¹ Les mystères de Mithra (3rd edn, 1913), 27.

² ibid., 17

³ For a survey, see Beck, R., ‘Mithraism since Franz Cumont’, ANRW 11.17.4 (1984), 2002–114Google Scholar, at 2071–5.

⁴ Will, E., Le relief cultuel gréco-romain (1955), 144–69Google Scholar; idem, ‘Origine et nature du Mithriacisme’, EM, 527–36, at 527–8; Colpe, C., ‘Mithra-Verehrung, Mithras-Kult und die Existenz iranischer Mysterien’, MS Vol. 2, 378–405Google Scholar, at 390–9; E. Schwertheim, ‘Monumente des Mithraskultes in Kommagene’, Kommagene, 63–8; idem, Mithras: seine Denkmäler und sein Kult, Antike Welt Sondernummer (1979), 13–24; Gordon, R. L., ‘The date and significance of CIMRM 593’, JMS 2 (1978, reprinted in Gordon 1996), 148–74Google Scholar; idem, ‘Who worshipped Mithras?’ JRA 7 (1994), 459–74, at 469–71 (more cautiously); Boyce, M. and Grenet, F., A History of Zoroastrianism Vol. 3 (1991), 468–90Google Scholar. On the limitations of the Anatolian evidence, see Beck, op. cit. (n. 3), 2018–19; Gordon, op. cit. (above, 1994), 461–2.

⁵ Will, op. cit. (n. 4, 1955), 164 f.; Turcan, R., Mithra et le mithriacisme (2nd edn, 1993), 25–6Google Scholar.

⁶ Vermaseren, M. J., ‘Mithras in der Römerzeit’, in Vermaseren, M. J. (ed.), Die orientalischen Religionen im Römerreich (1981), 96–120CrossRefGoogle Scholar, at 96–103; Merkelbach, R., Mithras (1984), 77, 160–1, 146–9Google Scholar; Clauss, M., Mithras: Kult und Mysterien (1990), 31–2Google Scholar; idem, Cultores Mithrae: Die Anhängerschaft des Mithras Kultes (1992), 253–5; more tentatively, W. Liebeschuetz, ‘The expansion of Mithraism among the religious cults of the second century’, SM, 195–216, at 199–200. There is a telling critique in Gordon, op. cit. (n. 4, 1994), 467–8. From the publication dates, it will be apparent that this is the later of the two trends; indeed, when I made my survey of post-Cumontian scholarship (op. cit. (n. 3), 2074) it was still something to be desired. There is a third trend, that typified by G. Widengren (‘The Mithraic mysteries in the Greco Roman world, with special regard to their Iranian background’, in La Persia e il mondo greco-romano, Accademia Nazionale dei Lincei, Anno 363, Quaderno 76 (1966), 433–55; idem, ‘Reflections on the origins of the Mithraic Mysteries’, in Perennitas: Studi in onore di Angelo Brelich (1980), 645–68; cf. Beck, op. cit. (n. 3), 2065–6, and 2013–4 with n. 14), which sees in the Mysteries essentially a continuity from Iran. I pass it by, not because it is negligible, but because by definition it postulates no new genesis of the Mysteries as part of the process of east–west transmission of Mithra-worship.

⁷ Geschichte der griechischen Religion (3rd edn, 1974 Vol. 2, 675 fGoogle Scholar.

⁸ An approach to history much the same as that later described by Veyne, P. in Comment on écrit l'histoire (1971)Google Scholar: ‘Rien qu'un récit véridique’ (ch. I, title).

⁹ I have in mind particularly the work of Manfred Clauss. This is not to belittle the great contribution made by Clauss in Cultores Mithrae (above, n. 6) on the basis of the cult's epigraphy (see my review, Phoenix 48 (1994), 173–6CrossRefGoogle Scholar, and Gordon's review article, op. cit. (n. 4, 1994)), but to sound a note of caution against carrying a reasonable scepticism concerning iconographic interpretations too far. See the retrospective and programmatic statements in the introduction to Mithras: Kult und Mysterien (above, n. 6), 7–9: of Cumont, ‘in der Annahme, die Religion bestehe essentiell in ihrer Theologie, vernachlässigte er den Kult’ (7). In a sense, Clauss renews the counter trend to Cumont's approach which, in Cumont's own day, was typified by Toutain, J., Les cultes paiens dans l'empire romain, I: Les provinces latines (3 vols, 1907–1920)Google Scholar; MacMullen, R., Paganism in the Roman Empire (1981), 116Google Scholar, with n. 11, has an interesting perspective on this earlier debate, favouring Toutain.

¹⁰ This final, illogical step is taken by Swerdlow, N. M., ‘On the cosmical mysteries of Mithras’, CP 86 (1991), 48–63Google Scholar, a review article of Ulansey, D., The Origins of the Mithraic Mysteries (1989)Google Scholar. Swerdlow's dismissal of the rich body of astrological evidence, in both the texts and the monuments, because of the excesses of its recent interpreters, leads him to the further dismissal, as contemptuous as it is ill-considered, of the Mysteries and their initiates alike: ‘… those who ask “What was Mithraism, anyway?” just may conclude that it was nothing much, and perhaps not a serious religion after all’ (62).

¹¹ To cite a work of each, Vermaseren, M. J., Mithra, ce dieu mystérieux (trans. Léman, M. and Gilbert, L., 1960)Google Scholar; Bianchi, U., ‘The religio-historical question of the mysteries of Mithra’, in Bianchi, U. (ed.), Mysteria Mithrae (1979), 3–60CrossRefGoogle Scholar; Turcan, op. cit. (n. 5); Merkelbach, op. cit. (n. 6).

¹² loc. cit. (n. 7). It is worth pointing out that, although unusual in classical scholarship, this sort of hypothetical reconstruction is both commonplace and fundamental to the study of primitive Christianity (or Christianities). New Testament Form Criticism, for instance, works by reconstructing from the texts (viz. the Gospels) the early Christian or proto-Christian communities whose needs those texts were intended to serve. This method even involves second-order hypotheses: e.g., a certain type of community is hypothesized for the common source text of Matthew and Luke, called ‘Q’; but ‘Q’ itself is hypothetical in that, although generally accepted as an actual text by New Testament scholarship, it is neither extant nor directly attested. The title of a recent work by one of my Toronto colleagues is illustrative of this method: Vaage, L. E., Galilean Upstarts: Jesus' First Followers According to Q (1994)Google Scholar.

¹³ S. Wikander, ‘Études sur les mystéres de Mithras’, Vetenskapssocieteten i Lund, Årsbok (1951), 5–46; P. Beskow, ‘The routes of early Mithraism’, EM, 7–18.

¹⁴ Blakely, J. A. et al. , Caesarea Maritima: The Pottery and Dating of Vault I, Joint Excavation Report 4 (1987), 62, 103Google Scholar; cf. Painter, R. J., Mithraism and the Religious Context at Caesarea Maritima, dissertation, Southern Baptist Theological Seminary (1994), 99–163Google Scholar.

¹⁵ Roll, I., ‘The mysteries of Mithras in the Roman orient: the problem of origin’, JMS 2 (1977), 18–52Google Scholar; Downey, S. B., ‘Syrian images of Mithras Tauroctonos’, EM, 135–49Google Scholar; Beck, op. cit. (n. 3), 2013–17.

¹⁶ From the perspective of dating, the dossier should be regarded as a composite of the certain and the highly probable. Most of its elements are conveniently set out in Merkelbach, op. cit. (n. 6), 147–9; see also Clauss, op. cit. (n. 6, 1990), 31–2; idem, op. cit. (n. 6, 1992), 251–2; Gordon, op. cit. (n. 4, 1994), 460–2, 467–8, 470. I have also consulted material prepared by Richard Gordon for a book on Mithraism which we are writing together. I am greatly dependent on, and grateful for, his expertise in this early phase of the Mysteries, especially in Germany where the picture is extremely complicated.

¹⁷ Gordon, op. cit. (n. 4, 1994), 461, with references in n. 8.

¹⁸ I. Huld-Zetsche, Mithras in Nida-Heddernheim (1986), 33–6. See also Vermaseren (below, n. 22), no. 1117; Schwertheim, E., Die Denkmäler orientalischer Gottheiten im römischen Deutschland (1974), no. 61; Clauss, op. cit. (n. 6, 1992), 115–17.CrossRefGoogle Scholar

¹⁹ Horn, H. G., ‘Das Mainzer Mithrasgefäβ’, Mainzer Archäologische Zeitschrift I (1994), 21–66Google Scholar, at 31–2 (see also Merkelbach, R., ‘Das Mainzer Mithrasgefäß’, ZPE 108 (1995), 1–6Google Scholar). Strictly, we have to do with the dating of a ritual cup of a certain pottery type (Wetterau ware); unfortunately, the mithraeum where it was discovered could not be systematically excavated. This remarkable object is decorated with seven figures, representing cult members engaged in two scenes of ritual performance. Though some of the figures are grade holders (the Pater and the Heliodromus are readily identifiable), finding a corresponding grade for every figure (let alone one-for-one correspondences with each of the seven grades in the hierarchy) is problematic. I am currently working on an explication of the cup's two scenes. What is undeniable is that the cup documents a developed, indeed sophisticated, ritual and ideology at a very early date in the cult's life.

²⁰ Garbsch, J., ‘Pons Aeni’, Bayerische Vorgeschichtsblätter 50 (1985), 355–462Google Scholar, at 428–35. See also Clauss, op. cit. (n. 6, 1992), 135.

²¹ Above, n. 14.

²² Where applicable, monuments are cited by their number in M. J. Vermaseren, Corpus Inscriptionum et Monumentorum Religionis Mithriacae (2 vols, 1956–60), prefixed with ‘V’. On the dossier of the earliest epigraphy, see Gordon, op. cit. (n. 4, 1994), 467. Gordon rightly draws attention to a certain fluidity of language in Mithras' cult-titles both in these early inscriptions and in the relatively few inscriptions from Anatolia: op. cit. (n. 4, 1978), 159 f.

²³ Respectively, by the cavalryman Tacitus and the centurion C. Lollius Crispus: Schwertheim, op. cit. (n. 18, 1974), nos 59i and o; Huld-Zetsche, op. cit. (n. 18), 55–6, nos 8 and 9; Merkelbach, op. cit. (n. 6), 149, nos 6 and 7; Clauss, op. cit. (n. 6, 1992), 116; Gordon, op. cit. (n. 4, 1994), 460.

²⁴ By the centurion C. Sacidius Barbarus: Merkelbach, op. cit. (n. 6), 148–9, no. 5; D. Schön, Orientalische Kulte im römischen Österreich (1988), no. 50; Clauss, op. cit. (n. 6, 1992), 157. See also C. M. Daniels, ‘The role of the Roman army in the spread and practice of Mithraism’, MS Vol. 2, 249–74, at 250–1.

²⁵ By Melichrisus, slave of P. Caragonius Philopalaestrus, conductor of the publicum portorium Ripae Thraciae: Merkelbach, op. cit. (n. 6), 148, no. 4; Clauss, op. cit. (n. 6, 1992), 224; Gordon, op. cit. (n. 4, 1978), 153–4; idem, op. cit. (n. 4, 1994), 267–8. Melichrisus' name may be esoteric, alluding to the purification of Mithraic Lions with honey (Porphyry, De antro 15). A peculiar detail of this monument is that the deity Cautopates carries an upside-down cockerel, balancing the cockerel carried upright by Cautopates' twin, Cautes, on the opposite side.

²⁶ By the veteran and pater sacrorum T. Tettius Plotus: Clauss, op. cit. (n. 6, 1992), 225.

²⁷ Respectively, by the imperial freedman T. Flavius Hyginus Ephebianus and the slave Alcimus, the vilicus of T. Claudius Livianus (in all likelihood, the praefectus praetorio under Trajan): Gordon, op. cit. (n. 4, 1978), 151–3, 155–6; Merkelbach, op. cit. (n. 6), 147–8, nos 3 and 2; Clauss, op. cit. (n. 6, 1992), 19–20 (see nn. 32, 35, 54). Hyginus makes his dedication ‘through his father (dia … patros idiou), Lollius Rufus’, and most scholars take ‘father’ not in the natural but in the esoteric hierarchic sense. The tauroctony dedicated by Alcimus has some significant idiosyncratic features: the torchbearers are grouped together on the same side of the composition, i.e., behind the bull's tail to the viewer's left, rather than one on each side; the ears of wheat appear not as growths on the bull's tail but as patterns of blood flowing from the wound struck by the god.

²⁸ I have included the one Anatolian monument which (a) belongs to this early period and (b) cannot be dismissed, because of iconographic or other dissimilarities, as definitely not a monument of the Mysteries. To exclude it, on the grounds that it must none the less belong to some collateral branch of Mithra-worship (since nothing about it absolutely compels us to attribute it to the Mysteries), would have begged the question. It is a dedication of one Midon, son of Solon, to Helios Mithras, and shows the bust of the god in a Phrygian cap. See, most recently Gordon, op. cit. (n. 4, 1994), 470. We shall return below to the dossier of Anatolian monuments.

²⁹ The well-known allusion by Statius to Mithras subduing the bull in the ‘Persian cave’: ‘seu Persei sub rupibus antri/ indignata sequi torquentem cornua Mithram’. See Gordon, op. cit. (n. 4, 1978), 161–4; Merkelbach, op. cit. (n. 6), 147, no. 1. If Statius is drawing, directly or indirectly, on the cult icon, he has either taken liberties with the standard iconography or else is replicating a non-standard (prestandard?) exemplar: Mithras, on the monuments, typically holds the bull by the muzzle, not the horns.

³⁰ The social appeal of Mithraism to conformists is agreed by all; see esp. R. L. Gordon, ‘Mithraism and Roman society’, Religion 2 (1972), 92–121 (reprinted in Gordon 1996); Merkelbach, op. cit. (n. 6), 153–88; Clauss, op. cit. (n. 6, 1990), 42–50; idem, op. cit. (n. 6, 1992), 262–75; Turcan, op. cit. (n. 5), 37–41; Liebeschuetz, op. cit. (n. 6). All stress the importance of familiae and of bureaucratic and military structures in the propagation of the Mysteries. The permission, indeed the encouragement, of superiors is likewise assumed. Liebeschuetz (203–6) rightly observes that the civilian groups frequently belong to a second tier of dependency, as the freedmen/slaves of freedmen/slaves.

³¹ As can readily be appreciated from the status of the dedicators: above, nn. 23–7.

³² Although, of course, considerable disagreement persists about the precise nature and extent of the Iranian component and the manner of its metamorphosis into the stuff of the Mysteries. The most thoroughgoing and effective critique of Cumont's interpretation of Mithraism as transmuted Mazdaism is Gordon's (‘Franz Cumont and the doctrines of Mithraism’, MS Vol. 1, 215–48), the most successful redefinition of the Mysteries' continuing Iranian ethos—at least in my opinion—Turcan's: ‘Le sacrifice mithriaque: Innovations de sens et de modalités’, in Le sacrifice dans l'antiquité classique, Entretiens sur l'Antiquité classique, Fondation Hardt (1981), 341–80; ‘Salut mithriaque et sotériologie néoplatonicienne’, in Bianchi, U. and Vermaseren, M. J. (eds), La soteriologia dei culti orientali nell' impero romano (1982), 173–91Google Scholar; ‘Le dieu et le divin dans les mystères de Mithra’, in van de Broek, R. (ed.), Knowledge of God in the Graeco-Roman World (1988), 243–61Google Scholar. On the maximalist view of Widengren and others, see above, n. 6.

³³ Scepticism is undoubtedly warranted by the implausibility of much of the astronomical/astrological interpretation of the cult-icon: Swerdlow, op. cit. (n. 10); Turcan, op. cit. (n. 5), 105–8; R. Beck, ‘In the place of the Lion: Mithras in the tauroctony’, SM, 29–50, at 32–40; cf. idem, op. cit. (n. 3), 2081–3. But the body of astrological data, occurring in both the monuments and the texts and concerning not only the icon but also the mithraeum and the grade hierarchy (see below), is not to be denied merely because of the perceived inadequacies of its interpreters or because of its difficult, somewhat rebarbative nature: Beck, R., Planetary Gods and Planetary Orders in the Mysteries of Mithras (1988), ix–xiiCrossRefGoogle Scholar. Patient, more sophisticated, and methodologically sounder evaluation of the evidence is certainly to be desiderated, but so is a recognition of its pervasiveness and complexity. Practitioners of this line of inquiry and their critics alike have a long but, one hopes, rewarding road to travel. There is an interesting story to be told—though not here—as to why Cumont, himself no mean scholar of ancient astrology, so persistently undervalued and marginalized the astrology of the Mysteries.

³⁴ One might, for example, envisage a figure such as Paul of Tarsus, who defined a certain type of Christianity even as he transmitted it.

³⁵ op. cit. (n. 10). Ulansey proposes a group of Tarsian Stoics who, at some time after the middle of the second century B.C., transmuted the astronomer Hipparchus' highly technical hypothesis about the precession of the equinoxes into the foundation doctrine of the new cult. As I have demonstrated (Beck, op. cit. (n. 33), 37–9), other than a few professional astronomers, almost no one took cognizance of Hipparchus' discovery (Origen and Proclus are the two exceptions, with a single reference apiece), and no one at all was interested in the historical reconstruction of the equinoxes of past epochs that Ulansey's account postulates. To imagine that people in Antiquity might have turned such matters into a religion is an egregious anachronism.

³⁶ The story of the dynasty and its fortunes is well and fully told by Sullivan, R. D., ‘The dynasty of Commagene’, ANRW II.8 (1977), 732–98Google Scholar. We are concerned with its later phases (Sullivan, 785–98), particularly with the times of the last reigning king, C. Iulius Antiochus IV Epiphanes (A.D. 38–72; PIR ² 4.138–40, no. J 149), and his son of the same name (PIR ² 4.140–1, no. J 150; RE 10.1.159–63, Iulius no. 66). The acme of Romanization was reached in the next generation by C. Iulius Antiochus Epiphanes Philopappus (he of the monument in Athens), suffect consul in 109 and Arval Brother (PIR ² 4.141, no. J 151). The kingly title was retained in the latter two generations, as was a proud dynastic memory (see below on Julia Balbilla). The dynasty had, of course, long been Hellenized, tracing its pedigree to Alexander and the Seleucids and interweaving Greek with Iranian in the pantheon of its cult (see next note). Highly germane, from our perspective, is the family's connection with the astrologer and high Roman functionary, Ti. Claudius Balbillus. The specifics of the link and the prosopography of Balbillus are complicated and controversial (less so now than formerly); they will be discussed briefly below. However interpreted, the relationship adds an unusual cultural dimension to the dynasty's Romanization. Was there a ‘trickle down’ of astrological doctrine to those in the dynasty's household to whom our scenario traces the origins of the Mysteries?

³⁷ See esp. Merkelbach, op. cit. (n. 6), 50–72. The scholarly literature on the Commagenian royal cult is considerable; I cite the two most recent comprehensive studies, both excellent: M. Boyce, op. cit. (n. 4), 309–51; Waldmann, H., Der kommagenische Mazdaismus (1991)Google Scholar. This is also the point at which to recognize the contribution to the study of Commagenian religion of another distinguished Belgian scholar, Jacques Duchesne-Guillemin: ‘Iran and Greece in Commagene’, EM, 187–204. On Mithras in particular in the context of the royal cult, see Dörner, F. K., ‘Mithras in Kommagene’, EM, 123–33Google Scholar; Schwertheim, op. cit. (n. 4, 1979), 13–18; Beck, op. cit. (n. 3), 2017–18. Important here because he has more to say than most about the dynasty and the cult subsequent to Antiochus I is Wagner, J., ‘Dynastie und Herrscherkult in Kommagene’, Istanbuler Mitteilungen 33 (1983), 177–224Google Scholar (see esp. 208–24). The articles and illustrations in the Antike Welt Sondernummer 1975 devoted to the kingdom (= F. K. Dörner (ed.), Kommagene) furnish an excellent overview.

³⁸ In its final phase, under Antiochus IV, the kingdom included portions of maritime Cilicia (Dio 59.8.2, Jos., , AJ 19.276Google Scholar). In A.D. 52, Antiochus campaigned against some wild tribes there (agrestium Cilicum nationes) which had been harrowing the coastal cities (Tac., , Ann. 12.55Google Scholar). Pacification was achieved by isolating and killing the chieftains and ‘settling the rest leniently’ (ceteros clementia composuit). It is worth considering whether the germination of the Mysteries might not have taken place when Commagenian and Cilician Mithra-worship coalesced at the exposure of Commagenian administrators and military to the rites of the Cilician tribes. On this scenario, the Mysteries would indeed have been transmitted from the teletai of Cilician outlaws, as Plutarch's testimony declares (above, n. 5), but at a different time and by a different route than scholars have supposed. These rites, one may rather postulate, were not carried abroad and perpetuated by Pompey's resettled pirates—Plutarch does not in fact say that they were—but instead lingered in their homeland of Cilicia until taken up by the Commagenians more than a century later.

³⁹ Josephus, , BJ 7.219–43Google Scholar, gives a full account of this episode and its aftermath.

⁴⁰ In the Civil Wars, on the side of Otho against the Vitellians in the battle twelve miles from Cremona (Tac., , Hist. 2.25.2Google Scholar); in the Judean War, at the siege of Jerusalem (Tac., , Hist. 5.1.2,Google ScholarJos., , BJ 5.460–5Google Scholar). On both occasions the Commagenians were led by Antiochus, the king's son (above, n. 36); on the latter occasion, he volunteered his crack detachment of ‘Macedonians’ in a gallant—or foolhardy—assault on the walls. There had been earlier co-operation, and hence presumably contacts, between Commagenian and Roman forces in Corbulo's Armenian campaigns (Tac., , Ann. 13.7.1Google Scholar, 37.2).

⁴¹ Daniels, op. cit. (n. 24), 250–2; Turcan, op. cit. (n. 5), 32. On Caesarea as a likely military contact point, see Painter, op. cit. (n. 14), 45–9, 115–19. Braund, D., ‘New “Latin” inscriptions in central Asia: Legio XV Apollinaris and Mithras?’ ZPE 89 (1991), 188–90Google Scholar, is properly sceptical of the interpretation of a Latin-alphabet inscription found in Northern Bactria which has a detachment of XV Apollinaris worshipping Mithras in a cave there some time during this period(!).

⁴² Jos., , BJ 7.243Google Scholar (‘… and there they remained (katemenon), treated with every respect’). Antiochus IV was no stranger to the city; it was presumably there that, together with Herod Agrippa of Judaea, he ‘associated with’ (syneinai) Caligula, a relationship which the Romans observed with dismay, considering the pair of client princes ‘mentors in tyranny’ (tyrannodidaskalous): Dio 59.24.1. The acquaintance of his eventual kinsman by marriage, Balbillus (above, n. 36), was most likely made in Rome.

⁴³ Painter, loc. cit. (n. 14).

⁴⁴ On the dossier of Anatolian material, Cumont, F., ‘Mithra en Asie Mineure’, Anatolian Studies in Honour of W. H. Buckler (1939), 67–76Google Scholar; Will, op. cit. (n. 4, 1955), 154–6; Beck, op. cit. (n. 3), 2018–19; Gordon, op. cit. (n. 4, 1978), 159–60; idem, op. cit. (n. 4, 1994), 461–2, 469–70. See above, n. 28, on V 23; also n. 22, on the variety in forms of dedication, to which Gordon rightly draws attention as evidence of a certain fluidity in the types of Mithra-worship and the early Mysteries there.

⁴⁵ It is in this category that I would place the Kerch terracottas, on which see Beck, op. cit. (n. 3), 2019.

⁴⁶ See above, n. 30 and the studies cited there; see also Beck, R., ‘The Mysteries of Mithras’, in Kloppenborg, J. S. and Wilson, S. G. (eds), Voluntary Associations in the Graeco-Roman World (1996), 176–85Google Scholar, at 179. For a recent sociological perspective on the growth of new religious movements through social networks (with applications to the ancient world), see Stark, R., The Rise of Christianity (1996), 13–21Google Scholar.

⁴⁷ Attempts have indeed been made to project Mithraism back into the royal cult, notably on the basis of the artificial cavern at Arsameia on the Nymphaeus as a precursor of the mithraeum: Schwertheim, op. cit. (n. 4, 1975); idem, op. cit. (n. 4, 1979), 13–18; Dörner, op. cit. (n. 37, 1978), 132–3; contra, H. Dörrie, Der Königskult des Antiochus von Kommagene im Lichte neuer Inschriften-Funde (1964), 192–4. Waldmann, op. cit. (n. 37), 182–4, gives an appropriately sceptical overview of the question ‘Mithrasmysterien in Kommagene?’ Actually, Waldmann's question can now be answered in the affirmative, though in a different sense than he intended. Very recently (late summer 1997), a mithraeum was discovered in a natural cave at Doliche, with a representation of the bull-killing Mithras cut into the rock. So far, however, there is nothing to suggest a particularly early date or that this is other than a standard Roman mithraeum. Nevertheless, the discovery is of great importance, and I return to it in a postscript.

⁴⁸ By ‘invention’ I mean, rather in the rhetorical sense, the discovery of the bull-killing as a divine fact of supreme relevance and its subsequent elaboration in myth and doctrine. By ‘salvation’ I mean only the effect of the act for good, however defined, on the world and, as mediated through the cult, on the initiates. The specifics of that good I here leave undefined. If Mithras as bull-killer was indeed the ‘invention’ of the founding cultists in the middle of the first century A.D., then it is not surprising that the search for his Iranian original has proven so unsatisfactory (see Beck, op. cit. (n. 3), 2068–9). (The exploration of the underlying Iranian concepts of sacrifical killing, whether by god or mortal, is of course another matter.) The reason that we hear no hint of a bull-killing Mithras prior to the late first century A.D. is, quite simply, that he did not exist until shortly before that time. If I were to suggest an antecedent for the motif of the bull-killing, I would locate it close to Commagene—in Tarsus and the image of the bull-vanquishing lion, prominent in the earlier coinage of that city. In fact, this antecedent has already been proposed, inter alia, by A. D. Bivar, ‘Towards an integrated picture of ancient Mithraism’, SM, 61–73, at 64–5. I would construe it, however, not as a forerunner of the bull-killing Mithras but rather as a trigger to his invention, a pre-existent motif, quite unrelated to him, which might have given both impetus and local legitimacy (the latter through the appearance of traditional depth) to a new religious creation. Again, I would pull the moment forward in time to the age of Antiochus IV and the expansion of Commagene into Cilicia (above, n. 38—though Tarsus, of course, remained outside his realm).

⁴⁹ The monuments of the royal cult make this equation both in text and in iconography. Mithras is called Apollo-Mithras-Helios-Hermes in the foundation text of the primary hierothesion on Nemrud Daǧ (V 32 = OGIS 383 = Waldmann, H., Die kommagenischen Kultreformen (1973), 62–9CrossRefGoogle Scholar, line 55), and elsewhere likewise, though with variation in the order of names. On the reliefs which show him in dexiosis with the king, one extant on the west terrace of Nemrud Daǧ (V 30) and two at Arsameia on the Nymphaeus, his Persian tiara is surrounded with the rayed solar halo: Waldmann (above), pls 22.3, 30.2–3; Dörner, Kommagene, 41 Abb. 42, 56 Abb. 82; idem, op. cit. (n. 37), pls 3–6; Schwertheim, op. cit. (n. 4, 1979), 17 Abb. 16, 20 Abb. 19; Merkelbach, op. cit. (n. 6), 266–7, Abb. 4–5. That the equation of Mithras with the Sun was formulated in the context of the royal cult and not inherited ready-made is proved by a different dexiosis relief, discovered in 1974 at Sofraz Köy, which belongs to an earlier phase of the cult and predates the identification of the two divinities. Here the god with solar rays and halo is named simply Apollo Epēkoos, his iconography is entirely Hellenic, and he is neither called nor does he carry any of the attributes of Persian Mithra: J. Wagner, ‘Neue Funde zum Götter- und Königskult unter Antiochos I. von Kommagene’, Kommagene, 51–9, at 54–9 with Abb. 77; Wagner, J. and Petzl, G., ‘Eine neue Temenos-Stele des Königs Antiochos I. von Kommagene’, ZPE 20 (1976), 201–23Google Scholar; Schwertheim, op. cit. (n. 4, 1979), 20, Abb. 20; Wagner, op. cit. (n. 37), 192–4, 198–208, pl. 49.4; Beck, op. cit. (n. 3), 2018. (Very similar is the Samosata relief of Helios: Waldmann, op. cit. (above), pl. 5.) Solar Mithras thus represents a continuity from the royal cult of Commagene to the Mysteries of Mithras, but it is not a long-standing one, being an ‘invention’ of the royal cult itself in the same sense that Mithras as bull-killer was the ‘invention’ of the Commagenian founders of the Mysteries a century or so later (see preceding note). In our account, then, the ‘invention’ of Helios-Mithras in the Commagenian royal cult is sufficient causal explanation of the solarity of Mithras in the Mysteries: Mithras is the Sun in the Mysteries because the Commagenian founders of the Mysteries received him in that identity. Further theories about the remoter Iranian origins of Mithras' solarity (a much vexed question: see Beck, op. cit. (n. 3), 2068; Boyce, op. cit. (n. 4), 479–82) thus concern not so much the Mysteries themselves as the antecedent royal cult: on what precedent, if any, did Antiochus I build his identification of Mithras with the Sun? The chronology of the monuments of the royal cult and the development of the various divine equations recorded there are complicated—and still very open—questions. Necessarily, I have simplified, but only to what commands general agreement. For discussions of these issues see, in addition to the works of Wagner cited above, Dörner, op. cit. (n. 37); Duchesne-Guillemin, op. cit. (n. 37); Boyce, op. cit. (n. 4), 317–49; Waldmann, op. cit. (n. 37), esp. 55–9 (note that, for relative chronology, this work replaces Waldmann's 1973 work cited above).

⁵⁰ On the phrase, see below, n. 53.

⁵¹ Beck, op. cit. (n. 33, 1994); see also above, n. 33.

⁵² V 31. The stars are arranged in the pattern of Leo the constellation, three larger stars above the lion's back are identified by inscription as the planets Mars, Mercury, and Jupiter, and there is a crescent moon on the lion's chest. There is general agreement, following O. Neugebauer and Hoesen, H. B. Van, Greek Horoscopes (1959), 14–16Google Scholar, no. 61, that the monument functions as a horoscope whose designated date is 7 July 62 B.C.; also, following Dörrie, op. cit. (n. 47), 201–7, that the accompanying dexiosis reliefs are to be interpreted in the light of this horoscope as alluding to the successive conjunctions (within a few days of each other) of the three planets and the moon with the principal star of Leo, Regulus, symbolized on the lion monument by the large star cradled in the lunar crescent on the lion's chest. The dexiosis reliefs thus also carry an astrological message, that the planetary gods on a particular occasion came to greet and be greeted by the king's celestial surrogate, ‘the royal star at the heart of the lion’ (phrase, Pliny's, NH 18.235Google Scholar, 271). In their ensemble, they both establish and validate the theological equations of the royal cult by reference to what had actually occurred in the heavens. (While essentially correct, the current astrological interpretation of the monuments requires modification, in part because Dörrie's conjunction data were erroneous and present a misleading and oversimplified picture of the underlying celestial events and configurations, and in part because Neugebauer and Van Hoesen, while, of course, accurate in their data, were severely limited in their treatment of the horoscope within its unusual religious context. I am presently at work on a revised interpretation, as also on the astrology of the cult's later monuments, notably the Karakuṣ site, a hitherto neglected line of inquiry, though obviously worth exploring if one is arguing for a continuity in astrological thinking from the royal cult to the Mysteries of Mithras.)

⁵³ Porphyry, De antro nympharum 6. On the way in which this ideal is exemplified in actual mithraea, principally the Mithraeum of the Seven Spheres in Ostia, see Beck, R., ‘Cosmic models: Some uses of Hellenistic science in Roman religion’, in Barnes, T. D. (ed.), The Sciences in Greco-Roman Society (1995) = Apeiron 27 (1994)Google Scholar, no. 4, 99–117. In principle and in nomenclature, though not in practice, every mithraeum is a ‘cave’ because the cave is a symbol of the universe (Porph., ibid.). The Mysteries, in a sense, went underground, in sharp contrast with their royal antecedent, which was a cult of high, open places. It is conceivable that the artificial cavern tunnelled out at the Arsameia site might have prompted the Mysteries to take that direction, but it is certainly not the prototype of a mithraeum (see above, n. 47). On the mithraeum recently discovered in a natural cave at Doliche, see above (n. 47), and below (Postscript).

⁵⁴ Beck, op. cit. (n. 33, 1988), esp. 1–11. The phrase (klimax heptapylos) is taken from the Mithraic symbolon given in Origen, Contra Celsum 6.22 (Beck, ibid., 73–85); the images of the grades and their tutelary planets are arranged in this ladder form in mosaic up the aisle of the Felicissimus Mithraeum at Ostia (V 299).

⁵⁵ For example, arrangements of the planets were used in both systems to define tutelary gods. But the products were very different. Whereas in the Commagenian royal cult the conjunctions of certain planets with the star Regulus in a certain year had been used to define the identities of the king's divine peers (above, n. 52), in the Mysteries a unique spatiotemporal sequence of the full seven was constructed to organize and validate the hierarchy of grades and to characterize progress through it: Beck, op. cit. (n. 33, 1988), I–II. The example is instructive in another respect, for it points up a key difference between the two religions: the royal cult focused the universal on the particular—on a particular moment in time, on particular local circumstances, and on one particular individual and his dynasty; in the Mysteries the ‘images of the universe’ were made universally applicable, functioning, at least potentially (and with the notorious restriction to the male sex), for the salvation of all. Because of its particularity, the royal cult was, finally, a non-exportable dead end; Mithraism injected a measure of egalitarianism into the cosmos, and so succeeded.

⁵⁶ See the important, though ultimately misdirected, 1931 article in which Cumont linked the newly discovered and idiosyncratic Dieburg relief (V 1247) to the magian hymns of Dio, , Or. 36 (39–61Google Scholar) to postulate a common eschatology transmitted from the ‘mages occidentaux’ to the Mysteries: ‘La fin du monde selon les mages occidentaux’, RHR 103 (1931), 29–96Google Scholar; contra, Gordon, op. cit. (n. 32), 237–41; Beck, op. cit. (n. 3), 2036–7; idem, ‘Thus spake not Zarathuštra: Zoroastrian pseudepigrapha of the Greco Roman world’, in Boyce and Grenet, op. cit. (n. 4), 491–565, at 539–48.

⁵⁷ Textes et monuments figurés relatifs aux mystères de Mithra, Vol. 1 (1899), 109, 120, 301 (‘c’ est là la doctrine capitale [viz. astrological fatalism] que Babylone a introduite dans le mazdéisme'); likewise much of Mithras' solarity: ibid., 200, 300 (‘il y a en réalité dans les mystères deux divinités solaires, l'une iranienne qui est l'héritière du Hvarè perse, l'autre sémitique qui est le substitut du Shamash babylonien, identifié à Mithra’), 303.

⁵⁸ (1929), 129–30.

⁵⁹ Above, n. 36.

⁶⁰ It is unnecessary to adopt here more than the minimalist position that Balbillus the astrologer was the grandfather of the Julia Balbilla who caused a poem to be inscribed on the colossus of Memnon in which she claims as her other grandfather Antiochus IV of Commagene (the grandfathers are styled respectively ‘Balbillus the wise’ and ‘Antiochus the king’): Bernand, A. and E., Les inscriptions grecques et latines du Colosse de Memnon (1960), 86–92Google Scholar, no. 29, lines 13–16, with commentary. The Bernands' correct reading of line 15 of this poem excludes, in my view, an earlier and widely current interpretation which gave Balbillus himself ‘a royal mother, (?)Aka’ (see also Gagé, J., Basileia (1968), 75–85Google Scholar). Balbillus' Commagenian mother was promoted by Cichorius, C., Römische Studien (1922), 390–8CrossRefGoogle Scholar, in the same prosopographical package which also made him the (unnamed) son of Thrasyllus mentioned by Tacitus, (Ann. 6.22Google Scholar) as predicting Nero's principate. One may, of course, retain the filiation while rejecting the earlier Commagenian marriage connection. That Balbillus himself married royalty (whether Commagenian or other) was argued by Schwartz, J., ‘Ti. Claudius Balbillus (préfet d' Égypte et conseiller de Neron)’, Bull. Inst. Franç. d'Archéol. Orient. 49 (1950), 45–55Google Scholar, at 48, and Gagé, op. cit. (above), 84. On Balbillus' daughter (Julia Balbilla's mother), see PIR ² 1.262, no. C 1086 (Claudia Capitolina). As a practising astrologer, Balbillus has the invidious distinction of advising Nero to divert the evil omen of a comet on to members of the nobility as surrogate victims (Suetonius, Nero 36). Balbillus was also among the ‘best’ astrologers whom Vespasian consulted, favouring him to the extent of allowing the Ephesians to institute games in his honour (Dio 66.9.2; on the records of these ‘Balbilleia’, see L Moretti, Iscrizioni agonistiche greche (1953), 184 and index s.v.). On the theoretical side, a fragment from Balbillus' astrological works is preserved (CCAG 8.3.103–4; 8.4.233–8, 240–4); it is concerned principally with the rather dangerous topic of length of life. Cumont, as historian of astrology, devoted a short study to him: 'Astrologues romains et byzantins: Balbillus', I., Mél. d'Archéol. et d'histoire … de l'École Franç, de Rome 37 (1918–1919), 33–8Google Scholar; see also W. and H. G. Gundel, Astrologumena (1966), 151–3. Balbillus, it is generally agreed, also had a varied and distinguished career as an equestrian functionary, a point of some significance if one is to cast him as a sort of godfather to the Mithraic Mysteries. Here again, I follow a minimalist consensus which identifies him with, inter alios, (a) the Balbillus who was the subject of a procuratorial career recorded in an Ephesian inscription (Keil, J., Forschungen in Ephesus 3 (1928), 127–8Google Scholar, nos 41–2 = AE 1924, 78) and (b) the Balbillus who was prefect of Egypt from 55 to 59 (Tac., , Ann. 13.22Google Scholar). For brief biographies which reflect at least this consensus (in addition to those by Cumont, Moretti, the Bernands and the Gundels cited above) see Magie, D., Roman Rule in Asia Minor, Vol. 2 (1950), 1398–400Google Scholar; R. Syme, Tacitus (1958), 508–9; Pflaum, H.-G., Les carrières procuratoriennes équestres sous le haut-empire romain, Vol. I. (1960), 34–41Google Scholar, no. 15. A ‘separatist’ position, breaking apart the various identities, was taken by Stein, A., ‘Balbillus’, Aegyptus 13 (1933), 123–36Google Scholar; and less insistently in PIR ² 1.184–5, no. C 813; cf. ibid., 349, no. B 38; cf. Schwartz, op. cit. (above) (less radical). A maximalist (or ‘unitarian’) biography, following Cichorius (above) and exploiting all possible identities and both Commagenian marriage connections, is woven into his history of Roman astrology by Cramer, F. H., Astrology in Roman Law and Politics (1954)Google Scholar, index s. ‘Balbillus’. Finally, one should mention, though with reservations, Gagé's picture of Balbillus and the Commagenian dynasty as deeply involved in the formulation of a ‘royalist’ ideology focused on the person of the emperor: op. cit. (above), 75–85, 108–17, 143–9, 155–63; more tentatively, Wagner, op. cit. (n. 37), 216–17.

⁶¹ In Balbillus' favour, one might point to certain Egyptian (or Egyptianizing) elements which appear to have entered Mithraic ideology, most notably in the person of their lion-headed god: Pettazzoni, R., ‘The monstrous figure of Time in Mithraism’, in Essays in the History of Religions (trans. Rose, H. J., 1954), 180–92Google Scholar. It is difficult to account for these motifs in most current scenarios of transmission. Balbillus, the curious polymath who prior to his governorship of Egypt had served as head of the ‘Museum and Library at Alexandria’ (see preceding note on his procuratorial career), would be a fine example of the type of conduit which must be postulated. Indeed, his whole persona resonates remarkably with that ‘bricolage’ of encyclopaedic learning which, as Gordon has so perceptively demonstrated (op. cit. (n. 4, 1978); ‘Reality, evocation and boundary in the Mysteries of Mithras’, JMS 3 (1980Google Scholar, reprinted in Gordon 1996), 19–99), characterizes the Mysteries. For example, Seneca, (Nat. Quaest. 43.2.13–15Google Scholar) reports an eye-witness account by Balbillus, whom he decribes as ‘the best of men and uniquely accomplished in every genre of literature’, of a battle between dolphins and crocodiles at the Heracleot mouth of the Nile; paradoxically, the more pacific creatures were the victors. Such animal lore is the stuff of the Mysteries: see Gordon, op. cit. (above, 1980).

⁶² Winter, E. and Schütte-Maischatz, A., ‘Neue Forschungen in Kommagene’, Historisch-Archäologischer Freundeskreis: Rundbrief (1997), 31–7Google Scholar. I am most grateful to Simon Price for alerting me to this discovery and sending me the initial report (above).

⁶³ On caves and rock-cut reliefs in Mithraism, see Lavagne, H., ‘Importance de la grotte dans le mithriacisme en occident’, EM, 271–8Google Scholar; Beck, R., ‘The rock-cut mithraea of Arupium (Dalmatia)’, Phoenix 38 (1984), 356–71CrossRefGoogle Scholar. On artificial caves in the Commagenian royal cult, see above n. 47.

⁶⁴ Mithraists from IV Scythica are attested at Dura (V 53, 62).