diller (n. 3), 50 describes their content as ‘quite miscellaneous: philosophical, grammatical, lexical, recipes, proverbs, anecdotes etc.’.

⁷⁶ Herodian is the source of the excerpt on the Spartan ephebia, published by Diller himself in his 1941 article (cf. above [n. 3]; in his 1974 book [n. 1] Diller, at 51 n. 15, quotes some of the bibliography on this text), and of the one anticipated by Diller at 51 n. 16, and fully published by D'Alessio, G.B. in ‘Le Ὧραι e le πέμϕιγεc: fr. 43.40–41 Pf. (= fr. 50 M.)’, in Bastianini, G., Casanova, A. (edd.), Callimaco. Cent'anni di papiri. Atti del convegno internazionale di studi. Firenze, 9–10 giugno 2005 (Florence, 2006), 101–17.Google Scholar

⁷⁷ Cf. Diller (n. 3), 50. For fairly recent updates regarding some aspects of the multifarious transmission of this work, cf. Nickau, K., ‘Zur Geschichte der griechischen Synonymica: Ptolemaios und die Epitoma Laurentiana’, Hermes 118 (1990), 253–6Google Scholar and ‘Schiffbruch in der Wüste der Sinai: Zu Herennios Philon, Neilos von Ankyra und dem Ammonioslexikon’, Hermes 128 (2000), 218–26Google Scholar (several versions of this and other lexica of synonyms have been published over the years by Nickau himself and by V. Palmieri); Villani, E., ‘Un nuovo lessico polisemantico bizantino (Ambr. C 222 inf., ff. 210v–212r)’, Aevum 78 (2004), 441–515Google Scholar, with further bibliography. If I am not mistaken, the Strabo excerpts have not been used by the scholars who have published on this issue.

⁷⁸ Among the ones for which no source is identified Diller (n. 3), 51 also mentions an excerpt which refers to Anticlides' Nostoi (FGrH 140), as well as the one discussed here.

⁷⁹ It may be useful to remark that, if the excerpts reflect the original arrangement of their source, a possible alphabetical sequence (limited to the first letter) would be conceivable only starting from the Latin terms. Our text, however, has been far too much abbreviated, and I do not think we can draw any positive conclusion from this feature.

⁸⁰ Cf., for example, and from different points of view, Dubuisson (n. 32); Rawson (n. 31), 55–6; Gruen, E., Culture and National Identity in Republican Rome (London, 1992), 234–5Google Scholar; Farrell, J., Latin Language and Latin Culture: From Ancient to Modern Times (Cambridge, 2001), 38–9CrossRefGoogle Scholar; Stevens (n. 32), 131–9 (with previous bibliography); Ascheri (n. 32).

⁸¹ Famously, but obviously coincidentally, the name of the town had been used by the proponents of the Aeolic theory as an example of the lost digamma: see e.g. Dion. Hal. Ant. Rom. 1.20.30 and 1.90.1 and Gabba, E., Dionysius and The History of Archaic Rome (Berkeley, Los Angeles and Oxford, 1991), 138–9Google Scholar. The poet Statius uses the older Greek form Hyele at Silv. 5.3.127.

⁸² On Statius the Elder see McNelis, C., ‘Greek grammarians and Roman society during the Early Empire: Statius' father and his contemporaries’, ClAnt 21 (2002), 67–94Google Scholar, at 69. For the careers pursued by his students, cf. Silv. 5.3.185–90 and McNelis, 90.

⁸³ Cf. Cassio (n. 37), 85; McNelis (n. 82), 83–6 on grammatical work on Greek dialects in the early Imperial period and 88–9 on the role these dialects may have played for Latin-speaking members of the elite.