1 Wünsch, R. (ed.), Defixionum Tabellae, IG III iii [Appendix] (Berlin 1897)Google Scholar. This volume was never absorbed into IG II², but was reprinted by Ares Publishers, with different pagination, in Inscriptiones Atticae, Supplementum Inscriptionarum Atticarum, I (Chicago 1976)Google Scholar. Sixty-one of them (i.e., most of the better-preserved ones), including some mere lists of names, are reprinted with translation in Eidinow, E., Oracles, curses and risk among the ancient Greeks (Oxford 2007) catalogue pp. 352–454 (352-86)CrossRefGoogle Scholar.

2 Curbera, J., “From the magician’s workshop: notes on the materiality of Greek curse-tablets,” in Boschung, D. and Bremmer, J. N. (edd.), The materiality of magic (Munich 2014)Google Scholar.

3 The text reads here: … Aves Nocturnae, Aves Harpyiae, Ortygiae …; Ortygia(e) is unexampled in this context. If we read aves with Ortygiae they would not, of course, be quails but, parallel to the Harpies (cf. Verg., Aen. 3.216: uirginei uolucrum vultus) = ‘Ortygian (birds)’. I know of no reference to such birds in the Underworld prior to Silius Italicus (Pun. 13.597-600), who describes the dirae volucres, including vultures, bubones, striges and Harpies, sitting on a yew-tree by the river Cocytus. This would fit neatly with the date assigned to our text. Ortygia is regularly Delos, which makes no obvious sense here, but we may have a direct (if mistaken) allusion to Ovid, Met. 5.639 f. (the nymph Arethusa speaking): Delia rupit humum caecisque ego mersa cauernis / advehor Ortygiam ([Artemis] clove the earth and I fell into its dark bowels and [so] reached Delos). Alternatives might be a false memory of the word ὠγύγιον (‘primeval’) applied to the river Styx by Hesiod (Theog. 805); cf. Seneca, Oed. 589: populi pestis Ogygii malum (‘pestilence, the doom of the Ogygian people’); Statius, Theb. 11.420: Ogygios … manis (‘the spirits of the dead in Hades’); or a mere confusion with Stygiae, scil. deae or sorores, the ‘Stygian goddesses/sisters’ = the Parcae (cf. Lucan, BCiv 9.837; Statius, Theb. 5.156 [152]; 10.833 [829];11.576 [574]).

4 Bevilacqua, G. and Colacicchi, O., “Una nuova defixio latina della via Ostiense,” NSA 9 17–18 (2006–2007) 303–49Google Scholar = AE 2007. 260 Google Scholar, with a French transl. (her subsequent commentary, Bevilacqua, G., “Aurora, Orchi soror,” ParPass 64 [2009] 46–67 Google Scholar, seems to me rather unbalanced, since it stresses disproportionately the significance of Aurora in the economy of the whole).

5 The fourth is the well-publicised discovery in 2000 of small lead cylinders in the cistern of the Fountain of Anna Perenna, north of the Aurelian Wall, containing poppets; the texts are Late Roman and, generally speaking, hard to read and of minor importance; see, e.g., Piranomonte, M. (ed.), Il santuario della musica. Il bosco sacro di Anna Perenna (site guide; Rome 2002)Google Scholar; ead., “La fontana sacra di Anna Perenna a Piazza Euclide: tra religione e magia,” MHNH. Revista Int. de Investigación sobre Magia y Astrología Antiguas 5 (2005) 87–104 Google Scholar; ead., “Anna Perenna a dieci anni dalla scoperta. Un riepilogo e un aggiornamento,” MHNH 9 (2009) 251–64Google Scholar; ead., “Religion and magic at Rome: the Fountain of Anna Perenna,” in Gordon, R. L. and Simón, F. Marco (edd.), Magical practice in the Latin West. Conf. Zaragoza (RGRW 168; Leiden 2010) 191–213 Google Scholar; J. Blänsdorf, “The texts from the Fons Annae Perennae,” ibid. 215-44. A rapid survey of the finds, with colour figures, can be found in Friggeri, R. et al. (edd.), Terme di Diocleziano. La collezione epigrafica (Rome 2012) 617–39Google Scholar. It has, however, been recognised that Jesus Christ is invoked in 4 cases: G. Németh, “Il demone e Gesù Cristo,” ibid. 619.

6 Witteyer, M., Das Heiligtum für Isis und Mater Magna. Texte und Bilder. Schauraum zur Ausgrabung (Mainz 2004)Google Scholar; there is also an English edn. The fullest publication so far, which of course is not listed in Blänsdorf’s bibliography, is Witteyer, M., “Rituelle Niederlegungen im Heiligtum für Isis und Magna Mater in Mainz,” in ead. and Schäfer, A. (edd.), Rituelle Deponierungen in Heiligtümern der hellenistisch-römischen Welt (Mainzer Archäologische Schriften 10, 2013) 317–52Google Scholar. We await the definitive publication. The site was earlier known as the Lothar-Passage, and can be visited (see www.roemisches-mainz.de; www.roemerpassage.com).

7 The epigraphic material was edited by G. Alföldy and G. Rupprecht, apud Witteyer, Heiligtum (supra n.6) 15-21; 34 f. The earliest inscription is probably AE 2004. 1014, read as: [-----Primi]genius [-----] / [--- Imp. Ve]spasiani Aug. / [----]atoris a[r]carius / [Isidi et?Matri] deum ex im[p]erio / [eius ---] posuit. I consider the conjecture eius highly doubtful (as does L. Bricault in “RICIS Suppl. 2,” apud Bibliotheca Isiaca 2 [Bordeaux 2011] 290 f. ad 609/0501); Witteyer’s supplement procur]atoris arcarius is impossible — there was no such post. As the parallel cases in Clauss-Slaby (EDCS = Epigraphik-Datenbank Clauss-Slaby: at www.manfredclauss.de) show, we must allow room for the name of the dispensator (possibly including Caes. n. ser(vi)) between Primi]genius and the mention of Vespasian. Both men were imperial slaves, since it was usual not to award freedom to men in ‘clerical’ grades with responsibility for handling Imperial funds until their (late) 30s, or even later: Weaver, P. R. C., Familia Caesaris: a social study of the emperor’s freedmen and slaves (Cambridge 1972) 202 CrossRefGoogle Scholar. However, the existence of an arcarius dispensatoris in Mogontiacum under Vespasian, a post that existed only in large financial bureaux, proves that the settlement had become the administrative centre for what later became Germania Superior much earlier than the late Flavian period, as used to be thought. The date of AE 2004. 1015–16Google Scholar, which record dedications respectively to Mater Magna and Isis by Claudia Aug. lib. Icmas and Vitulus Caes. (servus), must remain doubtful in view of the initial formula pro salute Augustorum, for which there is no unproblematic solution; see the exhaustive discussion, long antedating this find, in Chantraine, H., Freigelassene und Sklaven im Dienst der römischen Kaiser. Studien zu ihrer Nomenklatur (Wiesbaden 1967) 225–63Google Scholar. Interpreting Augustorum as Augusti + Augustae, Alföldy and Rupprecht dated these texts to late in the reign of Domitian, almost certainly wrongly. However, neither Icmas’ nomen nor the presence of Neronian coins (and even a couple of Republican ones) in the sacrificial fire-pit can provide an assured date ante quem: imperial freedmen and -women may have been active up to 40 years after the death of the last possible manumitting emperor (in this case, Nero); and good coins might remain for decades in circulation.

8 This is the first temple of Isis to have been excavated in the Germaniae. The discovery of the Mater Magna temple shows that the cult of this deity was introduced there a full century earlier than had been thought (Blänsdorf 16 here). Joint temples of these deities are excessively rare, though odd individuals are known to have held priesthoods of both simultaneously: e.g., L. Valerius Fyrmus, sacerdos Isidis Ost(i)ens(is) et M(atris) d(eum) Tra(n)stib(erinae) (CIL XIV 429 = ILS 4406 = RICIS 503/1123); L. Pacilius Taurus at Brindisi ( CIL IX 6099 Google Scholar = ILS 4178 = RICIS 505/0301 [I disagree with Bricault's judgement that he was not a priest also of the sacror(um) Isidis]; and C. Iulius Severus at Falerii ( CIL XI 7484 Google Scholar = ILS 6587 = RICIS 511/0401). Note also the schola dedicated to both deities by the [cultores Ser-] apis (sic) at Portus: CIL XIV 123 Google Scholar = RICIS 503/1218.

9 Witteyer, M., “Curse tablets and voodoo dolls from Mainz. The archaeological evidence for magical practices in the sanctuary of Isis and Magna Mater,” MHNH 5 (2005) 105–24, fig. 3 (p. 110)Google Scholar. The temple of Isis was built directly over a tumulus of the Hallstatt period.

10 Witteyer, M., “Verborgene Wünsche. Befunde antiken Schadenzaubers aus Mogontiacum-Mainz,” in Brodersen, K. and Kropp, A. (edd.), Fluchtafeln. Neue Funde und neue Deutungen zum antiken Schadenzauber (Frankfurt 2004) at 42 Google Scholar.

11 It is not clear to me why these half-timbered buildings are said probably to have been used as banqueting spaces (Witteyer [supra n.9] 106).

12 After the loss of the right bank of the Rhine in the mid-3rd c., protective walls were built around Mogontiacum, connected to the existing walls of the legionary fortress. The town, which seems to have received even the status of a municipium only in the mid-4th c., was re-fortified after the abandonment of the fortress, probably under Valentinian I.

13 The Isiac inscriptions are AE 2004. 1017–23Google Scholar = RICIS Suppl. 1 nos. 609/0501-0510 (apud L. Bricault, “RICIS. Supplément I,” in id. [ed.], Bibliotheca Isiaca 1 [Bordeaux 2008] 77-130). Two of them mention decuriae of pausarii, whom Witteyer interpreted as bearers of Isiac images who marked the stops during processions (Heiligtum [supra n.6] 21). This was rightly doubted by Bricault (ibid. 120), who thinks they were marines serving on the patrol-boats on the Rhine and Main. J. Scheid has recently argued that the Egyptian cults were made official at Rome already in A.D. 70, if not 69: “Le statut du culte d'Isis à Rome sous le Haut-Empire,” in C. Bonnet et al. (edd.), Les religions orientales dans le monde grec et romain: cent ans après Cumont (1906-2006). Bilan historique et historiographique (Brussels 2009) 173-86. Numerous legionary brickstamps attest to the fact that at least some of the material used in the building came from official sources, but no precise inferences are possible until the final publication appears.

14 Plan: Blänsdorf here at 5 fig. 4; col. pls. on p. 6, figs. 5 f.

15 DTM 10 inner face 9 f.: -]d[i]liquescant, quat{m}/modi hoc liquescet [---; 11 inner face 7-11: sic illorum / membra liquescan(t), / quatmodum hoc plum/bum liquescet, ut eo/ru(m) exsitum sit; 12 inner face 2: qu[omo] di hoc liquescet … Note the use of the future tense, which, like the regular script of many tablets, proves that they were normally written out beforehand, presumably at home or in a workshop; some of the tablets are carefully cut into a rectangle and flattened with a hammer; others, especially the shorter examples, are simply cut out of any old lead sheet lying to hand. One of the finds (DTM 31) forms an almost perfectly round disc, though the writing makes no coherent sense.

16 See Witteyer (supra n.9) 122, figs. 11a-b. Some of the 140 tablets from the Romano-British shrine of Mercury at Uley were found in the débris of buildings in the temenos: Woodward, A., in ead. and Leach, P., The Uley shrines: excavation of a ritual complex on West Hill, Uley, Glos., 1977-9 (London 1993) 113 Google Scholar.

17 Cf. R. S. O. Tomlin, “carta picta perscripta: Anleitung zum Lesen von Fluchtafeln,” in Brodersen and Kropp (supra n.10) 11-29.

18 For example, there are 19 separate images of DTM 3 (photographs and drawings). There is also a full list of the various letter-forms for each of the 34 texts (Tables 1-4). The list of words is virtually complete, though I note that such an important, though short, word as per is missing and that by no means all the occurrences of, e.g., ut are listed.

19 In his careful publication of the 18 curse-tablets from the Roman-period sanctuary of Demeter at Corinth, R. S. Stroud provided excellent drawings but only black-and-white images, which provide an idea of the original but are not clear enough to allow the readings to be confirmed or questioned: Corinth XVIII.6 Google Scholar. The sanctuary of Demeter and Kore. The inscriptions (Princeton, NJ 2013) 81–157 Google Scholar.

20 Blänsdorf rightly observes here (42) that these texts were written by the principals themselves.

21 This fundamental difference was totally ignored in the early days of scientific study: for example, Wünsch, who was editing a coherent set of documents from Attica of the Late Classical and Early Hellenistic period, quite innocent of Graeco-Egyptian techniques (see n.1), prefaced them with an ill-digested collection of mostly much later material, as though he could see no difference — it was all timeless ‘magic’. A. Audollent in his fundamental collection DTAud (1904) likewise muddled up these different modes by choosing to arrange the material geographically instead of by period and style (his special interest was in the Roman-period texts from Carthage and Hadrumetum). Yet even a century later it is all too often obscured by modern treatments; a particularly glaring example is Martin, M., Sois maudit! Malédictions et envoûtements dans l’Antiquité (Paris 2010)Google Scholar.

22 This is the only case in which, if the reading is correct, members of the armed forces appear: one of the targets is a (legionary) tesserarius, the other a cavalryman (eques).

23 At Uley, indeed, 87 out of 140 found.

24 See http://www.inrap.fr/archeologie-preventive/Actualites/Communiques-de-presse/Les-dernierscommuniques/Communiques-nationaux/p-13288-Offrandes-et-maledictions-dans-un-edificecultuel-sur-le-site-des-Jacobins-au-Mans.htm. W. van Andringa says the others also carry squiggles.

25 There is, however, only one instance of this among the tablets recovered on the site: DTM 16.

26 I include 8 texts here, DTM 8-13, 16 and 28 (nos. 10-12 were found among the tiles laid over the site c.A.D. 130; 28 is only very partially legible). In all cases, there are uncertainties in the readings, which I pass over here.

27 Cassius Fortunatus e[t] bona illius et Lutatia Restituta: necetis e[os] (DTM 13). No. 12 l.4 likewise includes the target’s peculium (which need not imply that she was a slave).

28 DTM 10 l.5 seems to mention an arc(h)igallus, proving that these priests existed well before their supposed (re-)introduction during the reign of Antoninus Pius. Blänsdorf, relying upon a long-sinceexploded theory of J. Carcopino, claims here that the archigallate was (re-)introduced under Claudius (117, ad loc.); it is usually (but with what justification I do not know) ascribed to Antoninus Pius: cf. Vermaseren, M. J., Cybele and Attis. The myth and the cult (London 1977) 98 Google Scholar, cf. 107. To cite Pliny, NH 35.70 in support, as Blänsdorf does, brings nothing: that passage cites an otherwise unknown author ‘Deculo’, to the effect that Tiberius kept a painting of an ‘archigallus’ ascribed to the painter Parrhasios (5th c. B.C.) in his private apartments (cf. G. Lippold, s.v. Parrhasios 3, in RE 18.4 [1949] 1874-80 [1876 §17]); that proves only that the word was known in mid-1st c. Rome (which no one disputes, since it was a well-known institution in Asia Minor). Of the 9 men recorded in Latin inscriptions as archigalli (see Clauss-Slaby [supra n.7]), none must predate the 2nd c.; two certainly belong to the late 2nd/first half of the 3rd c.

29 These are DTM 1-7 and 15. I fail to understand why Blänsdorf places no. 15, one of the best-written and carefully-composed texts, so far down his batting order.

30 DTM 6 and 15; the rhetorical structure is set out by Blänsdorf on 105 and 139. No. 15 is also beautifully written in capitals with serifs.

31 See also, e.g., DTM 5 ll.3-6: per omnia te rogo …; 1 l.1 f.: Mater magna te rogo, per tua sacra et numen tuum. Per maiestate(m) tua(m) recurs in AE 1988. 727 Google Scholar = HEp 1991. 227 Google Scholar (Belo in Baetica), a curse addressed to Isis Muromem (sic). For such expressions in Latin entreaties, prayers and oaths, cf. Livy 30.12.13: precor quaesoque per maiestatem regiam … ; Sall., , BJug. 14.25 Google Scholar: per maiestatem populi Romani; Apul., Met. 6.15: quodque vos deieratis per numina deorum, deos per Stygis maiestatem solere. On maiestas as a function of divine distance: Apul., De mundo 27.

32 DTM 6 l.4 f.: ita ut galli … absciderunt concideruntue se, sic illi abscissa sit fides …;

33 Bellonarii and magali: DTM 2 l.10 and 12, with Blänsdorf’s comments here on p. 70, suggesting a corruption of megalephori, a term otherwise attested only in the cult of Isis, and then just once; DTM 6 l.3 Google Scholar invokes the galli and bellonarii, but not the magali.

34 The word cistas has also been read in DTM 1 l.43, but the context is unintelligible.

35 E.g., DTM 2 ll.16 f.Google Scholar: d[e]mando tibi rel[igione], ut me votis condamnes et ut laetus libens ea tibi referam, si de eo exitum malum feceris. The same expression seems to occur in one of my middle group 28 l.3 f.: uoto me condem[n]e[s… and in a curse from Alcácer do Sal in Lusitania ( AE 2001. 1135)Google Scholar: (h)ostia(m) quadripede(m) dono Attis (sic) uoueo si eas iure inuenero: dom(i)ne Attis, te rogo per tu(u)m nocturnum ut me quam primu(m) compote(m) facias (ll.4-8). A nocturnus/m is probably the night-long vigil commemorating the death of Attis.

36 DTM 2 ll.3 f.: Mater deum, tu persequeris per terras, per [maria per locos] ar(i)does et umidos; 4 l.2-4: in megaro eum (i.e., the target, who has just been named) rogo te, Mat[e]r magna, megaro tuo recipias; cf. Blänsdorf’s comment (85 f.); the megaron that receives Attis recurs in the curse from Salacia cited in the previous note, though for some reason AE has never recorded the re-reading MEGARE for the editors’ MEGALE by F. Marco Simón (“Magia y cultos orientales: acerca una defixio de Alcácer do Sal [Setúbal] con mentión de Atis,” MHNH 4 [2004] 79-94). I have elsewhere tried to show that these and other ideas are derived from wider knowledge of Metroac narratives and iconography: Gordon, R. L., “ Ut tu me vindices: Mater Magna and Attis in some new Latin defixiones ,” in Mastrocinque, A. and Scibona, C. Giuffrè (edd.), Demeter, Isis, Vesta and Cybele: studies in Greek and Roman religion in honour of Giulia Sfameni Gasparro (Stuttgart 2012) 195–212 Google Scholar.

37 Blänsdorf (69), rightly correcting many of the entries in Clauss-Slaby (supra n.7), where they seem all to be taken as proper names or agnomina — at any rate, they are all spelled with an initial upper-case B.

38 This text has already adverted to this theme at ll.12-15. One notes here the fine use of numen — i.e., the active, interventionist aspect of a deity, anticipated already at l.2: p[e]r …numen tuum. The others cases are DTM 2 l.8 f.: neque auro neque argento neque ille solui re]fici redimi possit; 5 ll.11-14: neque se possit redimere nulla pe{r}cunia nullaque re neq(ue) abs te (i.e., Attis) neque ab ullo deo nisi ut exitum malum. In DTM 2 l.16 the principal imagines the target coming the very next day to the temple and admitting he has committed a crime against the deity: cr[a]s [ueniat] et dicat se admisisse nef[a]s.

39 DTM 2 l.14; 3 l.9 and 4 l.7.

40 Blänsdorf 71 f. here.

41 Aginare: the sole reference offered by OLD is Petron. 61.9, and the translation offered is “do one’s best by hook or by crook”; Georges s.v.: “sich drehen und wenden”. The word recurs in DTM 4 l.6; Blänsdorf (80) provides further references in the late-antique (bilingual) dictionaries. An agina is defined by Paul., Festus p.9.12-14 L. as the hole into which the beam of a steelyard is fixed → aginator, a small-scale money-grubber or monger.

42 DTM 7.3-5: quo]mod[i] hoc grapphio auerso, quod minime, uti solet, sic [eum] auersum …, where the infringement of the norm is made explicit; a similar thought in 6 ll.4 f. and 10 f., where the idea of inversion is applied to the target’s mind and undertakings. For reversal of the direction of writing as a persuasive analogy in Latin curses, see C. A. Faraone and A. Kropp, “Inversion, adversion and perversion as strategies in Latin curse tablets,” in Gordon and Marco Simón (supra n.5) 381-98. This text has another persuasive analogy quomodi [.e]t ho[.] sucus defluit e[--- (l.6) but the sense cannot be made out — perhaps a reference to the melting of the lead, perhaps to a liquid offering.

43 This tablet is written as two blocks of text (ll. 6 f. and 7-14) enclosed within a frame-sequence containing 12 words (ll. 1-5). Blänsdorf understands ll. 6 f. as the completion of this latter sequence.

44 References will be found via the index of Latin words; see also Blänsdorf’s commentary (101) to DTM 6 l.6 Google Scholar.

45 See especially the recent publication by W. van Andringa and his team of part of the Porta Nocera cemetery at Pompeii: Mourir à Pompéi. Fouilles d’un quartier funéraire de la nécropole romaine de Porta Nocera, 2003-2007 (CollEFR 468, 2013), with review in this issue.

46 This phrase occurs at DTM 2 l.18 and 5 ll.7 and 14. DTM 6 l.9 demands the death within a year.

47 E.g., DTM 1 ll.31-34: nisi ut illas vorent canes, uermes adque alia portenta.

48 Here Blänsdorf (60 f.) thinks rather of the famous cases of those who were eaten alive by ‘worms’, such as Antiochus IV Epiphanes and Sulla.

49 I am thinking of the volume edited by Asirvatham, S. R. et al., Between magic and religion: interdisciplinary studies in ancient Mediterranean religion and society (Lanham, MD 2001)Google Scholar.

50 E.g., Versnel, H. S., “Beyond cursing: the appeal to justice in judicial prayers,” in Faraone, C. A. and Obbink, D. (edd.), Magika hiera: ancient Greek magic and religion (Oxford 1991) 60–106 Google Scholar; id., “Prayers for justice in East and West: recent finds and publications,” in Gordon and Marco Simón (supra n.5) 275-355.

51 See in more detail my paper “Gods, guilt and suffering: psychological aspects of cursing in the northwestern provinces of the Roman Empire,” Acta Classica Univ. Scient. Debrecensis 49 (2013) 255–81Google Scholar. It is the great merit of Eidinow, E., Oracles, curses and risk among the Ancient Greeks (Oxford 2007)CrossRefGoogle Scholar, to have shown the value of thinking about defixiones as strategies of handling risk.

52 Cf. Stewart, P. J. and Strathern, A., Witchcraft, sorcery, rumors and gossip (Cambridge 2004) 29–58 Google Scholar.

53 One of Versnel’s difficulties is the sheer number of what, from his point of view, are ‘intermediate cases’ which do not correspond to his ideal types. In my view, the very idea of an ideal-type ‘magical’ or ‘malign’ defixio is untenable. Not only is it just a strategy to obtain divine justice or redress in a given emergency; the sheer range of means employed and ideas expressed make it hard to set up a convincing ideal-type that is not based on precisely the assumptions that need to be questioned. The excellent catalogue of Latin defixiones assembled by C. Sánchez Natalías, a member of our project on magic based in Zaragoza (about to be published by Archaeopress in Oxford), amply demonstrates the sheer variety of ideas and strategies deployed: cf. her El contenido de las defixiones en el Occidente del Imperio Romano (diss. Zaragoza–Verona 2013)Google Scholar.

54 My remarks here solely concern ‘do-it-yourself’/‘common or garden’ defixiones, not those in the learned Graeco-Egyptian tradition which were written by ritual specialists.

55 To me it seems special pleading to talk about ‘first strikes’ (especially in judicial cases) as a device to maintain the idea that defixiones are evidence of malign magic. In my view, those who chose this route had every reason to believe that the other side was busy suborning witnesses and twisting ‘the truth’. It is not an accident that the great majority of this type of defixio is found in late Classical and Hellenistic Attica, with their chaotic ochlocratic system of justice.