¹ D. M. Robinson, Excavations at Olynthus, Part XII: Domestic and public Architecture (Baltimore, 1946), 344–357, pl. IGoogle Scholar.

² Metzger, H., Recherches sur l'Imagerie Athénienne (Paris, 1965), 59Google Scholar, no. 14, and 63–65, pl. XXVII, 4. Cf. also the references in Matz, , Sarkophage II, 229, notes 61–63Google Scholar.

³ Jeanmaire, H., Dionysos, Histoire du Culte de Bacchus (Paris, 1951), 351–372Google Scholar, on the Dionysiac legends as a prototype for Alexander. Note, however, that in the pompa of Ptolemy Philadelphus there are several images of Dionysus, of which the first is standing in a cart pouring a libation; the second is the representation of the Indian Triumph proper, with Dionysus reclining on an elephant driven by a satyr (Athenaeus V, 198, c, and 200, c-d). Diodorus (III, 65, 7–8, and IV, 3, 1) also describes Dionysus as celebrating his Indian triumph riding on an elephant. I do not know of any Hellenistic example in which the representation of Dionysus in his chariot is amalgamated with that of his Indian triumph; but the arguments of Matz, , Sarkophage II, 228–230Google Scholar, in favour of a prototype of the Hellenistic period for the Roman examples of the theme seem to me convincing.

⁴ The Triumph was also apparently represented on a lost panel from the Tomb of Praecilius at Constantine (Inv. Alg. 218); it is known only from a description which mentions a nude female, a draped man holding a (?) thyrsus, and a draped female: Victory, Dionysus, and Ariadne or a bacchante(?).

⁵ The exception is El Djem II, where Dionysus is nude; this mosaic is earlier than the others, and, as will be seen below (pp. 54, 58), belongs to a slightly different tradition. The long robe is found as a regular costume for Dionysus much earlier: on the Attic and Apulian pottery referred to in note 2 he normally wears it, and Hellenistic descriptions give him a purple robe and a golden girdle (e.g., Athenaeus V, 200; the same costume still in Nonnus, , Dionysiaca xx, 229–230)Google Scholar. But the effeminate nature of Dionysus is particularly emphasised in late antiquity: see Lenzen, V., The Triumph of Dionysus on Textiles of Late Antique Egypt (Berkeley and Los Angeles, 1960), 2–4Google Scholar.

⁶ Picard, G. Ch., ‘La Datation des mosaïques de la maison du Virgile à Sousse’, Atti del VII Congresso Internazionale di Archeologia Classica, III (Rome, 1961), 244–247Google Scholar.

⁷ Ibid., 245.

⁸ E.g., Inv. Sousse, 57.099, p. 48; cf. also Foucher, L., Hadrumetum (Paris, 1964), 216–217, note 805Google Scholar.

⁹ Picard, op. cit. (note 6), 245–246.

¹⁰ Thus the El Djem mosaic has very close parallels on other mosaics from that city, notably in the Maison de Silène, Foucher, op. cit., 27–28, pls. XI–XII; on the Cherchell pavement the scene forms an emblema-type panel in the centre of a rich carpet-pattern of a sort found at Timgad, Djemila, and Hippo Regius, but different from anything produced in the eastern regions of the African provinces at this time; and the Saint-Leu mosaic, like the others from the same house, has a very harsh, crude style, quite unlike that of the more sophisticated eastern centres.

¹¹ Picard, op. cit. (note 6), 245, note 16.

¹² The freely spreading vine-branches are found on African mosaics from: Lambaesis, Inv. Alg, 191; Oudna, Villa of the Laberii, room 33 and a fragment from room 31, Inv. Tun. 376 and 373; Hippo Regius, Marec, E., ‘Deux Mosaïques d'Hippone’, Libyca I (1953), 102–107Google Scholar, fig. 4; El Djem, Maison de Tertulla (El Djem III), all probably earlier than El Djem I, as well as on numerous later mosaics. For the origin and history of the formal scroll, see Toynbee, J. M. C. and Ward-Perkins, J. B., ‘Peopled Scrolls: a Hellenistic motif in Imperial art’, PBSR xviii (1950), 1–43Google Scholar, and Levi, D., Antioch Mosaic Pavements (Princeton, 1947), 489–500Google Scholar.

¹³ Compare the leopard drinking from a crater and ridden by a winged child on a mosaic from Constantine, A. Berthier, BAC 1938–40, 212–214.

¹⁴ See below, pp. 59, 65.

¹⁵ Foucher, L., ‘Un hypogée romain à Sousse’, Karthago iv (1953)Google Scholar, 88, pl. Ia.

¹⁶ Guidi, G., ‘Orfeo, Liber Pater e Oceano in mosaici della Tripolitania’, Africa Italiana vi (1935), 144Google Scholar; Aurigemma, S., I Mosaici della Tripolitania (Rome, 1960), 25Google Scholar.

¹⁷ Severus is flanked by his two sons, in contrast to the single figure of Victory behind Dionysus on the Sousse mosaic; and the attitudes of the horses do not in fact correspond to those of the tigers, as it is the nearest horse on the relief which turns its head back, the second tiger from the front on the mosaic (and on all the other mosaics where this motif occurs, except that of Cherchell, where the scheme is reversed to allow the further tiger to drink from the crater).

¹⁸ Picard, op. cit. (note 6), 245.

^18a Also at Torre de Palma, in Portugal; but the published photographs (C. de Azevedo, ILN December 24 1955, 1101–3) do not allow conclusions to be drawn about the overall composition of the scene.

¹⁹ Bellido, A. Garcia y, AJA liii (1949), 156Google Scholar, pl. XXIIIb. The style and type of composition suggest a date in the late Severan period. The Spanish mosaics on this theme are collected by Freijeiro, A. Blanco, Mosaicos Antiguos de Asunto Baquico (Madrid, 1952), 7–24, 28–36Google Scholar, figs. 4, 13, 14, 15.

²⁰ H. Sichtermann, Arch. Anz., 1954, cols. 383–388, Abb. 58. Freijeiro, Blanco, op. cit., 34, dates it 220–270Google Scholar; Becatti, G., Scavidi Ostia, Vol. IVGoogle Scholar: Mosaici e Pavimenti Marmorei (Rome, 1961), 197Google Scholar, no. 377, note 2, less convincingly, considers it Severan.

²¹ Blanco Freijeiro, op. cit. (note 19), 28, fig. 13.

²² Ibid., 34–36, fig. 15.

²³ Ibid., 35. I do not see why it is impossible that the figure should be a Victory; it is certainly female. The object in its hand has been restored as a round object or a disk on a stake, but only the end of the stake appears from the photograph to be original.

²⁴ From the Insula del Dioniso, Reg. IV, v, 9; Becatti, , Ostia IV, no. 377Google Scholar, p. 197, Tav. LXXXIII. The Triumph forms the central panel in a design of acanthus scrolls, with satyrs framed in scrolls on the diagonals. Becatti dates the pavement towards the end of the second century.

²⁵ From the Walramsneustrasse: Parlasca, K., Die römischen Mosaiken in Deutschland (Berlin, 1959), 40–41Google Scholar, Taf. 40–41. Second quarter of the third century.

²⁶ Gentili, G. V., Mansuelli, G. A., Susini, G. C. and Veggini, A., Sarsina: La Città Romana; il Museo Archeologico (Faenza, 1967), 56–60Google Scholar, Tav. XX–XXI. End of the second or early third century.

²⁷ Levi, D., Antioch Mosaic Pavements (Princeton, 1947), 93–99Google Scholar, pl. XVI, c. Late Antonine (?).

²⁸ Broneer, O., AJA, xxxix (1935), 61Google Scholar, pl. XVII, 2. The date is uncertain, but the ornamental motifs could suit a date in the later third century.

²⁹ Levi, op. cit. (note 27), 98, note 29.

³⁰ These sarcophagi are collected and discussed in detail by Matz, Sarkophage II, 3, b, α, 212–244, nos. 94–104B; also I, 165–166, nos. 58, 58A. See also the discussions of Turcan, R., Les Sarcophages romains à Représentations Dionysiaques (Paris, 1966), 238–249, 441–472Google Scholar, and Lehmann-Hartleben, K. and Olsen, E. C., Dionysiac Sarcophagi in Baltimore (Baltimore, 1942), 70–72Google Scholar.

³¹ Matz, , Sarkophage II, 274–278Google Scholar, nos. 138–141, pls. 160, 165, 161,1, 158,2. These all belong to the Severan period, mostly to the earlier part of it.

³² Ibid., 248–249, no. 105, pl. 134; date c. 190. Note that on the other sarcophagi with centaur-drawn chariots Dionysus is normally nude; only here does he wear the long robe. But the long robe had appeared earlier in the iconography of the centaur-chariot, for it is Dionysus' costume, though in a simpler form, on the early second-century mosaic from Acholla.

³³ Ibid (note 31), 228–229.

³⁴ See the references to these other works in Matz, , Sarkophage II, 224–226Google Scholar. All are very much simpler than the sarcophagi; like the mosaics they confine themselves to the principal figures, but preserve for these the main outlines of the scheme.

³⁵ Levi, Antioch Mosaic Pavements, p. 98: ‘Summing up, we may say that the first group of African mosaics examined has appeared as a certain derivation from an ancient and famous composition, the one with the Indian Triumph, although with variations and intrusion of motifs from different compositions …’

³⁶ Matz, , Sarkophage II, 226Google Scholar; Ashmole, B., A Catalogue of the Ancient Marbles at Ince Blundell Hall (Oxford, 1929), 93Google Scholar, no. 249, pl. 44. The rectangular form of the chariot here seems to me by no means as certain as Matz thinks, since its edge is hidden behind the bacchante in the left foreground; the top edge is represented frontally, but seems to be slightly curved.

³⁷ Some of the appearances of the theme in other media may antedate its use on sarcophagi; for example, a bronze medallion of M. Aurelius of the year 149, Gnecchi, F., I Medaglioni Romani Vol. II (Milano, 1912), 36Google Scholar, no. 72, pl. 65,7. Gnecchi describes the main figure here as Venus, ‘sotto i tratti di Faustina collo scettro, accompagnata dalla Fortuna pronuba (?)’, but in fact it must be an effeminate Dionysus with a thyrsus, and probably a satyr behind; the other figures, Pan leading the procession and dancing bacchantes, conform to the usual types. This antedates all the sarcophagi except perhaps the Vatican fragment and the Ince Blundell relief; it preserves the main line of the composition, but does not show the frontal stance of Dionysus, the turned-back head of the further leopard/tiger, or Victory on Dionysus' right and the crossed thyrsus and palm-branch. The handle of a silver dish in the Metropolitan Museum, said to come from Iran, should date to the mid second century, if genuine (Alexander, G., ‘A Roman Silver Relief’, Metropol. Mus. of Art, Bulletin xiv (1955), 64CrossRefGoogle Scholar; Matz, , Sarkophage II, 221Google Scholar, note 37, describes this as a probable forgery); it has the general composition of the prototype, with Dionysus frontally posed, but alone in the chariot.

³⁸ The wreath is not always clearly shown on the sarcophagi, where Dionysus' head often reaches right to the top of the panel; sometimes it interrupts the crowning border above it (at Lyons and on the Lateran sarcophagus with the elephant chariot); but it is omitted on the Capitoline sarcophagus, and represented in very low relief behind Dionysus' head on the sarcophagus of the Casino Rospigliosi. If the Sousse mosaic were derived from the prototype directly, or through the same intermediary as the other mosaics, the omission of the wreath on it would be inexplicable, since there is no lack of room for it above Dionysus' head. But it is comprehensible if the sarcophagi, with their long, narrow format, have intervened between the original and the mosaic; the omission is much more likely to be due to the designer of the sarcophagus cartoon than to be an error on the part of the mosaicist himself.

³⁹ The frontal design was not itself original; it had been used on the ceiling stucco from Sousse (above, note 15), which is perhaps late second century. But the design there is very much simpler, consisting just of Dionysus alone in the chariot; it is the adaptation of the whole triumphal group to a circular design which is original on the El Djem III mosaic.