¹ While the default meaning of παῖς is ‘boy’, without more context it cannot be established whether δαλιοχεῖν is gender specific, or whether it could equally be used of intercourse with a girl. Furthermore, while the pederastic sense of δαλιοχεῖν is clear from the gloss, we cannot exclude the possibility of pederastic incest (on which cf. n. 26 below), even though we might expect the article (τῶι/τῆι παιδί) in such a context. For συνεῖναι, see Henderson, J., The Maculate Muse ² (Oxford, 1991), 159Google Scholar. I am grateful to the anonymous reader for CQ, whose comments forced me to clarify several important points.

² Cf. Dion. Thrax, GG I/1 5–6 Uhlig, for whom γλῶσσαι were a subset of γραμματική, the purpose of which was the ἐμπειρία τῶν παρὰ ποιηταῖς τε καὶ συγγραφεῦσιν ὡς ἐπὶ τὸ πολὺ λεγομένων, ‘the empirical knowledge of that which is said for the most part by poets and prose writers’; for the interpretation of Dionysius’ words, and the variant preserved at Sext. Emp. Math. 1.57, see Pagani, L., ‘Pioneers of grammar: Hellenistic scholarship and the study of language’, in Montanari, F. and Pagani, L. (edd.), From Scholars to Scholia: Chapters in the History of Ancient Greek Scholarship (Berlin and New York, 2011), 14–64Google Scholar, at 18. There was properly something of a distinction between glossography, an aspect of grammar, and dialectography, which was in some ways an aspect of ethnography. Yet, the two traditions were complementary; thus the glossographical tradition took an ethnographic approach in categorizing attributions by city or region; cf. Σ. Dion. Thrax, GG I/3 302–3 Uhlig (attributed to Trypho) Δωρὶς γὰρ διάλεκτος μία, ὑφ’ ἥν εἰσι γλῶσσαι πολλαί, Ἀργείων, Λακώνων, Συρακουσίων, Μεσσηνίων, Κορινθίων (‘Doric is one dialect, of which there are many γλῶσσαι [sub-dialects or regional variations], Argive, Laconian, Syracusian, Messenian, Corinthian’). On ancient glossography, see Latte, K., ‘Glossographika’, Philologus 80 (1925), 136–75Google Scholar; for the practice of dialectology in antiquity, see Hainsworth, J.B., ‘Greek views of Greek dialectology’, TPhS 66 (no volume number in print edition) (1967), 62–76Google Scholar, who notes that ‘Hellenistic glossography was … overwhelmed by mythology, history and literature, and … had little impact on the conception of the dialects’ (72); see also Cassio, A.C., ‘Parlate locali, dialetti delle stirpi e fonti letterarie nel grammatici greci’, in Crespo, E., Ramón, J.L. García and Striano, A. (edd.), Dialectologia Graeca. Actas del II coloquio internacional de dialectología griega (Madrid, 1993), 73–90Google Scholar. Pfeiffer, R., History of Classical Scholarship: From the Beginnings to the End of the Hellenistic Age (Oxford, 1968), 202Google Scholar expresses a high regard for Aristophanes of Byzantium's glossographical work, which he imagines went beyond ‘mere bookishness’; yet, we can feel confident that his research did not begin to encompass fieldwork in any recognizable modern sense. For a recent detailed treatment of linguistic science in antiquity, see Pagani (this note).

³ The likelihood of Athanadas’ History of Ambracia (FGrHist 303 F 1, mentioned only at Anton. Lib. 4.5) containing dialect glosses, particularly ones of a sexual nature, is extremely slim (though cf. the case of Derkyllus at n. 41 below).

⁴ For Corinthian and the Saronic dialects generally, see Buck, C.D., The Greek Dialects ² (Chicago, 1955), 164–6Google Scholar and 293–7; Schmidt, R., Einführung in die griechischen Dialekte (Darmstadt, 1977), 34–9Google Scholar; Colvin, S., A Historical Greek Reader: Mycenaean to the Koiné (Oxford, 2007), 142–50Google Scholar. Early inscriptions from Ambracia are woefully scant; for the evidence of the polyandrion elegy (SEG 41.540A), see below with nn. 33–5.

⁵ For the polis, see Hansen, M.H. and Nielsen, T.H., An Inventory of Archaic and Classical Poleis (Oxford, 2004), 354–6Google Scholar, no. 113; D. Strauch in Der neue Pauly s.v. Ambrakia. Callimachus related the myth of the death of the Ambracian tyrant Phalaecus early in Book 3 of the Aetia (Phalaecus Ambraciotis), frr. 62a–c Harder; cf. Gallazzi, C. and Lehnus, L., ‘Due nuovi frammenti delle Diegeseis. Approssimazioni al III libro degli Aitia di Callimaco’, ZPE 137 (2001), 7–18Google Scholar.

⁶ See Hartwig, A., ‘The evolution of comedy in the fourth century’, in Csapo, E., Goette, H.R., Green, J.R., Wilson, P. (edd.), Greek Theatre in the Fourth Century B.C. (Berlin and New York, 2014), 207–27Google Scholar, at 214–15.

⁷ T 1.49–51 K.–A. = Proleg. de com. XXVIII 54–5 Koster ἔγραψε κωμωιδίαν τινὰ Κώκαλον, ἐν ὧι εἰσάγει φθορὰν καὶ ἀναγνωρισμὸν καὶ τἆλλα πάντα, ἃ ἐζήλωσε Μένανδρος, ‘he wrote a play called Cocalus, in which he introduced rape and recognition scenes and everything else that Menander emulated’.

⁸ Satyrus’ Life of Euripides fr. 39 col. 7 = P.Oxy. 1176; see Schorn, S., Satyros aus Kallatis. Sammlung der Fragmente mit Kommentar (Basel, 2004)Google Scholar.

⁹ Nesselrath, H.G., ‘Parody and later Greek comedy’, HSPh 95 (1993), 181–95Google Scholar.

¹⁰ Cf. Hunter, R., ‘The comic chorus in the fourth century’, ZPE 36 (1979), 23–38Google Scholar, at 35 n. 61.

¹¹ IGUR 218 = CIG I 230 = IG XIV 1098; editio princeps in Oderico, G.A., De Marmorea Didascalia in Urbe Reperta (Rome, 1777)Google Scholar. See Bergk, T., ‘Verzeichniss der Siege dramatischer Dichter in Athen’, RhM 34 (1879), 292–330Google Scholar, at 328–30; Capps, E., ‘The Roman fragments of Athenian comic didascaliae’, CPh 1 (1906), 201–20Google Scholar; Luppe, W., ‘Zu den griechischen didaskalischen Inschriften in Rom’, ZPE 8 (1971), 123–8Google Scholar.

¹² The words that follow Ἀμπρακιώτιδι in line 3, γʹ ἐν [ἄστει, introduce the following list of titles, and do not refer to Ἀμπρακιῶτις.

¹³ See Chantraine, P., Dictionnaire étymologique de la langue grecque (Paris, 1968–1980), 165Google Scholar and Frisk, H., Griechisches etymologisches Wörterbuch (Heidelberg, 1960–1972), 220Google Scholar s.v βαρδῆν. Amongst those who suggest a non-Greek source are Blumenthal, A., ‘Illyrisches und Makedonisches’, IF 49 (1931), 169–83Google Scholar, at 178–9, who suggests IE *b^her- (now reconstructed as *b^herH–), ‘work with a sharp tool, cleave’; Pisani, V., ‘Zur Sprachgeschichte des alten Italiens’, RhM 97 (1954), 47–68Google Scholar, at 62 n. 14 opts for *b^her–, ‘carry’; the Diccionario Griego-Español suggests a connection with the root *mer– (i.e. *merh ₂–/*mr̥h ₂–), ‘grab forcefully, crush’. One might also suggest *b^herh ₂–, ‘move quickly’. ‘Alles hypothetisch’, as Frisk (this note) observes.

¹⁴ Beekes, R.S.P., Greek Etymological Dictionary (Leiden and Boston, 2010)Google Scholar, 1.127.

¹⁵ See Chantraine (n. 13), 105 s.v. ἄρδα; Bechtel, F., Die griechischen Dialekte (Berlin, 1921–1924)Google Scholar, 2.282. β for ϝ occurs already in fourth-century inscriptions, representing an underlying fricative pronunciation /f/, /v/, for which see Buck (n. 4), 47; ϝ for β occurs in an early inscription from Corinth, IG IV 212 = CEG 360 (c.510–500). Beekes's statement (n. 14), 1.127 that a connection of ἄρδα with ἄρδω, ‘to irrigate’, ‘is impossible because ἄρδα has a short initial ἀ–’ is nonsensical. The second syllable of the acc. sg. ἄρδαν is scanned short at Pherecrates, fr. 58, while Herodian tells us that the initial ἀ– of ἄρδω is long (GG 3/I 522.9, cf. 3/II 109.15; see Chantraine [n. 13], s.v.), but that is the sum of our knowledge.

¹⁶ Phot. Lex. a 2796 Theodoridis ἄρδα καὶ ἄρδαλος· τὸ μόλυσμα. καὶ ἀρδαλῶσαι τὸ μολῦναι ἀσχημόνως (‘ἄρδα καὶ ἄρδαλος: a taint/impurity. And ἀρδαλῶσαι means to defile in an obscene manner’); Galen 19.85.11 Kühn καὶ ἀρδαλῶσαι τὸ ῥυπᾶναι (‘ἀρδαλῶσαι means to defile’). Hesychius’ glosses δ 259 Latte δάρδα· μόλυσμα and δ 261 Latte δαρδαίνει· μολύνει (cf. also α 4795 Latte) are likely to be corruptions of an unfamiliar *ϝάρδα, *ϝαρδαίνει, perhaps on analogy with the Homeric δαρδάπτω, reduplicated from δάπτω. While δ is not a readily expected graphic or phonological development of ϝ (β and υ are more common, and cf. Masson, O., ‘Sur la notation occasionnelle du digamma grec par d'autres consonnes et la glose macédonienne abroûtes’, BSL 90 [1995], 231–9Google Scholar), A. Willi, BMCR 2011.02.40 n. 3 has identified other plausible instances of δ for ϝ in Hesychius. We hardly have ‘substrate’ interchange of δ/ø, as Beekes (n. 14), 1.127 muses. Another crumb might be provided by Hesychius’ enigmatic gloss ε 33 Latte ἐαρδάλη· ἐπλησίασεν, where we might again have corruption of an original *ϝαρδάλη, while the sexual sense of πλησιάζω (for which, see LSJ s.v.) is well established.

¹⁷ That is, is the word to be divided as βαρδ–ίσαγνος or βαρδί–σαγνος or βαρδίσ–αγνος.

¹⁸ On the status of initial /w/ in Corinthian and Ambraciot in the early fourth century, as well as the treatment of the thematic infinitives in βαρδῆν and δαλιοχεῖν, see further below.

¹⁹ We would typically expect –άζω in a denominative formation to an a-stem, although the use of –ίζω with consonant stems (e.g. ἑλληνίζω ← Ἕλλην, Dor. Ἑλλᾶν) could have led to analogical remodelling, particularly in a case such as βαρδίσαγνος which, it is suggested below, was likely an ad hoc formation.

²⁰ On the many and fascinating uses for (and associations of) the chaste-tree in ancient Greece, see generally Staden, H. von, ‘Spiderwoman and the Chaste Tree: the semantics of matter’, Configurations 1 (1993), 23–56CrossRefGoogle Scholar, with specific reference to the Thesmophoria and the ‘chaste’ associations of the tree at 37–41; see also Versnal, H.S., ‘The festival for Bona Dea and the Thesmophoria’, G&R 39 (1992), 31–55Google Scholar, at 35; Lloyd-Jones, H., ‘Alcaeus fr. 130 B, 1–2 Voigt’, ZPE 108 (1995), 35–7Google Scholar.

²¹ Cf. Plin. HN 24.59 Graeci lygon uocant, alias agnon, quoniam matronae Thesmophoriis Atheniensium castitatem custodientes his foliis cubitus sibi sterunt, ‘the Greeks call it (the uitex) lygos, otherwise agnos, because the wives of the Athenians, guarding their chastity at the Thesmophoria, spread its leaves on their beds’.

²² Thus rightly Pavese, C.O., ‘δαλιοχός e δάλλω’, Glotta 72 (1994), 75–81Google Scholar. For the formation as ὀχέω, cf. ὀχέωνται at Aratus, Phaen. 1070. Whether the formal source of ὀχεύω is ἐχώ < *seĝ^h– or ἐχώ < *ṷeg^h– is debated, although it seems evident that the semantic range of both ὀχεύω and ὀχέω have fallen together under the influence of the highly productive ἐχώ < *seĝ^h–.

²³ Similarly, βαίνω, a word usually reserved for animal copulation, can also be used of humans (cf. Henderson [n. 1], 155; for the compound εἰσαναβαίνω, see Watkins, C., ‘The golden bowl: thoughts on the New Sappho and its Asianic background’, ClAnt 26 [2002], 305–25Google Scholar, at 311–15). Pollux, in his extensive list of verbs of sexual congress at 5.92–3, lists both ὀχεύω and βαίνω under the heading of τὸ μίγνυσθαι ἐπὶ τῶν ἀλόγων, ‘copulation amongst animals’.

²⁴ I assume a derivation from *delh ₁–, ‘split, cleave’, cognate with Latin dolēre; cf. Frisk (n. 13), s.v. δηλέομαι and Hackstein, O., Die Sprachform der homerischen Epen (Wiesbaden, 2002), 219–20Google Scholar (the first laryngeal, and thus the connection with dolēre, is suggested by the reflex δάλερος discussed in n. 25). Pavese (n. 22) tries to connect δαλι– with θάλλω < *d^halh ₁– (or *d^heh ₂lH– for those who do not believe in PIE *a), ‘grow, flourish’. The semantic connection is unproblematic, given the root's connections with fertility and fecundity. However, *d^h– invariably yields the voiceless aspirate θ in Greek, and thus Pavese (n. 22), observing the Albanian reflex dal, must postulate a borrowing scenario from Illyrian or Macedonian. Pavese (n. 22) also tries to connect the lemma δάλλω in Hesychius (δ 160 Latte), glossed ἡ ἀπόπληκτος. οἱ δὲ τὴν ἔξωρον παρθένον, ἢ γυναῖκα πρεσβυτέραν, ὅταν συμπαίζηι ταῖς παρθένοις. ὑπερῆλιξ (‘a senseless/crippled woman; otherwise an unmarried woman past her prime, or an older woman, when she frolics around with young girls; past his/her prime’; for the image cf. Anaxandrides, fr. 67 K.–A. ὑπερήμεροί μοι τῶν γάμων αἱ παρθένοι, ‘my daughters are past their prime for marriage’). Hesychius’ glosses, however, unambiguously support the connection with δαλός, ‘ember’, < δαίω.

²⁵ Cf. φρενοδαλής, ‘mind-destroying’, at Aesch. Eum. 330 and πανδάλητος, ‘completely ruined’, at Hipponax, fr. 4 IEG ², for which see Masson, O., Les fragments du poète Hipponax (Paris, 1962)Google Scholar, ad loc. Following Pavese (n. 22), we likely have a Caland-like formation in δαλιοχός (the Caland system is a complex set of rules governing IE nominal derivation in which, most basically, adjectives with a suffix in *–ro– substitute an *–i– when forming the first member of compounds, and an *–es– when forming the second member of compounds; cf. e.g. κυδρός ~ κυδιάνειρα ~ ἐρικυδής; see generally Rau, J., Indo-European Nominal Morphology: The Decads and the Caland System [Innsbruck, 2009], 65–186Google Scholar). Compare δαλιοχός with φρενοδαλής in Aesch. Eum. 330 (above). This would presuppose a form with a –ρο– suffix, e.g. δάλερος. Happily, such a form is attested as a hapax in the MSS of a fragment of Empedocles quoted at Plut. Quaest. conv. 663A = 31 B 90.2 DK = D 68 Laks and Most, although in the text of DK and Laks and Most the word has been replaced by Diels's conjecture *δαερός (unattested). This is unnecessary. Far more likely is that δάλερος, a cognate of δηλέομαι, has converged with the many cognates of δαίω that contain a –lo– suffix (e.g. δαλός, ‘ember’), a convergence facilitated by the complementary senses of ‘destructive’ and ‘burning’. The short –ᾰ– in φρενοδᾰλής and δᾰ́λερος (both guaranteed by metre), and likely in πανδάλητος and δαλιοχός, is readily explained via a zero grade *dl̥h ₁–V– of the root *delh ₁–; cf. Hackstein (n. 24), 219–20, who suggests that the long –ᾱ– (η) of δηλέομαι is a secondary development within Greek. (More precisely, the short –ᾰ– in δᾰ́λερος specifically is likely analogical with predicted Caland formations in *–i–, as well as with possible verbal formations such as *dl̥h ₁–ei̯^e/ _o– [cf. καλέω < *kl̥h ₁– ei̯^e/ _o–], as the predicted outcome of a Caland form *dl̥h ₁–ro– would be ^×dlē–ro– [cf. λῆνος < *ṷl̥h ₂–neh ₂–; κλητός < *kl̥h ₁–to–]). For the compound noun formed from two underlying verbal roots, cf. e.g. Homeric ὀλοοίτροχος, ‘a roller, a rolling rock’, with the first member either cognate with εἴλεω < *ṷel– (i.e. *ṷel– 2 in LIV), or from ὀλοός < *h ₃elh₁–ṷo– (see G.P. Shipp, Studies in the Language of Homer ² [Cambridge, 1972], 122–3 and Janko on Il. 13.136–8), and the second element deverbative from τρέχω. The agent noun δαλιοχός is rightly accented on the ultima; cf. τόμος, ‘a cutting’ (nomina actionis) / τομός, ‘a cutter’ (nomina agentis).

²⁶ Pederasty is a common element in comedy of all periods, and requires no comment here. Incest was one of the taboos that even comedy did not revel in portraying too much, yet it is referenced for one effect or another in all periods of Attic comedy. Strepsiades expresses horror at Euripides’ depiction of incest in the Aeolus (Nub. 1371–6; cf. Aeschylus’ words at Ran. 1194 with Dover), while the subject is glossed over at Vesp. 1178. Niceratus adduces incest as amongst the worst sexual crimes at Men. Sam. 495–7. Teleclides, fr. 52 παιδοφίλης παιδέρως Ζεύς is presumably a reference to Zeus's rape of Ganymede. Cratinus, fr. 311 describes incest in language strikingly similar to Hesychius’ glossing of δαλιοχεῖν: ‘to lie down with one's father’, συγκαθεύδειν τῶι πατρί. See further D. Clay, ‘Unspeakable words in Greek tragedy’, AJPh 103 (1982), 277–98.

²⁷ See, above all else, Colvin, S., Dialect in Aristophanes (Oxford, 1999)Google Scholar, as well as Arnott and Hunter on the passages from Alexis and Eubulus cited immediately below.

²⁸ For dialect and non-Athenian speech in the fragmentary poets of Old Comedy, see Colvin, S., ‘The language of non-Athenians in Old Comedy’, in Harvey, D. and Wilkins, J. (edd.), The Rivals of Aristophanes (London, 2000), 285–98Google Scholar.

²⁹ Though in this case somewhat unsuccessfully. In καὶ λεπαστά μ’ ἁδύοινος εὐφρανεῖ δι’ ἁμέρας (‘a limpet-cup of sweet wine will cheer me through the day’) we might expect ϝαδύϝοινος (μ’ ἁδύ– guaranteed by metre; for ϝ– < *sw– in Cretan cf. ϝέκαστος < *swe–kn̥– and ϝός < *swo– in the Gortyn Law Code, IC IV 123–71, and see Bile, M., Le dialecte crétois ancien [Paris, 1988], 113–15Google Scholar) and ἀμέρας. The dialect is neither Cretan nor Laconian, and might reflect only the sort of Doric colouring found in choral lyric. Another instance of indeterminate Doric occurs at Philyllius, fr. 10.

³⁰ A regular feature of Syllabic Cypriot, e.g. po–to–li–(se/ne/wi) in the Idalion Bronze, no. 217 in Masson, O., Les inscriptions chypriote syllabiques (Paris, 1961)Google Scholar; see Egetmeyer, M., Le dialecte grec ancien de Chypre (Berlin and New York, 2010)CrossRefGoogle Scholar, 1.199, §217.

³¹ Cf. Colvin (n. 28), 291 on τρηχύς at Eupolis, fr. 341 and Ar. Pax 1086; Millis, B.W., Anaxandrides: Introduction, Translation, Commentary (Fragmenta Comica Bd. 17) (Heidelberg, 2015)Google Scholar, on fr. 45 suspects epic parody here. Note that καὶ in the grammarians and scholia is often used simply to introduce a quotation or an author's name, and thus there is likely no copulative sense in κεῖται δὲ καὶ παρὰ κωμικῶι Ἀναξανδρίδηι.

³² It should also be noted that there are no non-Athenians in Old Comedy who do not speak in a regional dialect.

³³ On the inscription generally, see Cassio, A., ‘I distici del polyandrion di Ambracia e l’“io anonimo” nell'epigramma greco’, SMEA 33 (1994), 101–17Google Scholar; Athanassoudi, I., ‘Du bon usage d'un texte poétique par le dialectologue: le cas des distiques du polyandrion’, Verbum 18 (1995/1996 [1997]), 217–25Google Scholar; del Bario Vega, M.L., ‘Observaciones lingüísticas sobre el polyandrion de Ambracia’, in Gabaudan, F. Cortés and Dosuna, J.V. Méndez (edd.), Dic mihi, Musa, uirum. Homenaje al Profesor Antonio López Eire (Salamanca, 2010), 47–54Google Scholar. The communis opinio is that the dialect is essentially Corinthian, although most Corinthianisms are graphic (e.g. use of a distinct grapheme [commonly transliterated as E] to represent the monophthongization of the diphthong *ey, and AE to represent *ay [καΕ and πολῖταΕ in line 9, and κατέχΕ for κατέχει in line 8]; cf. Athanassoudi [this note], 220–2). The ‘mild’ Doric of Corinthian is closer to the literary Doric of Stesichorus and Pindar than Laconian or Cretan is, and—graphic representation aside—there is nothing in the polyandrion inscription that could not be found in Doric lyric.

³⁴ Note for example that Pindar will use both ξένος and ξεῖνος metri causa.

³⁵ An instructive comparandum is CEG 145, a roughly contemporary funerary verse inscription from the nearby Corinthian colony of Corcyra, in which we find such non-epichoric forms as ἐπ’ Ἀράθθοιο ρhοϝαῖσι (a-stem dat. pl. in –αισι[ν] are alien to Corinthian and Doric generally), as well as plentiful digammas in e.g. στονόϝεσσαν ἀϝυτάν. Another Corcyran inscription of the mid sixth century, SEG 146, provides us with Ξενϝάρεος.

³⁶ The handbooks are not overly consistent with respect to the status of wau in Saronic, and the older ones tend not to differentiate between inscriptions from the colonies and inscriptions from the mother city. Thus for Corinthian Buck (n. 4), 164 states that digamma is retained in early inscriptions in all positions; Schmidt (n. 4), 38 implies retention in initial position, with hεκατεροι in a late sixth-century inscription from Selinous as an exception; most recently, Colvin (n. 4), 45 records that it is lost early in Corinthian, with the exception of post-consonantal digamma, while he notes its retention in initial position in fifth-century Corinthian inscriptions from Corcyra (Colvin [n. 4], 144) and in Megarian from Selinous (Colvin [n. 4], 147).

³⁷ See Colvin, S., ‘Aristophanes: dialect and textual criticism’, Mnemosyne 48 (1995), 34–47CrossRefGoogle Scholar, at 41–2. At 156, Attic παριδών would not fit the metre. The question of the interpretation of the manuscript evidence for dialect forms that are not guaranteed by metre is of course a complex one, for which see again Colvin (this note).

³⁸ Or ‘On the Dialect’, Περὶ τῆς Ἑλλήνων διαλέκτου καὶ Ἀργείων καὶ Ἱμεραίων καὶ Ῥηγίνων καὶ Δωριέων καὶ Συρακουσίων at τ 1115 Adler.

³⁹ Hainsworth (n. 2), 74.

⁴⁰ Cf. Hainsworth (n. 2), 71.

⁴¹ Cf. e.g. Etym. Magn. 391.12 on the Argive historian Derkyllus (FGrHist 288), who evidently represented the loss of intervocalic /s/ (an areal feature in Argolic and Laconian) in his work (ὡς ἐπὶ τοῦ ποιῆσαι ποιῆἁι κ.τ.λ. κέχρηται τούτωι τῶι εἴδει τῆς δασείας καὶ Δέρκυλλος, ‘for example, ποιῆἁι in place of ποιῆσαι, etc.; Derkyllus uses this type of aspiration’). We may doubt, however, that this was a regular feature of his prose style, but was rather noted in passing in his Argolica. See generally Cassio, A.C., ‘Storiografia locale di Argo e dorico letterario: Agia, Dercillo, e il Pap. Soc. Ital. 1091’, RFIC 117 (1989), 257–75Google Scholar.