1. Armor types: Waurick 1983, 265ff. At pp. 277, 288, and 292 (the BR cup) Waurick characterizes this armor as a Hellenistic-style Lederpanzer worn by higher-ranking military personnel, though unsure whether this is just artistic convention; I see no reason why it should not reflect actuality. As for the possible Praetorian identification, the same armor is worn by the figure in the central group on the Primaporta Augustus cuirass. One can reasonably suggest that he is one of the emperor's special Praetorian attendants. Héron de Villefosse compares Trajanic coins with a similar figure behind the seated emperor at a military audience ( BMCRE III, pls. 20.10, 19.19; MonPiot 5 [1899]: 139). Primaporta Augustus soldier: Kähler (1959, 17-18, figs. 11, 16) wavers between Tiberius and a personified populus Romanus, two dominant opinions; Simon 1986, 55, s.v. pl. I, "Mars." Although Kähler et al. call the band across his cuirass a general's band ( Feldherrnbinde ), the BR guard, clearly no general, has one too. The dog does mark the figure as symbolic of, for example, the Praetorians (?), as do his shoes: see Goette 1988a, 410 ("Mars"). Not the personified Roman people: there is no parallel, nor can I see why the available personification of the Genius of the Roman People should have been superseded in this one instance. Not an imperial prince: all such portraits known are bareheaded. The profile rendering and plain helmet rule out Mars.

2. An early example of a favorite spatial device of later historical relief, e.g., the Domitianic Cancelleria adventus panel; Strong 1988, fig. 71. Cf. the Julio-Claudian Arcus Novus fragment (fig. 12) and the figures flanking Augustus on the Ara Pacis.

3. A thick shaft terminates in a rounded knob over the figure's left shoulder. Not the right shape or length for a standard or a weapon, perhaps it is a trumpet; the Gallic carnyx has a similar profile, though with an animal-head bell.

4. A similar contrast is evident between the documentary and allegorical relief fragments from the Arcus Novus, from Ravenna, and on the Ara Pacis and Sorrento base (figs. 8, 12, 15, 71f.).

5. Arretine ware matrix: of M. Perennius. Compare the tableau-like quality and comparative lack of meaningful narrative structure in the Aquileia dish (fig. 17). There is a matrix for this cup in the Metropolitan Museum and another in the Boston Museum of Fine Arts, as well as an impression in the Louvre. Fragments are common in museum catalogues. The New York matrix illustrated by Zanker (1988, figs. 45a-b) has been shown now by Francesca Porten Palange to be a forged version, but she assures me that the scheme does exist in authentic versions; the examples of this important scheme need a full investigation. I thank Joan Mertens of the Metropolitan Museum for directing me to Dr. Porten Palange's work, which tries systematically to distinguish original Arretine from modern reproductions of what remain, however, authentic schemata. See F. P. Porten Palange, "Fälschungen in der Arretinischen Reliefkeramik," Archäologisches Korrespondenzblatt, RGZM-Mainz 19 (1989): 91-99.

6. LIMC II (1984), s.v. "Ares/Mars," 383.

7. Drusus: Héron de Villefosse in the original publication ( MonPiot 5 [1899]: 134-40, 154), cited only by a few who are likewise predisposed to see individual portraits (see Kiss 1975, 96). Vermeule (1963, 35) and Pollini (1978, 289-90) speculate on Drusus very briefly; Küthmann (1959, 77) (after Willers) saw Gaius Caesar, taking the scene as the submission of the Sugambri and stretching Dio 55.6.4 to imply that Gaius was present with Tiberius and Augustus at that formal submission in 8 B.C. and presided over it (!) (on modern bias toward Gaius and Lucius see chap. 8); Simon (1986, 245) sees also a Julio-Claudian prince (cf. his bare head) but proposes Germanicus by comparision with the Gemma Augustea (Augustus-Germanicus-Tiberius). Most make no portrait identification nor report earlier attempts (Hölscher, Gabelmann, Kleiner, Ryberg, Zanker).

8. The cuirass type and its significance: Waurick 1983, 298-99.

9. See pp. 165f. Due to his premature death, portraits of Drusus are rarer than those of his brother Tiberius. Many are products of retrospective by his son Claudius, though based on prototypes from his own lifetime; cf. Fittschen and Zanker 1985, 29. Also, identification of Julio-Claudian princes is notoriously difficult, as their features are often very similar. Identification often hinges on close observation of little locks of hair fringing their faces in different, set patterns. This tool cannot be used here, given the scale of the cup. Nevertheless, the features here outlined do correspond to typical portraits of Drusus the Elder.

10. Ex Coll. Albani. For thorough plates (front, back, and both profiles), see Kaisersaal 1986, 60-61 (Tiberius), 62-63 (Drusus). See Fittschen and Zanker 1985, cat. 10, 22; Fittschen 1977, cat. 13, n. 16; Jucker 1977, 217f., fig. 4; Rose 1987, chap. 2, "Drusus the Elder."

11. My work, now quoted by Rose 1990, 460. "Submission": Zanker 1988, 229; Baratte 1986, 72; Simon 1986, 144; Gabelmann 1984, 127-31, 133, cat. 41; Gabelmann 1986, 285f.; Hölscher 1980, 282f.; Pollini 1978, 285-86. The most recent work on imperialism in Vergil adduces the panel as "Augustus receiving the homage of a varied group of conquered barbarians, men, women [ sic ] and children" (Hardie 1986, 368). Baratte says the children throw themselves at Augustus' feet! See pp. 99-100, 164-65 below. On the Torlonia relief, not part of this debate, see chap. 7, pp. 166-67 and p. 289 n. 19 (fig. 89).

12. Ryberg 1967, 61f., fig. 44; Gabelmann 1984, cat. 80; Strong 1988, fig. 132. This relief in its own way "ennobles" the barbarian father and son, by borrowing a composition developed to show the wounded Aeneas; the mourning Ascanius and stoic Aeneas have, however, been crumpled and twisted into anguished figures of defeat. The prototype Aeneas panel or painting must have been in Rome; the extant 4th Style copy from Pompeii would of course not have been available to the second-century artist. Casa di Sirico VII 1.23, Naples, MN 9009; Coll. Napoli (1986), 63 (col.), cat. 209; Alinari P.2.N. 12012.

13. Gabelmann 1984, 182-88; Brilliant 1963, 157, on fig. 3.132; see chaps. 5 and 7. These sarcophagi, like the Antonine panel, try to exemplify the virtus and clementia of the emperor/protagonist toward "zumeist besiegten Feinden" (Gabelmann, 128, on the BR scene).

14. See chap. 7. This naked emphasis on conquest characterizes other Julio-Claudian depictions of barbarians with their women and/or children; cf. the lower exergues of the great cameos (Gemma Augustea, fig. 16; Grande Camée de France). This is also true of cuirass statues, as they are by nature triumphal monuments. See p. 166 for the three Julio-Claudian cuirass statues (Copenhagen, Grosseto, Pal. Colonna) with child captives.

15. In triumphal processions captives may walk calmly, but the context here is not exactly benevolent. Consider the Julio-Claudian architrave (?) fragment from Rome in figure 84 ( ex Coll. Farnese; Naples MN 6722 [7516]; all heads restored; see also pp. 166, 289 n. 23): in a triumphal procession walk two Celtic captives (tunic, sagum, braccae ), each with a hand on the shoulder of a boy (tunic, sagum ) before them in the foreground plane; Koeppel 1983a, cat. II, fig. 12. Children were in fact sometimes led in triumphs—the infant Juba II, for example, in Caesar's North African triumph (App. BCiv . 2. 101).

16. Several claim that something like the BR scene is allegorized on the so-called Hoby cup, where Priam begs Achilles for the dead body of Hector—Vermeule 1968, 136; Gabelmann 1984, 142, cat. 46, pl. 14.1; 143, cat. 47, pl. 14.2, an Arretine ware reproduction.

17. Cf. the similarly deviant lid panel of the Ludovisi sarcophagus (fig. 88), which paraphrases the prototype of the BR scene; see chap. 7.

18. Simon 1967, 21, pls. 3, 17-19, 21; Simon in Helbig ⁴ II, 686. Simon (1986, 73, fig. 90a) called the south wall child simply a "Barbarenprinz," a hostage (74) to guarantee peace as in RG 32.

19. Contra Pollini (p. 264 n. 27) this is not an exomis ; the sleeve shoulder strap has simply slipped.

20. For this unpublished detail see the negatives DAI neg. 8817-18.

21. Picard 1957, pl. 9. Infants' dress in many cultures mixes aspects of the adult dress of either sex.

22. The relief may quote the Ara Pacis; the baby staggers toward his mother with arms out and head (now mostly gone) thrown back, in a not dissimilar fashion. His mother's hairstyle seems Tiberian or Claudian. Espérandieu 1938 (suppl. XI.2), cat. 7649.

23. A Gallic cult-figure type is a bearded male divinity with a child across his lap; one elaborate example shows a torqued male baby; Espérandieu 1911, cat. 2882, Auxerre (compare, for instance, cat. 3017). In Augustan art: a torque frames some of the Jupiter shield images of the Forum of Augustus, alternating with the better-known Jupiter-Ammon heads (fig. 69); see p. 82. On children: cf. a marble torso of a naked baby in Providence (RISD Mus. inv. 26.158); Ridgway (1972, 93-94) links Simon's Ara Pacis "princes" to this torso, identified as a barbarian child resident in Rome. This piece may exemplify friendly interest in northern children but might also belong to a trophy group of captives.

24. Some Hellenistic females wear a bracelet on the upper arm, but for males it seems a Gallic trait, observable on Gallic honorific statues in the Greco-Roman period. See the fine Augustan warrior (1.9 m; headless) with sagum and shield from Mondragon (Avignon, Mus. Calvet), and two right arm fragments from a multifigure limestone warrior group at Entremont. Espérandieu 1907, cat. 271; 1966, cat. 8662, 8665, pl. 12; Eydoux 1962, fig. 63 (Mondragon).

25. Simon (see p. 263 n. 18) followed by Ridgway 1972; Gercke 1968, 136-40; Rose 1987, chap. 2, "Gaius Caesar." My own arguments were presented at the AIA-APA 1986-87 meetings as "Lost Episodes in Augustan History"; at the same convention Rose accepted my identification of the Gallic child, identifying the other as from Pontos. Simon called both Eastern. Pollini (1986a, 453 n. 3) now calls both Gallic and (1987, 27 n. 49) publishes the agreement with my views that he had previously given in conversation, citing my evidence, though without acknowledgment. Dodging the issue are Kleiner (1985, 110: "a boy") and Syme (1984), who does locate Gaius in the north wall boy with Julia (426-27). See now Rose's definitive 1990 article sorting out the "Princes" Gaius and Lucius from the "Barbarians''; for Koeppel 1987 and 1988, see p. 264 n. 29.

26. Torelli 1982, 60 n. 72: "It is perfect nonsense from the Roman point of view to imagine (as Simon . . . does) that these boys were barbarian (?) princes, 'guests' (?) of Augustus." He was then commended by his reviewer, Smith (1985, 226). Compare La Rocca on the inner altar personifications; see p. 265 n. 29. On Roman definitions of pax see pp. 88-93 above.

27. The view that Gaius (south wall Oriental) and Lucius Caesar (north wall BR child) are here costumed for the Lusus Troiae (cf. Vergil's two-line description of these games as conducted under Aeneas) was especially promulgated by Pollini's 1978 dissertation (retracted; see n. 25), which is often cited; cf., still, Zanker 1988, 217f.; La Rocca 1983, 24, 30-31; Gabelmann 1985, 522; Torelli 1982, 48 n. 72. One need not belabor here the fallacies of this reconstruction (or Pollini's hypothesis [1987, 27 n. 43] that this is Vipsania Agrippina): the BR comparison makes it clear that "Lucius" is a Celt. Gaius' portrait can be identified on the camillus on the north wall; Rose 1990; Pollini 1986a, 453 n. 3, 454 n. 8; Pollini 1987, cat. 4 Type I, 42-43 and 21ff.; Rose 1987, S.V., 280. The Oriental has no bulla (which would not be removed for the dangerous Lusus Troiae!), completely non-Julio-Claudian features, and a mother who wears a diadem.

28. Note that the scanty dress of this high-ranking baby definitively sets the occasion of the procession frieze at a very warm time of year!

29. Cf. Koeppel's catalogue articles on the Ara Pacis in his Bonner Jahrbuch series; Koeppel (1988, 104-5) notes that Rose's "forthcoming" article settles the question; the 1987 catalogue of figures (cat. 5.30, 31, south frieze; cat. 6.35, north frieze) had asserted the foreign status of the children and mother as probable.

30. Well visible in the color photo essay on the Ara Pacis in FMR 10 (January-February 1983): 106. The diadem: Smith 1988a, 34f., 38 (its adoption by the non-Greek rulers of Bithynia, Cappadocia, and Pontus), 43 (its use by queens).

31. See RE 2.70 (1961) 1226-75 (R. Hanslik).

32. Most recently on Agrippa's activities: Halfmann 1986, 26, 163-64, 165-66. He made a stay at the Hellespont in 16 B.C. and sailed through the Bosphoros to Sinope and Amisos in 14 B.C. Rose 1990, 453, 457f.

33. Gem portrait in Beazley 1920, cat. 98 and pl. 6; and see Rose 1990, n. 13,

34. Large torque with finely striated tubes, large round finials. With a Cybele dish, Berlin, Charlottenburg Antikenabt. 3779.3-4. Barr Sharrar 1987, 138H, 14-15, pl. 70; LIMC II (1984), S.V. "Attis," no. 345 (compare no. 368). In the Hellenistic period a torque was also used for the iconography of "Asiatic" religious figures, like the syncretic Dionysus-Amor wearing a braided torque with ram's-head finials on a gold medallion now in the Louvre, ex de Clerq Coll.; Hoffmann and Davidson 1965, 228-31, cat. 93, pl. VI, fig. 93a-b.

35. See Smith 1988a, pl. 77, esp. 6, 11-12, 13-15; 16-18, of Pharnakes II, Asander, and an unidentified Bosphoran king, are closest to our period. For this comparison see also Rose 1990, 459.

36. As pointed out by Rose (1990, 459) with reference to the colossal polychrome tripod bearers catalogued by Schneider 1986, KO 1-3 (Naples, MN 6117, 6115; and Copenhagen 1177) and assigned by him to a monument on the Palatine. In these, however, the tongue flap is long, narrow, and straight, ending in a single point; it seems to be a double flap; the top piece is bent up over the shoe knot and tucked into the top of the shoe; the one underlying it is left dangling; see especially Schneider, pl. II. (He did not discuss shoes.) Other Parthian figures are KO 9 (Vatican Mag. delle Corazze 3330), a rectangular flap; KO 11 (Madrid, Prado 366), flap tapers to a point; SO 24, a long, narrow tongue; SO 53, an oval flap; pl. 37, unnumbered figure, a flared trapezoidal flap.

37. Cf. Hor. Carm. 3.21.20, regum apices burlesqued in this symposiastic mock hymn for Messala Corvinus.

38. They transform the genre of decorative Attis figures. As Rose (1990, 459) noted, Attis' Oriental dress can include shoes with flaps (as it can include the Oriental torque; see n. 34), but again none have the same profile: cf. LIMC III (1986), S.V., nos. 46, 117, 126 (rectangle), 90, 128 (rounded, straight tongue), 262 (narrow, pointed leaf). Only 115 (a small bronze from Thrace, Louvre Br 493) flares and has three points, where scallops are cut into the edge.

39. New York, Met. Mus., Edith Perry Chapman Fund 1949, 49.II.3, .64 m; Baltimore, Walters Art Gall. 54.1330, .62 m. Said to be from Egypt in 1912. See now Herrmann in The Gods Delight 1988, 282, 288-93, cat. 51, and 294-25, cat. 52. I cannot see that these figures wear, as Herrmann asserts, two tunics, a sleeveless one over a long-sleeved one; they seem rather to wear a single, very voluminous garment caught at the shoulder and other points, having but a single skirt. For its volume untucked, see the dexiosis reliefs from Commagene (n. 40). Note the arrangement of secondary folds under the belt, and the central knot of cloth over it: these indicate a date in the 30s, by comparison with Augustus' Apollo Palatinus, which had the same drapery mannerisms; cf. Roccos 1989.

40. Arsameia sculptures: Smith 1988a, 102-4, pls. 58.1-4, 59.3-5; dexiosis relief: pl. 59.1; see "Kommagene," Ant. Welt 6 (1975): fig. 78; cf. pl. 59.2, dexiosis relief from Sofraz Köy (Smith, 104), where the sun-god greeting Antiochos wears this mantle too. Depictions of Armenia and Asia (e.g., Herrmann in The Gods Delight 1988, 290, fig. 51c, Arretine "Asia" by L. Avillius Sura, priv. coll.) tend to have a centrally pinned mantle.

41. See Herrmann in The Gods Delight 1988, 290-91.

42. On philhellenism in the Commagenian dynasty, see Sullivan 1978, 793. On Antiochus IV's later attempts to try to Orientalize his image, see Smith 1988a, 104 (Nimrud Dagh as "a . . . synthetic Greek version of Oriental dynastic art").

43. See Herrmann in The Gods Delight 1988, 293.

44. Proposed by Rose 1990, 458-59.

45. Dio 54.24.4. Roddaz 1984, 463-75; RE 5.2 (1905) 1879-80, s.v. "Dynamis" (Stein), suggests that she may be the daughter of Pharnakes, who was sent in 47 B.C. to Caesar in token of peace (App. BCiv. 2.91)

46. RE, s.v., notes that Polemo lived to 8 B.C. at the latest and in the interval had time to marry Dynamis' sister Pythodoris.

47. The dynasty of Commagene: Sullivan 1975; Sullivan 1978, 781f. I have not seen his Near Eastern Royalty and Rome, 100-30 BC (Chicago, 1989).

48. Dio 59.4.3 calls him a paidiskos ( paidiskoi et' onti ). A paidiskos could be a young teenager; Polybius (30.26.9) says Ptolemy VII Philometor was a paidiskos when Antiochos IV attacked him in 170 B.C., and he is likely to have been at least twelve if not older, for he was a young child already in 180 B.C., when he acceded with his mother as regent.

49. Sullivan (1975, 35) accepts this date in his authoritative stemma of the house of Commagene.

50. Iotape, wife of Antiochos IV, in chignon and diadem, on the only coins of a queen of Commagene: Tasyürek 1975, 43, no. VII, fig. 47; her head at Nimrud Dagh: Smith 1988a, cat. 100, pl. 59.5.

51. Weinstock 1960, 53: "Why then should one wish to part company with so many distinguished scholars? They have not so far produced one valid proof and, indeed, they did not realize that proof was required. And yet neither is Pax represented on the altar nor her symbol, the caduceus, nor is her name inscribed on it. . . . Pax ought to have been conspicuous"; see PP. 45-46 in Weinstock on the etymology of pax and its two meanings for Romans in the Republic and early Empire.

52. Further on war celebrated here: Gruen 1985, 61-62.

53. Compare the arrangements of the positions of Gaius and Lucius on the altar, put on the north frieze with their mother, Julia, partly to avoid the awkwardness of including them on the south frieze, where both their natural and adoptive fathers are present; Rose 1990, 464-65.

54. Cf. Cic. QFr. 1.134: Rome's imperium guarantees pax sempiterna and otium to its subjects, protecting them from external and civil war, as long as they are loyal and pay taxes. Compare Augustus' work in Gaul, pp. 118, 121 below; Ehrenberg and Jones 1976, no. 42 (p. 76 above): imp. Caesari Augusto pp. Hispania ulterior Baetica quod beneficio eius et perpetua cura prouincia pacata est; also Cic. Prov. cons. 39; later, Tac. Hist. 4.74, Ann. 13.56.1, and the cruel parody by a victim of empire in Agr. 30 ("They make a desolation and call it pax ").

55. These correspondences to Venus/Amor imagery bestow approval upon the visual evocation of "barbarian" fecundity—again, not what one would expect in the representation of a people that has just left off being hostile. On the role of children on the Ara Pacis see now also Rose 1990, 453: "In addition to serving as illustrations of Augustan social policy and legislation, the children were used to signify the establishment and future maintenance of the Pax Augusta "; see also his conclusion.

56. Cf. Zanker 1988, s.v. figs. 19, 245a-b; La Rocca 1974; see chap. 2, P. 51, s.v. the Anaglypha Traiani; Kleiner 1985, 116-17; p. 223 n. 42 and p. 235 n. 63 (Venus and babies, late Republic and early Empire).

57. Drusus' presence in a paludamentum is part of the proof that the procession is not for an augury, nor for an animal sacrifice to Pax or any other deity, but rather for a supplicatio: see chap. 5, pp. 124-25, 136ff, on the limitation of the sacrifice paludatus to certain military oath-taking ceremonies. Drusus' costume also strongly implies what one would anyway expect, that the ceremony shown is taking place outside the pomerium in the Campus Martius near or at the Ara Pacis's site.

58. Simon 1967, pls. 16, 23 (detail of head); clearly the features and hair cap of Gaius. See p. 264 nn. 25, 27 (Pollini, Rose). As son of Augustus and camillus, Gaius is cognate with the camillus Iulus/Ascanius on the end panel serving Aeneas' sacrifice; see Rose 1990.

59. Correspondence noted most recently: Rose 1990, s.v. fig. 8. The correspondence between coins and cup is noted by Banti and Simonetti (after von Bahrfeldt) in CNR (1974), 142, there described as depicting the surrender of the German Sugambri in 8 B.C. Banti and Simonetti cite L. Cesano, who did recognize "son-giving" but identified the barbarian as the Parthian Phraates handing over one of his sons. Müthmann 1936; Gabelmann 1984, cat. 38, 121-22, pl. 12.6; Gabelmann 1986, 282; Trillmich in Kais. Aug. 1988, 523-24, cat. 366; Burns and Overbeck 1987, 48: the coin "undoubtedly" documents the Sugambri's surrender after 8 B.C. Omitted by the two most recent commentators on the cups, Hölscher (1980, 282) and Kleiner (1983, 291).

60. This image qua submission: Gabelmann 1986, 282-83

61. Rex datus types commemorate those occasions when Rome formally confirmed a client-ruler in his position or gave a new king to a foreign people. Cf RIC 419 (61 B.C.), the "crowning" of Ptolemy V of Egypt by M. Aemilius Lepidus in 201; BMCRE I, pl. 28.1 (A.D. 38), Germanicus crowning Artaxias of Armenia. Josephus describes such ceremonies in our period, stressing the gracious reception of the client-ruler: 40 B.C., Herod before the Senate with Antony and Caesar ( BJ 1.284-85, details of presentation to Senate); 30 B.C., Herod before Augustus in Rhodes ( BJ 1.386-93, details of crowning ceremony at 387 and 392-93); 4 B.C., Archelaus before Augustus ( BJ 2.25 and 37, details of supplication and of meeting before Augustus' council). A rex datus ceremony is probably shown on Gaius' cenotaph at Limyra (A.D. 4); see chap. 8, n. 66.

Thracian coinage of Rhoemetalkes III uses the formulae of Roman state iconography to narrate Rhoemetalkes' confirmation by Caligula: Rhoemetalkes stands before a tribunal on which Caligula sits, togate on a sella curulis . By the Caligulan period, then, if not before, this rex datus type (emperor on tribunal, candidate below) was set for the later Empire, when it was the normal formula. Gabelmann 1986, 282 n. 7 (" submissio ").

62. E.g., Burns and Overbeck's apposition of the two types (1987, 42, cat. 100-101). All identify the coin with the submission of German tribes conquered by Tiberius in his campaigns of 8 B.C. against the Germans in general and the Sugambri in particular, whose rulers went to Lugdunum to negotiate with Augustus, were taken hostage, confined in various cities, and killed themselves (Dio 55.6 supplies this detail); cf. RG 32. So CNR IV, no. 129 and pp. 131-32; Mattingly, BMCRE I, cxiii, cxvi; Gross 1985, 43; Hölscher 1980 and Kleiner 1983 (see n. 59); Simon 1986, s.v. fig. 90a. Gabelmann (1984) notes that the child seems freely offered, remarking the barbarian's dignified upright stance; since he mistakes this for a submissio, it becomes for him evidence that representations of "proskynesis" were deliberately avoided in Augustan submission scenes, to prove that the BR submission scene is too flattering to have been formulated in Augustus' lifetime (128-31). Contra see, among others, Price 1987, 85: acts of obeisance and cult on the part of foreigners and provincials are fine.

63. This fills the gap asserted by Blamberg 1976, 4: "The pacification of Spain and Gaul ( RG 26) received no direct numismatic advertisement." Rose (1990, n. 40) notes that the issue is pendant to a type with a young equestrian Gaius, signaling his visit to the Gallic legions at this time; he sees them as complements, showing Gallic children going to Rome and Roman children going to Gaul, this free movement of children being itself a sign of pax .

64. Gros 1981, 160-65, fig. 1; Rolland 1977, pl. 24 (drawing), 50f. The arch: Bedon et al. 1988, 1: 178-80, 2: 116 (bibl.); Rolland 1977.

65. Cf. the mix of Italian tomb and Celtic rite by a Gaul loyal in opposition to Civilis' rebellion; p. 229 n. 10 above. Strabo 3.3.6: good Spaniards, loyal and Romanized, are " togati (or as you might say, peacably inclined), and have been transferred, clad in their toga robe, to their present gentleness of disposition and their Italian mode of life" (Knapp 1977, 160); Dio 46.55: Narbonensis is called "togata" since, being more pacified ( eirenikotera = "pacatiora''), its cities use Roman dress. A proud Rhaetian auxiliary under Augustus or Tiberius had his stele carved to show a statue of himself in military dress on a pedestal, next to it his son wearing a Roman toga; stele of Montanus from Andernach (Bonn); CIL XIII.7684; Espérandieu 1907, cat. 6207.

66. Rolland 1977, pl. 25 (drawing), 47f. Most call her Roma, but her fringed sagum and long gown make her Gallia. Because she sits upright and at ease atop (instead of among) a weapon pile—like Roma—she is not devicta but triumphans, cognate with the good Gaul opposite.

67. Stemmer 1978, 19, cat. I.19, pl. 10.1. Compare the Gemma Augustea exergue; see fig. 16.

68. See chap. 3 on the equation in status made in the province group of BR I:1, between Gaul and Spain on the one hand and "civilized" Africa and Asia Minor on the other.

69. Aymard 1961, 136-42; at p. 141 distinguished from royal offspring "freely" come to Rome.

70. Phraates' sons: Augustus RG 32.2; arena story: Suet. Aug. 43.4; cf. ILS 842 = Ehrenberg and Jones 1976, no. 183, a dedication by two of them. Cleopatra Selene: Dio 51.15; Plut. Ant. 87; Crinagoras (Gow and Page) 25, 28. Herod's sons—Agrippa: Joseph. AJ 18.143; Archelaus and Philippos, paideuomenoi in Rome: Joseph. BJ 1.602. Such children of client-kings, listed also by Kienast 1982, 407-10, are distinguished by Braund (1984, 16) from massed child hostages (p. 270 n. 76 below).

71. Cf. Juvenal's cruel parody in 2.162f., on the corruption by actual Roman upper-class mores of a young Armenian hostage, who had become the lover of an officer: "venerat obses,/hic fiunt homines. . . . mittentur braccae cultelli frena flagellum,/sic praetextatos referunt Artaxata mores" (165-69). Compare Caligula's seduction of hostages described in Suet. Gaius 36; Courtney 1980, 149-50. Courtney takes referunt to mean "bring back," as if the mores were spoils of war; it refers rather to the process of Romanization and transfer of Roman ways back home, expected of such hostages.

72. Episodes from the Social Wars of 90 B.C. demonstrate that the practice of sending highborn children as hostages and tokens of fides had become embedded in Romano-Italian political thinking. Adolescents were exchanged as obsides by the members of the Italian League as part of their initial moves toward covert confederation, namely, the meirakios homeros of App. BCiv. 1.38; Appian also narrates ( BCiv. 1.44) how the rebel general Popaedius pretended to ally himself with the Roman general Caepio by sending him his two infant children (actually slaves) dressed like Roman children of rank in the toga praetexta . To send one's own child (or pretend to!) as a token of pax occurs again in the context of civil war: Antony, in the tense period succeeding Caesar's assassination, sent his little son (who was about two years old) to the Senate as a pacis obses; Cic. Phil. 1.13.31, 2.26; Plut. Ant. 14.1-2, cf. Brut. 19.2; App. BCiv. 2.142. As Pelling (1988, 151) remarks, they "doubtless found their hostage a handful." Antyllus was sent again to Octavian after Actium (Dio 51.8)—not, contra Pelling (297), "implausible."

73. In another episode from Appian's account of the Social Wars Of 90 B.C., BCiv. 1.42, a rebel general makes opportune use of such a hostage to create a " rex " to seduce the loyalty of the prince's countrymen serving his Roman opponent as auxiliaries. Gaius Papius took Oxynta, son of Jugurtha, away from his Roman guard at Venusia, put him in the royal purple, and displayed him to the Numidian horse of Sextus Caesar, most of whom promptly went over "as if to their own king."

74. Livy 2.32.9-12, the apocryphal speech of Menenius Agrippa after the secession of the plebs, made to induce them to return to their place under the guidance of the Senate. Cicero ( Off. 3.5.22) uses the metaphor to demonstrate that natural law demands harmony rather than violence from the members of society toward each other. See ultimately Xen. Mem. 2.3.18: Socrates compares the fighting of friends or brothers to a pair of hands or feet whose twin members fight one another. Note how the Augustan Livy twists this model of concord into a model of hierarchy.

75. Cf. Knapp 1977, 108, 111f., and pt. 3, chap. 1, on Romanization and its mechanisms in Spain before 100 B.C.

76. Sertorius' institution of special schools for the children of native elites is an interesting tool to direct Romanization. Augustus and Tiberius were to maintain this practice in Gaul; see their school for young Gallic noblemen at Augustodunum (Tac. Ann. 3.43) and the ludus litterarius for the education of obsides near the Rhine, whose students were paraded by Gaius (Suet. Calig. 15). (Related, in effect if not purpose, according to Braund 1984, 14, was the characteristic stipulation in late Republican and Augustan treaties that hostages should be replaced with a fresh batch at fixed intervals of a few years' duration, thus exposing as broad a cross section of the elite as possible to Roman mores .) By the Flavian era the institution of such schools was a recognized tool of imperial expansion and the creation of new provinces: Tacitus ( Agr. 21.2) portrays Agricola's creation of such schools in Britain as part of the typical repertoire of devious imperialist tools for subjugation. For Kienast (1982, 403 n. 173a) the school foundations in Gaul and Britain were meant to combat Druidism, as counterweights to Druidic schools.

77. See n. 78. Strabo 4.3.1; Livy Per. 137; Dio 54.32.1; Suet. Claud. 2.1; Juvenal 1.43; Turcan 1982; Hänlein-Schäfer 1985, 14 and cat. A64; Zanker 1988, s.v. fig. 236. See Knapp 1977, 160: Republican generals in Spain tended to "lead natives away from their fractionalized political organizations" by encouraging tribal confederations and the coalescing of smaller into larger settlements.

78. Habicht (1972, 65-67) discusses the Western comparanda to the Greek East (55-64); organized imperial cult, like the provincial assemblies of the West, was not based on native tradition nor freely invented, but rather was inspired (66) and even directly overseen by the members of the imperial house (Drusus and the Lyons altar; 2 B.C., L. Domitius Ahenobarbus and the Elbe altar, Dio 55.109.2; A.D. 9, Varus and the Ara Ubiorum for Germany). Thus the imperial house established a dependence on themselves by the native elite, and a patron's responsibility toward such elites on their own part. According to Hänlein-Schäfer 1985, 14 n. 32, the imposition of such assemblies was an Augustan innovation, which was carried out in those areas most recently acquired or least integrated into the Empire; compare the regions listed by Augustus as made to take a special oath to him ( RG 25; 32 B.C.: Gaul, Spain, Africa, Sicily, Sardinia); Hänlein-Schäfer's cat. A65, the Ara Ubiorum. Cf. Zanker 1988, chap. 8. Replication the other way, from West to East: Oliver (1958, 475, 494-96) suggested that gerusiae in the East were encouraged as a sort of counterpart to western Augustales, formulated to strengthen the cities' loyalty to the Roman government; cited by Sherk with regard to Agrippa's letter to the Gerousia Argiva (1969, 324, cat. 63).

79. Though Braund (1984, 16) records the demanding of child hostages by Augustus, he sees it as part of the phenomenon of the mass taking of hostages, initiated by Caesar in Gaul (600 obsides: BGall. 2.15; 200: 5.4; 100: 6.4); this was a Roman response to the typical west-European societal structure of tribal elites with their widely available chieftainships, in perceived contrast to Hellenized political structures. Augustus took at one time 700 child hostages in newly conquered Dalmatia (App. Ill. 28). The BR event may indeed have been closer in inspiration to this phenomenon than to other distinct traditions of royal "son-giving"; what matters for us is that Augustan propaganda distinguished the BR event from an event like that in Dalmatia.

80. Its disruptive effects were aggravated by German subversion. Livy Per. 139; Dio 54.32.1; ILS 212.11.36-39. The loyalty of Gallia Comata and the Aedui in the face of this discontent earned them Claudius' commendation, discussed on pp. 120-21 below. The problem with the census was that it was a tool for tax assessment; Tacitus ( Ann. 3.40ff.) depicts resentment over taxes and tax debts as the spark kindling the revolt of Sacrovir in A.D. 21 following Germanicus' death (in which revolt, as Claudius diplomatically omitted to say, many Aedui took part).

81. Maurin 1986, 110f., noting that this is the only provincial assembly to cross province lines; Gallia Comata is the official term for the tres Galliae up to ca. the reign of Vespasian.

82. Census and altar cult were combined again as mechanisms for the pacification of the new province of Germany around A.D. 9; Tac. Ann. 1.39, 57. Cf. the combination of an altar ( bomos ) to Augustus with treaty alliances ( philia ) in regard to the trans-Elbe tribes immediately before 2 B.C.; Dio 55.10a.2. Cf. Vell. Pat. 2.107: a German crosses the Elbe to get a look at Tiberius, whom he knows so far from cult worship. See also pp. 270-71 n. 78.

83. The census: RE 3.2 (1899) 2703-19, at 2708-10, s.v. "Claudius" #139 (E. Groag) (Drusus). Drusus' mandate in Gaul: Rösger and Will 1985, 37f. The council of the primores Galliarum: Bellen 1984, 386-87, noting Drusus' special position as patron to the Gallic primores . The assembly was first organized in 12 (Dio 54.32.1), that is, for the initial consecration of the altar complex at Lugdunum prior to construction. It will have met again in 10 B.C. for the dedication of the completed complex, the award of the first priesthoods, etc. Bellen primarily examines the obsequies for Drusus voted and enacted by this assembly (in conjunction with the Rhine legions), on which see further Frenz 1985a, 395-97; Frenz satisfactorily identifies Drusus' cenotaph with the so-called Elchelstein at Mainz.

84. For details of his encounters with individual German peoples see Will 1987; he notes (46) how contemporary opinion at Rome treated Drusus' victories in battle as meaning successful, permanent incorporation of the territory traversed; cf. Livy Per. 140.

85. Just so Tiberius worked to consolidate Gallic affairs before moving against Germany itself when he later was sent to repair the effects of Varus' fall; cf. Vell. Pat. 2.120.1: "Gallias confirmat."

86. These connections are stated generally by Halfmann (1986, 27) in relation to the role of Augustus' "Stellvertreter" Drusus, Tiberius, and Germanicus. He notes Drusus' and Germanicus' census; Claudius' speech on the Aedui; a circuit by Tiberius of Gaul in A.D. 4 when he resumed his Rhine command; an ancient formulation of the role played by these imperatores, i.e., the commendation of Germanicus by the Senate at his death as "ordinato statu Galliarum" (Tab. Siar.; González 1984, 59, l. 15).

87. Dio 54.36.2 says that when the Dalmatians and Dacians attacked in 10 B.C., Tiberius was sent against them from Gaul, where he had gone with Augustus; 54.36.4 has Augustus start for Rome with Drusus and (cf. Suet. Tib. 7.3) Tiberius, the emperor having been in Lugdunum most of the time watching the Germans from near at hand. POxy. 3020 (Sherk 1969, no. 100) refers to an embassy from Alexandria that had to go to Gaul in 10 to present its case to Augustus. CIL VI.457 (Chisholm E.16) and ILS 92 (Chisholm J.1e) are altar dedications made by the emperor back in Rome from gifts given to him during his absence in 10.

See now Halfmann 1986, 158-62, for Augustus' visit in the winter 11 B.C. through 10 B.C. One hopes that historical literature in English, especially basic texts, will soon begin to supply more accurate chronologies and will cease to maintain the indifference of so many ancient sources to Augustus' physical and administrative interventions in the Western provinces. As an instructor of undergraduates, I note with disfavor that the otherwise laudable 1987 Penguin translation of Dio's Augustan book supplies in the category "The West" in its "Chronological Table" on pp. 304-5 no mention of any of Augustus' journeys nor of Drusus' census nor of the institution of the Lugdunum altar cult; the index on p. 335 notes only the first visit to Gaul, in 16-13 B.C.

88. Halfmann (1986, 159, 162) posits a last visit of Augustus to Gaul in 8 B.C.: Dio 55.6.1, when Tiberius crossed the Rhine Augustus stayed in the oikeia, the "hinterland"; Augustus returned to Rome in 7 B.C. an event documented by ludi votivi given for his return ( CIL VI.4, fasc. 3, no. 36789 and VI.1, no. 385; cf. Dio 55.8.3). However, oikeia can indicate North Italy as well as Gaul.

89. These will be identical with the families, and even with particular individuals, supplying the summus magistratus who led Gallic civitates through the early Tiberian period, when a less personal system of magisterial rule by councils of duoviri was instituted; see Will 1987, 12.

90. Dio 46.55: ekaleito . . . haute de de Komata hoti hoi Galatai hoi tautei es komen to pleiston tas trichas anientes episemoi kata touto para tous allous esan .

91. Neither physiognomy nor dress could be those of Orientals, contra Simon 1986, 143.

92. Brussels, Mus. Roy. inv. A 1145. Hölscher 1984b, figs. 1-4, 284 n. 44, attributing the head to a victory monument of Marius. An actual North European with such hair is the "Osterby Man" from a Danish bog: Glob 1969, the fifth plate after p. 112. Cf. a bronze of the second century A.D. in Romans and Barbarians 1972, cat. 4. On a stele from Worms, this chignon seems to have fallen down on a naked German being speared by an auxiliary—a long tail of hair falls down his back; Espérandieu 1922, 92-93, no. 6014; CIL XIII.6233.

93. RRC 448/2e, pl. 53.8; rev., naked Gaul fighting from a chariot. The companion issue 448/3, pl. 53.9, has a female bust with a long, disheveled hair (rev., Artemis with spear and stag). The shield behind the male bust and the carnyx (Gallic war trumpet) behind the female bust emphasize that these are trophy images. Compare the reduced but still vividly observed head of the captive by a trophy on a companion issue, RRC 452/4 (48/47 B.C.).

94. Trans. Chisholm and Ferguson 1981, K.23a; Braund 1985, no. 570.

95. Cf. Prop. 6, with reference to the Sugambri in l.77, celebrating Agrippa's holding ludi quinquinnales while Augustus was away in 16, as described in Dio 54.19.1, 8; on Lollius' campaigns, see RE 2, 4.1 (1931) 660, s.v. "Sugambri" (Schönfeld).

96. ILS 916 = Ehrenberg and Jones 1976, no. 198 (Forum Clodii) on Cn. Pullius Pollio, who had prior/further experience of the West as proconsul of Gallia Narbonensis. Maurin 1986, 110. Sources for Augustus' activities of 16-13 B.C.: Halfmann 1986, 27, 158-62. He left Rome after 29 June, when he dedicated the Temple of Quirinus (Dio 54.19.4); CIL VI.385 documents the ludi votivi pro reditu of 13 B.C.

97. Tiberius' own and inherited popularity in Gallia Narbonensis: Bedon et al. 1988, 1: 180.

98. On some principes of Gallia Comata, see Syme 1977/1987, 992-94. He notes the many "xx Claudii" who will have received citizenship from Tiberius and Drusus, and the Aeduan Julius Calenus (grandson of C. Julius Eporedorix), one of only two Gauls known to have been tribunes in a Roman legion. ILS 7008 (Aventicum) records a public funeral for the Aeduan C. Valerius C.f. Fabius Camillus, not, contra Syme, for the Julia C. Iuli Camilli f. Festilla (probably his sister) who funded it.

99. Habicht 1972, 68: his citizenship will have been granted by Caesar or Augustus to bind the loyalty of this near descendant of one of Caesar's worst enemies; similarly, at the cult of A.D. 9 at Ara Ubiorum one of the first priests is Segestes' son C. Iulius Segimundus (Tac. Ann. 1.57; cf. Strabo 7.291f.). Another early Aeduan priest: ILS 7014, . . . / Aed[uo]/sacerdos; Maurin 1986, 115. A later Augustan/Tiberian priest (who dedicated the amphitheater at Lugdunum), C. Julius Rufus, carefully boasts three generations of Gallic ancestry on an arch to Tiberius at Saintes; CIL XIII. 1036; Bedon et al. 1988, 1: 197; his fellow tribesmen of the Santoni, C. Julius Victor, proclaims himself Conconnetodubni f: CIL XIII.1042. Cf. Syme 1977/1987, 992-93, on these noble pedigrees.

100. This does not vitiate my comments on historicity. I firmly maintain the accepted view, that the procession friezes of the Ara Pacis depict the celebration of a particular ceremony (its nature is a different problem) on a particular occasion. It is evident that the friezes are also meant, like the Parthenon frieze, to function as a timeless commemoration of the strength of the Augustan settlement based on the primacy of Augustus' domus . To this end persons are included in the procession whose simultaneous presences are difficult to account for, such as Agrippa and his "hostage," who should have returned only in 12 B.C., Drusus, etc. Symbolic details do seem to have been added or reworked in the course of time, between 13 and 9 B.C.; the fact that the veiled Agrippa, alone on the frieze, does not wear a laurel wreath must somehow refer to his death in 12. (This feature would have been highlighted by the original painting and gilding of the altar, and by Agrippa's placement between two veiled figures who did have wreaths, Augustus and Livia.)

101. On the manipulation of tribal structures as structures for the enrollment of auxiliaries, and the role of militarization in assimilation, see Will on the German tribes in the period from Drusus' campaigns onward (1987, 17f.). Will notes (7) how the Batavii, for example, served in groups led by the nobilissimi popularium (Tac. Hist. 4.12.3); he also comments (46) on how a tribe would be assimilated as a gens foederata partly by citizenship grants to its leaders (here, Segestes of the Cherusci, grant from Augustus; Tac. Ann. 1.58.1).

102. Cf. Kienast 1982, 119-20, on Augustus' propaganda image as pater orbis .