¹ Provincial statuettes of bronze are also known, but will not be considered in this article which proposes to deal with a limited group of dedications and not with the iconography as a whole. In the main these statuettes seem to derive from the periphery of the region under discussion, although their provenances may not necessarily be considered the places of their execution. Five of these statuettes are known to me: (a) Louvre no. 655. Of unknown provenance. Ridder, Bronzes antiques du Louvre, I, no. 547, pl. 40, with bibl. The best reproduction is to be found in the Journal of Hellenic Studies, iii, 1882, pl. facing p. 207 (illustration to Waldstein article on pp. 107–110; he considered it the closest reproduction of the Hermes of Praxiteles, which is manifestly inaccurate), (b) Sofia, National Museum. From Gigen (the ancient Oescus). The child holds grapes in his right hand and a thyrsos in his left. Bulletin de l'Institut archéol. Bulgare, iii, 1925, pp. 248–253, fig. 67. (c) Zurich Antiquarium. From Vindonissa. Benndorf in Vorlegeblätter für archäol. übungen, series A, Vienna, 1879, pl. XII, 13. (d) Péronne. From Marché Allouarde near Roye (Somme). Danicourt, Revue archéol., iv, 1884², pp. 72–75, pl. 4. (e) Autun, Bulliot collection (?). From Champdôtre-les-Auxonne (Côte d'Or). Villefosse, Gazette archéol., xiv, 1889, pp. 95–101, pl. 19. This, more than any other provincial representation of the subject, can be related to a remote reflection of the Praxiteles statue, although the child derives from another tradition which I relate to Kephisodotos.

The problem of iconographic inspiration for these provincial sculptures will be discussed elsewhere. Despite the fact that scholars postulate dependence on the Hermes of Praxiteles in individual cases, there is generally no foundation for such connection. Of the monuments under discussion only two (nos. 1 and 6) portray Mercury with his right arm raised above his head, holding a bunch of grapes, and the child in all examples is turned in an entirely different pose from that of the Olympia statue. In the two reliefs which do recall that work, the ponderation of Mercury is in no way related. These representations go back to other classical prototypes and various currents are sometimes crystallized in the same sculpture.

² In addition to the monuments included in this list there are several problematical ones:

Ésperandieu, Récueil (hereinafter cited as E), XI, no. 544, gives an aedicula from Cannstatt in the Stuttgart Museum for which Paret (Germania, ix, pp. 6–13, fig. 10) reconstructs the figures of Mercury and Bacchus. This reconstruction is based on the dubious inference that the grapevine ornament of the aedicula refers to Bacchus. No piece of the child's body was discovered, and only minor fragments of Mercury and a goat.

E, V, no. 4483 quotes a fragment of a stele with Mercury and Bacchus which was included in Hildenbrand's catalogue of the Speyer Museum (Röm. Steinsaal, p. 31, no. 62d). Ésperandieu was unable to see it and suggests it might be in the storerooms. Said to be from St. Ingbert.

Henning, Denkmäler … Strassburg, p. 50, pl. xlvi, 7, gives a relief of Mercury who perhaps held a small Bacchus. The left side of this is very damaged and the photograph fails to reveal any traces of the child. From Görsdorf, Kr. Niederbronn.

³ E, V, no. 4413.

⁴ E, V, no. 4471.

⁵ E, V, no. 4491. Very damaged. Inscription (C.I.L. xiii, 4538): [D]eo [Mercurio …] sacra … v(otum) s(olvit) [l(ibens) m(erito)].

⁶ E, VII, no. 5653. He reproduces also a drawing from Ravenez, Alsace illustrée, iii, pl. XIII, 1 which shows fragments since lost. Inscription (C.I.L. xiii, 6034): Deo Me(rcurio) S. [S]everus ?] equoni e]x p(romisso) l(ibens) l(aetus) m(erito). Dates in second decade of the third century A.D.

⁷ E, VI, no. 5126.

⁸ E, VIII, no. 5908.

⁹ E, X, no. 5969.

¹⁰ E, VI, no. 5639; Henning, op. cit., p. 50, pl. xlvi, 8. Surface hacked away.

¹¹ E, VII, no. 5605. Fragment of the lower abdomen of Mercury, part of his left arm, and the drapery wrapped about it. There are said to be traces of Bacchus, but these are not distinguishable in photographs. It may be accepted on the basis of no. 8 from the same site of the Wasenburg Mercury Temple.

¹² E, VII, no. 5569; corrections and illustration in X, pp. 23–24. Inscription (C.I.L. xiii, 11684a): Mercurio sacrum, Gentilis, Africani f(ilius), v(otum) s(olvit) l(ibens) [m(erito)].

¹³ E, VII, no. 5494. Lower part destroyed in fire of 1870.

¹⁴ E, VII, no. 5528. Bacchus destroyed, but traces of his arm remain.

¹⁵ Germania Romana, iv, pl. XIX, 1; Mezger, Die röm. Steindenkmäler … im Maximilians-Museum, p. 20, no. xii. See below for problems in identifying the child.

¹⁶ Benndorf, Archäologische-epigraphische Mittheilungen, ii, 1878, pp. 1–9, pl. 1. Found with a votive altar to Mercury and four other monuments dedicated to the god.

¹⁷ E, XI, no. 294. Mercury's head and knees, part of his mantle and almost all of the goat at the left are restored. The child is also restored, save for the head and a piece of the torso. This Mithraeum dates from 210 A.D.

¹⁸ E, XI, no. 309. Inscription on socle (C.I.L. xiii, 11788a): D(eo) i(nvicto) M(ithrae), Mercurio, Q(uintus) P(…) Gemellus v(otum) s(olvit) l(ibens) l(aetus) m(erito).

This seems to be the unique example in major art of a seated Mercury holding the infant Bacchus, although similar iconography appears on a terracotta relief roundel in Turin (Heydemann, 3rd Winckelmannsprogramm, Halle, 1879, p. 43, no. 46 — not illustrated) and on Roman coins of Sagalassos in Psidia (Vorlegeblätter, series A, pl. XII, 5 a and c). For the coins see von Sallet, Zeitschrift für Numismatik, vii, p. 57. Neither coin agrees with the other and the looseness of the compositions argue against inspiration from a statuary representation.

¹⁹ E, XI, no. 629.

²⁰ Graeven, , Zeitschrift für bildende Kunst, xiv, 1905, p. 170Google Scholar. His figure 5 is a better reproduction of the stele than that of Ésperandieu.

²¹ Cavedoni, , Bullettino dell'instituto di corrispondenza archeologico, 1834, p. 108Google Scholar f. Cf. Schnoeringer (ibid., p. 44) who says no more than that it is a very curious allegory with Mercury holding a “sick child.”

²² Bonner Jahrbücher, xii, 1848, pp. 17–20, pl. V, 1.

²³ For example: Froehner, Terres cuites d'Asie de la coll. J. Greau, ii, pl. 91 — mould with caduceus and dolphin.

²⁴ Steindenkmäler, pp. 20–21.

²⁵ Arte romana sul Reno, p. 19 f., p. 186; cf. Arndt-Amelung, Einzelaufnahmen, no. 1065.

²⁶ If this winged figure could be interpreted as a soul, it would, on the contrary, substantiate the following theory of transmigratory significance.

²⁷ For example, in the Louvre vase G 755 (reproduced in Rizzo, Prassitele, pl. 105, 1) where the attendant figures are a satyr and a maenad. Similarly, the Reuss vase reproduced in Vorlegeblätter, pl. XII, 1a.

²⁸ Head, Principal Coins of the Greeks, pl. 24, no. 49; p. 42. Seltman, Greek Coins, pl. xxxv, 12; p. 166. Head dates these 400–336 B.C. Frequently the inscription is omitted.

²⁹ Furtwängler, Die antiken Gemmen, i, pl. 16, 54; p. 80; pl. 18, 12; pp. 87–88. Archaic style.

³⁰ For a provincial example of the type, see E, I, no. 111. Hanfmann, , American Journal of Archaeology, xliii, 1939, p. 231Google Scholar and note 2, fig. 2, discusses an interesting sarcophagus in Zagreb on which the reversed type is used for Mercury carrying Bacchus, with whom the deceased child is equated.

³¹ For example, the frescoes of the Tomb of Vibia: Garuci, Storia dell'arte Christiana, vi, pl. 494.

³² Cf. Lehmann-Hartleben and Olsen, Dionysiac Sarcophagi in Baltimore, p. 47, note 151. Dolphins appear in Etruscan tomb decoration and on innumerable Roman grave monuments.

³³ This statement is based upon preliminary research on the subject of stylistic development in the Roman provinces. For this group of sculptures, whatever there is of external evidence confirms the conclusions reached by stylistic analysis. The Augsburg and Carnuntum reliefs seem to be earlier, probably belonging to the late second century.

³⁴ Ritterling, Real-Encyclopädie, xii, col. 1653.

³⁵ Münzer, F. and Strack, M. L., Die antiken Münzen von Thrakien, p. 223, no. 427; pl. VI, 12 (Anchialos). Filow, Numismatische Zeitschrift, 1918, pp. 39–41; pl. IX, 1 (Pautalia). Ruzicka in Bulicev Zbornik, 1921, pp. 667–670; pl. XIV. These coins have been considered reproductions of the Hermes of Praxiteles (Ruzicka), of the Kephisodotan Hermes and Dionysos (Rizzo, op. cit., pp. 9–10), and of the same subject by Timarchos (Klein, Jahreshefte des österreich. archäol. Institutes, xiv, 1911, pp. 98–111). This is not the place to elaborate the controversy, but Klein seems at least justified in placing the original statue after both Praxiteles and Kephisodotos, since the composition borrows something from each. The type of Hermes and Dionysos reflected bears no relation to our provincial examples; Hermes leans against a colonnette on which the child is seated. The coins post-date the departure of the VIII legion from Moesia — i.e. in the known impressions — but may serve as an indication that statues of the subject were set up in these Thracian centers and may have been familiar to Moesian recruits. — Cf. Filow, op. cit., p. 41.

³⁶ For the rôle of the VIII legion, see Cumont, Textes et Monuments, i, p. 256. At least in the Rhine and Danube areas, the spread of the Mithras cult was achieved by the army. As Cumont says, there was good reason for one of the grades to be called milites — “Mithras,” Roscher's Lexicon, 11², col. 3033.

That this diffusion was accomplished without the intermediary of Rome is assured by the discovery of little bas-reliefs in Rhenish Mithraea which are of certain Pannonian or Moesian provenance, as well as by the fact that the cult monuments of the Rhone valley show no relations to the Rhenish ones. — Cumont, Les mystères de Mithra, p. 51.

³⁷ Wüst, “Mithras,” Real-Encyclopädie, xv, col. 2153.

³⁸ Lehner, Bonner Jahrbücher, cxxix, 1924, pp. 51–52.

³⁹ Cumont, Textes et Monuments, ii, p. 153, no. 423 from Recking near Heilbronn (dated 148 A.D.); p. 154, no. 430, an altar from the Mithraeum of Gross-Krotzenburg; p. 165, no. 506 from Geneva (dated 201 A.D.).

⁴⁰ Ritterling, Real-Encyclopädie, xii, col. 1654.

⁴¹ No. 580 in the Museum. Inscription on lyre: Ann(o) col(oniae) CLXXX invicto deo Mithrae sacr(um) G. Accius Hedychrus pater a(nimo) l(ibente) p(osuit) — 159 A.D. Lantier, Inventaire des monuments … de la péninsule Ibérique, pt. I, p. 3, no. 7; pl. IV, fig. 7, with bibl. For the discovery, see Archäol. Anzeiger, 1914, col. 377, fig. 52.

⁴² Fragment: E, XI, no. 292; E, XI, no. 295.

⁴³ E, XI, no. 146.

⁴⁴ Behn, Mithrasheiligtum zu Dieburg, pp. 32–33, nos. 8–9, figs. 33–34; pp. 37–38, no. 22, fig. 46 (small altar with inscription Deo Xancto Mercurio, etc.).

⁴⁵ Forrer, Mithra-Heiligtum von Königshofen, pp. 49–50, pl. XVI, 3.

⁴⁶ Cumont, Les mystères de M., p. 146, note 2.

⁴⁷ Cumont, Textes et Mon., i, p. 146.

⁴⁸ Ibid., p. 147; ii, inscriptions no. 15, 17, 19, 20, 22, 24, 147 and 522. Cumont and Wüst (Real-Encycl. xv, col. 2145) consider these as evidence of the cult alliance of Mithras and Bacchus. Actually, they prove no more than that one person might belong to — or even officiate in — more than one mystery cult. The inscription no. 522, from Spain, does imply that Mithras, Liber and a Genius were cult companions.

⁴⁹ Ibid., p. 320.

⁵⁰ Cumont, Les mystères, p. 146, note 2.

⁵¹ E, XI, no. 138. The main scene on the reverse shows Mithras and Sol standing behind a reclining bull. Mithras holds a drinking horn and Sol offers a bunch of grapes — Leipoldt, Die Religion des Mithra (Haas, Bilderatlas zur Religionsgeschichte, lief. 15), p. xi. Leipoldt calls the scene a sacred meal.

⁵² Cumont, Textes et Mon., i, p. 147; ii, no. 69. There is a strong possibility, it must be admitted, that the grapes held by Mithras may be restored. Cumont quotes Maionica (Archäol.-epigraph. Mittheilungen, ii, 1878, p. 39) as a proponent of authenticity.

⁵³ See above, note 18. Cf. Leipoldt, op. cit., no. 26–27, pp. xiii–xiv.

⁵⁴ C.I.L., ix, 425; x, 6219.

⁵⁵ Inscription from Apt (Vaucluse): Revue épigraphique, v, 1903, no. 1545. Cf. Cumont, Les mystères, p. 146, note 2.

⁵⁶ As early as 1727 Dom Martin had recognized “la métamorphose qu'avoient fait les Gaulois de Mercure en Mithras.” — Religion des Gaulois, ii, p. 105.

⁵⁷ Behn (Mithrash. zu Dieburg, p. 47) suggests a syllogistic explanation: Mercury was equivalent to the Germanic Wodan, whereas the latter was affiliated with Mithras; hence Mercury became identified with Mithras. This theory is elaborated in his article in ΑΓΓΕΛΟΣ ii, 1927, p. 165 f.

⁵⁸ Les mystères, p. 146, note 2. Cf. Höfer, “Psychopompos,” Roscher's Lexicon, iii², col. 3256–3257.

⁵⁹ Cumont, Les religions orientales dans le paganisme romain, 4th ed., p. 264, note 90.

⁶⁰ Julian, Caesars, 336c: “καὶ ἡνίκα ἂν ἐνθένδε ἀπιέναι δέῃ, μετὰ τῆς ἀγαθῆς ἐλπίδος ἡγεμόνα θεὸν εὐμενῆ καθιστὰς σεαυτῷ.”