1 See Marian, Wenzel, Ukrasni motivi na steccima (Ornamental Motifs on Tombstones from Medieval Bosnia and Surrounding Regions) (Sarajevo, 1965), Pl. GV–CVII Google Scholar. The researchleading to this book and to the present article, which is being extended with added material into book form, has been sponsored by the Bollingen Foundation, New York, to which I owe considerable thanks.

2 Wenzel, M, “A Mediaeval Mystery Cult in Bosnia and Hercegovina,” Journal of the Warburg and Courtauld Institutes, XXIV (London, 1961), 95.Google Scholar

3 Wenzel, M, “Bosnian and Herzegovinian Tombstones—Who Made Them and Why,” Sudost-Forschungen, XXI (Munich, 1962), 102–43.Google Scholar

4 See also M. Wenzel, Ukrasni motivi, p. 287, Pl. LXXVI, Figs. 2-3, at Glumina, in Hercegovina.

5 Two stars, or rosettes, accompany two horsemen at Strizirep, Sinj region, and Voštane, Trilj region, Croatian Dalmatia (ibid., p. 383, Pl. GUI, Figs. 2, 3). See also ibid., p. 385, Pl. CIV, Fig. 7. Although falling today into different Yugoslav republics, these sites have a certain geographical unity and are practically all in the karst hinterlands and plains formed by the Dinaric Alps.

6 See ibid., p. 387, Pl. CV, and p. 389, Pl. CVI. There are more examples known of the woman between horsemen than of opposing horsemen without the woman. The interesting problem presented by the woman turning into a deer between horsemen, notably at Čengić Bara, Kladovo polje, Ulog region, Hercegovina, where the woman is depicted with deer's antlers (ibid., p. 391, Pl. GVII, Fig. 1), does not enter into the scope of this paper.

7 In Fig. 2, c, a column surmounted by a star appears between the two horsemen instead of a woman, but on the opposite end of the same tombstone there is a woman between horsemen, and the woman is given a headdress with rays (ibid., p. 389, Pl. CVI, Fig. 8).

8 Ibid., pp. 347-59, Pl. XGIII-XCVI.

9 See also ibid., p. 355, Pl XCIV, Fig. 12, at Miruše, site “Mistihalj,” Bileća region, Hercegovina.

10 See also ibid., p. 353, Pl XCIII, Figs. 13, 17; p. 355, Pl XCIV, Fig. 5; p. 357, Pl XCV, Figs- 9. i3. 15; P- 359. pl. XCVI, Fig. 6.

11 See Tudor, D, “I cavalieri Danubiani,” Ephemeris Dacoromana, Annuario della Scuola Romana (Rome, 1937), VII, 189–356 Google Scholar; and Tudor, D, “Nuovi monumcnti sui cavalieri Danubiani,” Dacia, N. S. IV (Bucharest, 1960), 333–62.Google Scholar

12 Tudor, “Nuovi monumenti,” p. 346, Fig. 10, gives a typical example.

13 Ibid., p. 338, Fig. 3; p. 348, Fig. 11.

14 Tudor, “I cavalieri Danubiani,” p. 298, Fig. 14; p. 313, Fig. 36; p. 324, Fig. 56; p. 333, Fig. 65; p. 337, Fig. 68; p. 338, Fig. 69.

15 Intercisa II, Archaeologia Hungarica, XXXIII (Budapest, 1957), 383, the section “Bleigegenstande,” Pl. LXXIV ff.

16 Ibid., Pl. LXXVI, Fig. 8; and Edith B. Thomas, “Ólom fogadalmi emlékek Pannoniában : Poganytelki olomőntő műhely (Monuments votifs en plomb surle territoire dela Pannonie : La fondrie de plomb de Poganytelek)” (précis), Archaeologiai Értesítő, LXXIX (Budapest, 1952), 32-37. H- IV, Fig. 5, and Pl. V, Figs. 5, 7.

17 See also Thomas, Pl. V, Figs. 2, 4.

18 Ibid., Pl. VI, Fig. 7, from Brigetio (Kállay-gyuj-temény).

19 Ibid., Pl VI, Fig. 1, from Brigetio.

20 Ibid., Fl IV, Figs. 1, 4, 6.

21 Michon, E, “Miroirs Antiques de verre doublé de plomb,” Bulletin Archéologique du comité des travaux historiques et scientifiques (Paris), No. 2, 1909, pp. 231–50Google Scholar; Michon, E, “Nouvellcs Observations surles miroirs de verre,” ibid., No. 2, 1911, pp. 196–207 Google Scholar; Tudor, D, ” Le dépot de miroirs de verre doublé de plomb trouvé á Sucidava,” Dacia, N. S. III (1959), 415–32Google Scholar; Veličković, M.,” Olovni okviri antičldh staklenih ogledala iz Narodnog Muzeja,” Zbomik radooa Narodnog Muzeja, 1958-59 (Belgrade, 1959), II, 55–72.Google Scholar

22 F., Chapouthier, Les Dioscures au service d'une déesse (Paris, 1935), p. 287, Fig. 56, and p. 288Google Scholar; Tudor, “I cavalieri Danubiani,” pp. 343-48, cat. nos. 120-28; Giglioli, G. Q., “Due gemme basilidiane del Museo Archeologico di Perugia Archeologia Classica, III (Rome, 1951), 199 ff., Pl. XLIX, L.Google Scholar

23 Gems possessing this feature are one formerly in the Odeschalchi collection, sketched in 1752 (Fig. 6, c), and another in the collection of the German Archaeological Institute, Rome (Tudor, “I cavalieri Danubiani,” pp. 344, 345, cat. nos. 121, 124, Figs. 78, 81). A gem in the Kunsthistorisches Museum, Vienna (Fig. 6, a), shows the sticks with snakes butlacks the feature of their being directed toward the prone figures below.

24 Tudor, “I cavalieri Danubiani,” pp. 230, 231; D. Tudor, “Intorno al culto dei cavalieri Danubiani,” Dacia, N. S. V (1961), 338. A. Delatte and Ph. Derchain (Les Intailles magiques gréco-égyptiennes [Paris, 1964], p. 193Google Scholar), in discussing figures standing on other figures which are represented horizontally, remark : “Némésis, Hecate et Osiris interviennent done, dans cette attitude, évidemment en qualité de divinités qui vainquentl'ennemi, e'est-á-direle mal. II est probable qu'il faille donner cette signification ál'ensemble dela serie. En particulier, les deux exemplaires qui montrent une déesse entre deux cavaliers font songer á une curieuse association de Némésis et des Dioscures, qui pourraient avoir été invoqués ici contre ‘l'ennemi’ en qualité de dieux saveurs.” The inheritors of this iconographic arrangement, St. George and St. Theodore, shown spearing a serpent and a man, sometimes with Christ between, are automatically interpreted in this way ( Wenzel, M, “Some Notes on the Iconography of St. Helen,” Actes du XHe Congrès International des Etudes Byzantines, III [Belgrade, 1964], 415, 416, 418 Google Scholar, Fig. 4).

25 Tudor, “Intorno al culto dei cavalieri Danubiani,” p. 321; Tudor, “I cavalieri Danubiani,” p. 237.

26 Chapouthier, cat. nos. 2-22, 26-54, 57, 58, 61-70, 72-75.

27 See also ibid., cat. nos. 26-29, 32-36, 40, 41, 45-48, 50-58, 60, 66, 69, 73-75.

28 See also ibid., cat. nos. 14, 20, 21, 26-30, 35-37, 41, 42, 44, 47-50, 53, 54, 57, 58, 64, 66, 69 73-75-

29 Ibid., cat. nos. 1-23, 26-37, 60, 61, 64-69, 73, 74.

30 See also ibid., cat. nos. 37, 38.

31 Ibid., cat. nos. 52, 56.

32 M. C. Waites, “The Meaning of the Dokana,” American Journal of Archaeology (Norwood, Mass.), XXIII (1919), 7, Fig. 6. Two Etruscan mirrors with this detail, similar to those discussed by Miss Waites, are preserved as part of the Ravestein collection in the Musées Royaux d'Art et d'Histoire, Brussels, one from Vulci (Fig. 7, c) and the other (cat. no. R. 1290) from Palestrine. I am extremely grateful to J.-Ch. Baity for allowing me access to this material. Fig. 6, b, presents this feature on an engraved gem.

33 See also Chapouthier, cat. nos. 59, 76, 79-87, 92-94, 99, 101, and also p. 280, Figs. 53, 54-

34 Ibid., pp. 176, 177, Fig. 17.

35 Ibid., p. 241, Fig. 36, shows twin amphorae on a monument dedicated to the Corybantes, identified with the Dioscuri, from Pergamon. Daremberg-Saglio, Dictionnaire des antiquités (Paris, 1889), Vol. D-E, s.v. “Dioscuri,” p. 255, Fig. 2437, depicts Spartan bronze coins with the heads of the Dioscuri surmounted by stars on one side and twin amphorae wrapped round with snakes on the other. Taranto, a Spartan colony, had on its money twin amphorae surmounted by stars. Fig. 8, a (the votive relief dedicated by Argenidas to the Dioscuri), shows them standing beside twin amphorae.

36 Plutarch, “De fraterno amore,” Moralia, 478, A, B.

37 “The ancient representations of the Dioscuri are called by the Spartans ‘beam-figures’ (dokana); they consist of two parallel wooden beams joined by two other transverse beams placed across them, and this common and indivisible character of the offering appears entirely suitable to the brotherlylove of these gods.” Ibid. (“Loeb Classical Library” edition, trans. W. C. Helmbold [London, 1939], VI, 247). The word translated here as “offering,” aphidrúmata, is translated by Jane Harrison (Themis [Cambridge, Eng., 1927], p. 305) as anything set apart, a dedication.

38 Rendel Harris, J., The Cult of the Heavenly Twins (Cambridge, Eng., 1906), Pl V Google Scholar, gives a photograph of this monument demonstrating that there is only one crosspiece, and none across the top. Drawings are given in Harrison, p. 305, Fig. 84, and Cook, A. B., Zeus (Cambridge, Eng., 1925), Vol. II Google Scholar, Part 2, Appendixes and Index, p. 1062, Fig. 916. The inscription reads, “Argenidas, son of Aristogenidas, to the Dioscuri, a vow.” Under the two dokana is written “Anakeion,” implying that they formed a double sanctuary to the “Lords.“

39 A. Bouché Leclercq, L'Astrologie grecque (Paris, 1899), p. 135, s.v. “les gémeaux.” The zodiacal sign of the Dioscuri is usually shown as a kind of “ H “ with two crosspieces, the two upper terminations of the “ H “ bending slightly outward. The two principal stars have been interpreted differently as Castor and Pollux, Apollo and Hercules, Apollo and Bacchus, Amphion and Zethus, and the Cabiri of Samothrace.

40 See P. M. Fraser, Samothrace, the Inscriptions on Stone (New York, 1960), Appendix IV, pp. 112-16. The relief shown in Fig. 8, b (second century B.C.), which mentions alist of mystai from Cyzicus, is usually taken to be a representation of a round building with two bases and has not previously been recognized as a dokana, but it seems to resemble most closely the dokana as it actually exists today. Chapouthier, p. 177, Fig. 17, sketches the dokana rather differently from the photograph on which Fig. 8, b, was based and discusses a barely distinguishable snake entwined around one of the uprights.

41 Waites, p. 8. Further published examples are those given in E., Gerhard, Etruschische Spiegel (Berlin, 1843)Google Scholar, Part I : Abbildungen, Pl XLV-XLVI. The dating of these mirrors is discussed in Mansuelli, G. A., “Studi sugli specchi Etruschi IV : La Mitologia figurata negli specchi Etruschi Studi Etruschi (Florence, 1948-49), XX, 61, 87 Google Scholar. Those with the crosspicce comelate in the series, about the third century B.C. This is roughly the same date as a South Italian red-figure vase, showing Greek influence, on which the Dioscuri stand with staffs and stars on either side of a divided circle on which is inscribed an “ H. “ This vase is preserved in the Museum of Decorative Arts, Prague.

42 Pausanias, XVI, 1-3.

43 Lucian, Dialogi deorum, XXVI.

44 Cicero, De natura deorum, II, 5-6; Ovid, Fasti, I, 705-8; Pausanias, IV, 27, 1.

45 Homer, Odyssey, XI, 300; Oppian, Cynegeticus, II, 14.

46 Xenophanes, Diels-Kranz, A.39; Pliny, Nat. Hist., II, 101; Lucian, Dialogi deorum, XXVI; Horace, Odes, I, Ode III ; Plutarch, Moralia, 426 C. See also K., Jaisle, Die Dioskuren als Rrtter zur See bei Griechen und Romern und ihr Fortleben in christlichen Legenden (Tübingen, 1907), pp. 26, 30, 36 ff.Google Scholar; and Rendel Harris, J., Dioscuri in the Christian Legends (London, 1903), p. 5.Google Scholar

47 Plutarch, Moralia, 944 D.

48 Jaisle, p. 30, quotes Ampelius, in his Scholia to Germanicus’ Aratea, as saying that the Dioscuri were present to the initiates at Samothrace and protected them from tempests on ships. N., Lewis, Samothrace, the Ancient Literary Sources (New York, 1958)Google Scholar, presents a number of sources to this effect, among which is the following : Gat. no. 179, Servius, In Aeneidem, 3, 12, says the Penates may be the Great Gods, and also the Great Gods may be two male statues before the door in Samothrace, representing Castor and Pollux. Pausanias, X, 37, 7, says the Dioscuri, Curetes, and Cabiri are identified as Boy Lords. Lewis, cat. no. 214, Strabo, 10, 3, 7, tells that people confuse the Corybantes, the Cabiri, the Idaean Dactyli and the Telchines with the Curetes, and that the rites of the Samothracians and of those in Lemnos and several other places have a common relationship, because the divine attendants are called the same. 49 Vows of secrecy taken at Samothrace are mentioned in Lewis, cat. nos. 141, 174, 174a, 221, 222, 224, 225, 229.

50 Ibid., cat. no. 179, Servius, In Aeneidem, 3, 12.

51 Fortis, Abbe A., Travels into Dalmatia (London, 1778)Google Scholar, is one of the earliest sources.

52 Duboko was a noted mining area in Roman times, though many of its inhabitants probably were always stockbreeders, as today. The villagelies on a small tributary of the River Pek (ancient “Pincus“), which flows from alarge cave a few miles from the village. The river was reputed to have carried gold. At the point where the Pek enters the Danube there are remains of the Roman castrum “Pincum” (M. M. Vasić, “Dionisos i na5 folklor,” Glas Srpske Akademije Nauka [SAN], GCXIV [Belgrade, 1954], Odeljenje drustvenih nauka, N.S. 3, pp. 133, 134. The festival of Rusalje, as celebrated there and elsewhere in the Balkans, was identified by Miklosich with the Roman rosalia or dies rosae, which was a festival for the dead (F. Miklosich, “Rusalien,” Sitzungsberichte der kais. Akad. d. Wiss. Phil.-Hist. CL, XLVI [Vienna, 1864], 386-405). The rosalia is, however, not mentioned in Ovid's Fasti among official Roman celebrations, and there is confusion about its date. It may have been imported into Rome from the Balkans, rather than the Balkan ritual deriving from Rome. Certainly, modern authorities recognize that the problem of the Rusalje is far more complicated than admitted by Miklosich. See F. Tailliez, “Rusalile, les Rosalies etla Rose,” Cahiers Sextil Puscariu, Linguistique— Philologie—Literature Roumains (Valle Hermoso, 1952), Vol. I, fasc. 2, pp. 301-17. I attended the ritual on June 2, 3, and 4, 1963, in the company of Dimitrije Stefanovic and Olga Moskovljević, from the Music Institute and Ethnographic Institute of the Serbian Academy of Sciences. I am grateful for their various notes, of which they have kindly allowed me copies. I am also grateful to Milica Ilijin, also of the Institute of Music, who visited the ritual at Duboko in 1952 and saw many additional details.

53 A disconnected translation from the Rumanian, as recorded by Dimitrije Stefanovic on June 2, 1963, is : “That you should come to have a talk. Gome on, Dad, forlunch and supper. We have come to meet you, and we are all waiting for you. Here is water for you to clean yourself. For along time you haven't washed or bathed.” This was apparently supposed to have really aroused the deceased, because when we hesitated about accepting food offered to us that had been on the grave, they said, “Don't worry, he won't come after you.“

54 Vasic, p. 130, mentions as a possible variant that there need be only two young men.

55 The traditional order is theleader with a sword, one virgin, the second young man, two virgins, and the third young man, who is sometimes, but not always, also carrying a sword.

56 The effective diameter is frequently reduced by hanging a round honey-cake or cooky with a hole in the middle over the mirror, so that the mirror is seen through the hole. The mirrors that are used are square pocket mirrors, but the effect is that of a small, round mirror, strikinglylike those of the Roman period which were found in the same area. It is hard to say whether or not this is coincidence.

57 In the nineteenth century, and until fairly recently, the instruments employed were bagpipes, as recorded in the earliest written source mentioning this ritual (M. Dj. Milićević, Knezevina Srbija [Belgrade, 1876], p. 1085).

58 H. Schwartz, “The Mirror of the Artist and the Mirror of the Devout : Observations on Some Painting's, Drawings and Prints of the Fifteenth Century,” in Studies in the History of Art Dedicated to William E. Suida (New York, 1959), pp. 102-4. Schwartz has demonstrated that smalllead mirrors, sold to pilgrims as souvenirs, were held up to reflect the image of sacred relics and thus capture some of their holy power, so as to take it away with them.

59 P. Kemp, Healing Ritual : Studies in the Technique and Tradition of the Southern Slavs (London, 1955). PP. 78-81.

60 Aristophanes, Clouds, 745-50, mentions bringing down the moon and keeping it in a round boxlike a mirror.

61 Apuleius, Metamorphoses, I, 2; II, 4.

62 Plutarch, Moralia, 417 A, The Obsolescence of Oracles, tells how the women of Thessaly were said to draw down the moon. This notion gained credence when one of them, skilled at astronomy, always claimed at an eclipse that she was bewitching the moon and bringing it down.

63 Lucian, Dialogues of the Hetaerae, I : “Her mother Chrysarium is a witch; she knows Thessalian charms and can draw down the moon.“

64 Apuleius, XI, 9 : “Mulieres candido splendentes amicimine, variolaetantes gestamine, verno florentes coronamine, quae de gremio per viam, qua sacer incedebat comitatus; aliae quae nitentibus speculis pone tergum r ever sis venientu deaeobvium commonstrarent obsequium.” There is no evidence that this particular detail of mirrors worn on the back was a feature of the cult of Isis as practiced in Egypt, and it may have been a result oflocal syncretism.

65 In addition to the smalllead casts representing Isis-Fortuna (see note 20 above), there are a number of engraved gems with the same motif in the collections of the Archaeological Museum, Zagreb, and the National Museum, Belgrade. No reproductions of these have been published. Other monuments relating to the cult of Isis found in the Danube region are given in V., Wessetzky, Die Ägyptischen Kulte zur Römerzeit in Ungarn (Leiden, 1961), p. 32 Google Scholar, Pl. III, Fig. 5; Pl. XIV, Fig. 18; Pl. XVI, Fig. 20. Monuments dedicated to Isis were found in both Zadar and Solin, along the Dalmatian coast ( Jacob, Spon, Voyage d'ltalie, de Dalmatie, de Grèce et iu Levant [Lyons, 1678], p. 76 Google Scholar; George, Wheeler, A Journey into Greece [London, 1682], Part I, p. 11 Google Scholar; and J. Zeiller, “Surles cultes de Cybèle et de Mithra, á propos de quelques inscriptions de Dalmatie,” Revue archéologique, XXVIII [Paris, 1928], 219). That the Isis cult penetrated to the Stolac region of Hercegovina is demonstrated by a monument found in the Bregava River, with certain Illyrian names and representations of the sistrum of Isis ( Sergejevski, D, “Nove akvizicije odeljenja klasi£ne arheologije Zemaljskog Muzeja,” Glasnik Zemaljskog Muzeja [Sarajevo, 1948], N. S. III, 169, Pl. I, Fig. 1Google Scholar). Further, the “ankh” or sign of Life, frequently related with Egyptian cults and the cult of Isis in particular, is represented six times on certain rustic medieval tombstones to the north and south of the Bregava River, all in Hercegovina (Wenzel, Ukrasni motivi, p. 109, Pl. XXVI, Figs. 25-28; p. 15, Pl. XXVII, Figs. 16, 17). It is a known fact that imported deities in these regions became popular if they could be affiliated to somelocal goddess or god.

66 We were told that if the deceased is a young, unmarried person, theleader may carry a flag decorated with his name, photograph, and certain of his belongings, as if the danceleader were acting the part of the dead person himself. This was observed by Milica Ilijin when she visited Duboko in 1952. A photograph alone may substitute for the flag. On the Sunday following the Duboko festival, at the nearby village of Neresnica, a similar festival is held on the occasion of Mali Duhovi. Here, in the summer of 1963, I observed such a flag, but the dances had degenerated beyond recognition.

67 This waslast observed by Milica Ilijin in 1952, but in 1963 we were told that a good head of a household (dobar domaHn) would take care to have it done for himself or that it could be done for aliving husband or wife at the time of the celebration for the deceased spouse. John Fine, Harvard, has kindly sent me information of a ritual observed by Mme. Holenkoff in 1938, in a village of the Morava region, very near Kragujevac. A slava could be held for a man who was very old, but not necessarily dying. The man paid the priest for his part in the prayers and also paid for food and drink, although he himself did not eat or drink any of it. Instead, helay under the table, which was set up at the entrance to the church, but not inside the church. The guests made merry, with or without music, and talked of the good qualities of the man under the table. The man could not have another funeral banquet after his actual death. Both of these customs relate to a kind of mystery initiation, in which the initiate has a “trial run” at death to assure his safe passagelater on.

68 Slobodan Zečević, “Ljeljenovo kolo,” Narodno stvaralaštvo (Belgrade), No. 9-10, Jan.- April, 1964, p. 708, remarks that only certain rare Yugoslav dances are carried on at night. As everybody knows that the souls of the dead are abroad at night, people usually avoid dancing at that time and terminate their dancing at sundown. These night dances are definitely chthonic in character.

69 When, on the early morning of the third day, June 4, 1963, shortly before dawn, we attempted tolocate the mistress of the house where we had been sleeping, she was nowhere to be found, as she was dancing in the kolo, whose participants could not be distinguished in the dark. We were told that it was an evil housewife (crna domaHca) who would not be dancing in the kolo at that time.

70 The main accessible sources for this part of the ritual are Majzner, M, “Dubočke Rusalje, poslednii tragovi iz kulta velike majke bogova,” Godišnjica Nikole Čupića, XXXIV (Belgrade, 1921), 226–57Google Scholar; Milićević, pp. 1085-86; Vasić, pp. 129, 164-67, Figs. 1-5, including a small map of the area; Gligorić, D., “Rusalje, srbovlaski narodni običaji iz srpskog kraja u Zviždu u Srbiji,” Bosanska Vila (Sarajevo), No. 8, 1893 Google Scholar; Milosavljević, S., “Srpski narodni obicaji iz sreza Omoljskog,” Srpski etnografski zoornik, XIX (Belgrade, 1913), 315 Google Scholar (this article contains many other interesting customs from this area, including the recipe for moon-cake quoted above).

71 Majzner, p. 233, gives four different versions in Rumanian and Serbian.

72 Ibid., p. 234.

73 Milica Ilijin said that the prophesies, when made, were singularly uninteresting, usually that such and such a dead person was all right. It was possible that the trance-woman, upon re - covery, did not prophesy at all, but she would agree that she had been with the vile (nymphs) or that a vila had been inside of her. When questioned further about the vila, she wouldlaugh and say no more.

74 In roughly the same part of the Balkans there are a number of cases of people falling into trances in order to remove sickness from a sick person and transfer it to themselves. The sick person is then healed, automatically, and they themselves are brought out of their trances by ritual dancing not unlike that which is performed at Duboko for the dead ( Tomić, Persida, “ ‘Vilarke’ i ‘vilari’ kod vlaških cigana u Temniću i Belići,” Zbornik radova, Etnografski Institut SAN, IV [Belgrade, 1950], 237–62Google Scholar; Arnaudov, M, “Kukeri i rusalii (Studii vŭrkhu bŭlgarskitie obredi ilegendi),” Sbornik za narodni umotvoreniia i narodopis, XXXIV [Sofia, 1920], 165–69Google Scholar). An additional ritual exists in the Homolje district which concerns the shifting of death from an endangered person to an animal of his choice, which is then killed and buried under a rose bush, this bush being identified with the powerful female supernatural being—the vila or rusalja or whatever—who has been about to seize the person and take him away. The endangered person is then resurrected (Kemp, pp. go-92).

76 MacCulloch, J. A., Medieval Faith and Fable (London, 1932), p. 170 Google Scholar, quotes a passage from St. Augustine (De Civ. Dei, XXII, 8) in which recovery from a trance is called restoration from death. MacCulloch gives other early medieval sources in which visionaries in trances are said to have visited Purgatory.

76 The word rusalje (and its variants) has the confusing characteristic of meaning, in both Serbo-Croatian and Rumanian, the souls of certain dead, female supernatural beings connected with rivers, the festival of Whitsun or Pentecost (Duhovi), a kind of dance, and certain dancers who usually, but far from always, dance at Whitsun (Majzner, pp. 227, 228; Ljubica S. Jankovid and Danica S. Jankovi6, Narodne Igre (Belgrade, 1948), IV, 185; Tailliez, pp. 301-17).

77 Janković and Janković, IV, 184, mention that an inhabitant of the Duboko region told them one didn't dare bathe in the river for a certain period of time beginning a few days before Pentecost, because if one did, the rusalje would enter one from the river. The same man remarked that when the women fell into a trance, it was said alternately that they had been seized (1) by the rusalje (uzele suje rusalje) or (2) by the Duhovi (uzeli suje Duhovi; the word Duhovi can mean ghosts or spirits as well as Whitsun, the Holy Ghost being Sveti Duh), or (3) that they had been taken by the Holy Trinity (uzelaje Sveta Trojica). We were ourselves told that the trance-woman had been entered by a vila and that she had been with the vile. A Duboko wise woman or witch whom we visited the second day of Duhovi, Nejatasa KurCilovic, was unable to prophesy at that time, as she and others told us, because there were so many vile about just then that they would surely enter into her and cause her to fall into a trance, and she was too old for this. It was possible for her to prophesy at other times without falling into a trance.

78 See p. 368 above and note 46. It will be observed that at Duboko in earlier times the swordsmen were responsible for carrying the trance-woman across water (Majzner, pp. 229, 230). Another mythological pair of psychopompic characters whose icons show strong similarities to those of the Dioscuri with the woman between are Hermes and Orpheus flanking Eurydice. The group illustrated on a Thracian coin of Gordianus Pius (A.D. 238-44) shows a small figure under the feet of Eurydice, similar to that under the feet of the woman between the Dioscuri on certain gems, for example, that in Vienna (Fig. 6, a) j see Guthrie, W. K. G., Orpheus and Greek Religion (London, 1952), p. 21 Google Scholar, Fig. c. Gems with this feature are given in Tudor, “ I cavalieri Danubiani,” pp. 343, 344, Fig. 79; and Delatte and Derchain, pp. 194, 195.

79 E. Lilek, “Vadjenje zive vatre u Bosni i Hercegovini,” Glasnik Zemaljskog Muzeja (Sarajevo, 1893), V, 35-36.

80 There are many sources mentioning the new-fire ritual in the Danube region. In Rumania it is confined to the southernmost part of the country, but the practice extends through Serbia, Macedonia, and Bulgaria, one may say in the direction of Samothracc. It is preserved in some mountain parts of Bosnia and Hercegovina, but I have found no mention of it in Croatia or Montenegro. The sources most relevant to this article are Trojanović, S., Vatra (Belgrade, 1930)Google Scholar; Lilek, V, 35, 36; Titelbach, V, “Rozněcováni ziveho ohné u Jihoslavanu,” Česky Líd (Prague, 1896), V, No. 4, 341–46Google Scholar; Kemp, pp. 145 ff.

81 Kemp, p. 149. That the ritual should be general knowledge was the case with certain antique mystery cults, although not with that at Samothrace in the classical period. See Lewis, cat. no. 141, p. 64, Diodorus 5, 77, 3 : “The initiatory rite which is celebrated by the Athenians at Eleusis … and that of Samothrace, and the one practiced in Thrace among the Cicones, whence Orpheus came who introduced them—these are all handed down in the form of a mystery, whereas at Cnosus in Crete it has been the custom from ancient times that these initiatory rites should be handed down to all openly, and what among other people was not to be divulged, this the Cretans conceal from no one who may wish to inform himself upon such matters.“

82 Kemp mentioned a fire-making machine in a village near Plav, Bulgaria, that had been set apart as a monument, similar to that described by Plutarch at Sparta : “Alarge apparatus of the kind described above was made entirely of a tree struck bylightning and was set up in the middle of the village street. It required considerable effort to produce fire and was not often done. It was also apparently a rite adapted to the usages on which I could not obtain definite information” (p. 146).

83 When applied tolivestock, the procedure is called volovska bogomolja and as such may be found in the ethnographicliterature; see B., Drobnjakovic, Etnologija naroda Jugoslavije (Belgrade, 1960), I, 219–23Google Scholar.

84 Kemp, p. E50. 88Ibid., p. 151. The charm to be said on coming through the tunnel is usually : “ I have crossed water and not been drowned. I have gone through the fire and not been burnt. I have gazed into the fat and not been drowned.” That which is gazed into is replaced by a mirror in a variant ritual. See note 92 below.

86 Cicero, De natura deorum, I, 42, says that the festival connected with the Cabiri at Lemnos, which related to the rites at Samothrace, took place at night andlasted for nine days, during which all fires of the island, which were thought to be impure, were extinguished. Sacrifices were offered to the dead, and a sacred vessel was sent out to fetch new fire from Delos. During these sacrifices the Cabiri were thought to be absent with the sacred vessel, after the return of which the new fire was distributed and a newlife began.

87 Kemp, p. 155.

88 Ibid., p. 150.

89 Ibid., p. 124. The “falling sickness” (pripadciva bolest, or padavica) bears a name strikinglylike that of the trance-woman, the padalica.

90 Ibid., p. 153. It is often thought that those who have been unconscious along time may receive the power of healing others. Alocal Yugoslav ethnographer reported that a sick person “falls as it were into an ecstasy and becomes inspired (vetrovit) … and a healthy person cannot be a doctor (vrazalc)” (ibid., p. 209).

91 Ibid., p. 153. It is interesting that the Samothracian twin gods were thought by the Romans to be the same as the Lars (Lewis, cat. no. 173, Arnobius, Adversus nationes, 3, 43) or, more frequently, to have been identical with the Penates, “which were carried off by Aeneas from Samothrace and brought to Italy, wherefore the Samothracians are said to be kinsmen of the Romans” (Lewis, cat. no. 179, Servius, In Aeneidem, 3, 12; and also cat. nos. 180-89). Dionysius of Halicarnassus (Lewis, cat. no. 186, Antiquitates Romance, I, 68, 2-4, and 69, 4) reports that the holy objects brought into Italy by Aeneas were images of the Great Gods of Samothrace and that these objects were kept by the holy virgins in the temple of Vesta. Plutarch (Lewis, cat. no. 188, Camillus, 20, 6-7) says that these objects were two small jars stored away, visible to the vestal virgins only. In any case, it is clear from the sources that the Dioscuri or Samothracian twins, who could be represented by twin sparks or fires, had some objects sacred to them deposited at the temple of Vesta, which was sacred to the hearth. Further, passing under thelitter of a vestal virgin while it was being carried along the street was punishable by death (Plutarch, Numa, 10, 4). Might this have been sacrilegious duplication of some sacred rite, similar, perhaps, to that mentioned above?

92 Kemp, p. 125.

93 There is, for instance, the parallel of the sea voyage taken by Argenidas, presumably from the side of the water where are the fire-making machines to the other side where are the two jars, relating to the route taken by a fire-ritual initiate, between the fires, across the water, and finishing with the two jars.

94 Tudor, “ I cavalieri Danubiani,” p. 325, Fig. 56. The same feature appears again on a stone tablet from Almus, Bulgaria (ibid., p. 313, Fig. 36).

95 Tudor, D, “Der Kult der donaulandischen Reiter,” Das Altertum (Berlin, 1962), VIII, 240.Google Scholar

96 See also Kemp, facing p. 130.

97 Lewis, cat. no. 167, p. 77.

98 Pausanias, I, XIII.

99 They are of the Moscow and Novgorod schools, dating from the fifteenth and sixteenth centuries. See K., Onasch, Ikonen (Berlin, 1961), pp. 394–95 Google Scholar; and Kondakov, N. P., Die Russische Ikone (Prague, 1929), Vol. II, Pl. 88 Google Scholar. Russian peasants have a rituallike the volovska bogomolja (see note 83 above and Fig. 15) but involving horses, new fire, and SS. Florus and Laurus (J. Rendel Harris, Dioscuri in the Christian Legends, p. 63).

100 Onasch, p. 394.

101 The most notable example is the tombstone from Zborna gomila mentioned in note 7 above. It has been at present removed to Sarajevo, but the scene was originally on the north side, the horses pointing toward the west end, on which the woman between horsemen was depicted, and away from the east end, which represented paired horsemen with figures beneath their horses’ hoofs. Paired stars are on both ends. There is here only one horse, and it is beingled by horsemen. See Wenzel, Ukrasni motivi, p. 279, Pl. LXXII, Fig. 11. The same scene appears again in the same graveyard, on another stone also showing the woman between horsemen, but by a different hand (ibid., p. 383, Pl. CHI, Fig. 7). The end with the woman and horseman is given on p. 389, Pl. GVI, Fig. 9.

102 This stone tablet in the Narodni Muzej, Belgrade, photo No. B 1240, closely resembles a similar tablet represented among the Dal Pozzo-Albani drawings of classical antiquities in the Royal Library at Windsor Castle; see Tudor, “ I cavalieri Danubiani,” pp. 342, 343, Fig. 76. It may be assumed that there is a single individual working the machine with the aid of the bow mentioned before.

103 The reason for the confusion in classical sources and representations as to whether the dokana had one or two crosspieces has not previously been explained.

104 The detail shown in Fig. 18, c, may, of course, be an adaption from a classical model, because it is almost unique in Western iconography.