¹ See R. Seager, CQ 47 (1997), 253fF. As in that article, my chief concern is with Ammianus' perception of events rather than what actually happened, except in so far as that sheds light on the historian's distortions and omissions.

² Ammianus states clearly that the treason of the German officers was no more than a rumour (7: ut quidam existimabanl). It is argued by P. A. Barceló, Roms auswartige Beziehungen unter der Constantinischen Dynastie (306–363) (Regensburg, 1981), 25, 31–2, that they had in fact been sent to negotiate a peaceful settlement with their kinsmen. This is possible, but not compelling.

³ For surprise and its importance, see e.g. 20.10.2, 21.3.8, 29.4.2, 30.5.14, 31.10.12, 31.11.5.

⁴ For fear stimulating barbarians to make peace, see e.g. 17.1.12, 17.12.13, 17.13.2–3 (feigned), 21, 19.11.6, 15,27.5.3,29.6.16,31.12.12.

⁵ For self-abasement, see e.g. 16.12.15, 17.1.12, 17.8.5, 17.10.3, 9, 17.12.9–10, 13, 17.13.21, 19.11.6, 20.10.2, 27.5.7, 30.5.1, 30.6.1–2, 31.4.1, 4, 13, 31.10.17.

⁶ The task Constantius sets himself here is in essence similar to that performed in 369 for Valens by Themistius (Or. 10), though Constantius is more pragmatic, whereas Themistius descants on his favourite theme of philanthropia; see below, section 6.

⁷ Barceló (n. 2), 26.

⁸ Barceló (n. 2), 26.

⁹ On this practice, see Heather, P. in Pohl, W., Kingdoms of the Empire (Leiden, 1997), 69.Google Scholar

¹⁰ The text is defective, but the general sense is clear enough.

¹¹ Bitter, N., Kampfschilderungen bei Ammianus Marcellinus (Bonn, 1976), 59ff.Google Scholar, 76, 84; Seager, R., Ammianus Marcellinus (Columbia, 1986), 33ff.Google Scholar, 54ff.; Wiedemann, T. E. J. in Moxon, I. S., Smart, J. D., and Woodman, A. J., Past Perspectives (Cambridge, 1986), 194ff.Google Scholar

¹² For fear at the approach or first sight of Roman arms, see e.g. 16.11.8, 17.1.3, 6–7, 17.2.1, 17.10.3, 17.12.4,21, 17.13.3, 18.2.14, 27.2.6, 30.5.13, 31.7.7.

¹³ There seems to be no good reason to assume (despite Barcelo [n. 2], 28) that Constantius is being criticized here for avoiding the action and then claiming credit for others' successes. Contrast 16.12.70 on Julian in Gaul.

¹⁴ Barceló (n. 2), 62ff.

¹⁵ It is probably too subtle to interpret laetioris as ‘happier than usual’, implying that, although the campaign was a success, this was an exception to the rule. (Thus Sabbah, G., La méthode d'Ammien Marcellin [Paris, 1978], 188Google Scholar, n. 60. Contra: J. Szidat, Historia 21 [1972], 713, though of the two parallels he cites only 26.7.15 is apposite; 20.11.32 is a genuine comparative.)

¹⁶ The allusion is presumably to Constantius' generous treatment of the Alamannic kings in 354, but it is not certain that the comparative has pejorative overtones (as implied by Sabbah's translation; contra: De Jonge ad loc). Contrast the unquestionably hostile judgement of Julian (Ep. Ath. 280B).

¹⁷ Barceló (n. 2), 63.

¹⁸ Most editors read pauor ‘fear’ (E) for V's nonsensical patior, but the correction cannot be regarded as certain; see De Jonge ad loc.

¹⁹ On deditio at this time, see the various views of Schulz, R., Die Entwicklung des römischen Völkerrechts im vierten und fünften Jahrhundert n.Chr. (Stuttgart, 1993), 136ff.Google Scholar; G. Wirth in Pohl (n. 9), 16ff.; Heather (n. 9), 61ff. (the best treatment); E. Chrysos in Pohl (n. 9), 192ff.

²⁰ On Constantius' diplomatic policy of dismantling larger units in the interests of easier control and manipulation, see Barceló (n. 2), 64, 71.

²¹ The details of their liberation and the background of their subjection to the Limigantes are provided in 17ff., all in a manner complimentary to Constantius' power.

²² The Sarmatae had good cause; see Barceló (n. 2), 65.

²³ Ammianus treats this campaign in virtual isolation from what has gone before, alluding only indirectly (17.12.19) to the fact that Constantius intended to use Sarmatian troops against the Limigantes. See Barceló (n. 2), 65.

²⁴ This must imply that the Limigantes were already aware of the Roman plan to move them elsewhere. Hence their initial fear of a Roman attack must have been based on a lack of confidence in Roman good faith.

²⁵ Described at length in 15.4.

²⁶ Contrast the promise of obedience in 15.3.

²⁷ Lenius here echoes lenius in 2. Since at that point there was no suggestion of bad faith on Constantius' part, the repetition may be meant to indicate that he was still sincere.

²⁸ Noted by Barceló (n. 2), 68, in the course of an account of events which is extremely hostile to Constantius and the Romans.

²⁹ It is possible that his attitude was in part at least dictated by the need to conciliate the free Sarmatians.

³⁰ Thus Barceló (n. 2), 66ff.

³¹ Cf. 8: oculis … et uocibus truculenlis; 9:fuwris amentiam; acriter imminentes.

³² Specific contrasts: 11:nec … exorauit celerem mortem: 3: uilam precati; 11: alienis uiribus potiusquam conscientiae suae iudieio: 2: scelerum conscientia; 11: fortunae non meriti: 2: lenius… poscebat.

³³ Fastu prioris uitae here echoes cum genuine fastu (5), suggesting that their humiliation was well deserved.

³⁴ For the obedience of the elements to commanders blessed with felicitas, cf. Pan. Lai. 10.12.4; 11.9.1ff.; 8.7.3, 14.4–5, 15.1; 7.9.4; 6.13.3; 5.10.4; 4.32.6. Note also the Tiber's active support of Constantine against Maxentius (12.17.2, 18.1) and that of the Save for Theodosius against Maximus (2.34.4). See R. Seager, PLLS 4 (1983), 132, 135, 138, 144, 146, 148, 151, 155, 161.

³⁵ In 23 parere imperils underlines the difference in their attitude from that displayed by the Amicenses in 5: non iussa facturi; ut iubencla repudiarent.

³⁶ See Seager (n. 34), 145, 147, 149ff., 153f., 160ff.

³⁷ Sabbah (n. 15), 188, n. 60, sees alacrior solito as a further ironic deflation of Constantius' achievement. But see Szidat (n. 15), 713.

³⁸ They were still on their own side of the river, cf. De Jonge ad loc. and the translations of Rolfe and Seyfarth, against that of Hamilton.

³⁹ Genuino more here perhaps picks up the allusion to their native savagery in the advance notice at 17.13.23.

⁴⁰ This on the assumption that V's adtributis is retained. The paradosis is more pointed than the correction attributa, since it makes the Romans claim that the Limigantes had requested not merely a treaty but a new home, even though that claim may be false, since it goes beyond 17.13.23, where Ammianus says only sunt assensi ‘they agreed’.

⁴¹ Thus De Jonge ad loc; contra: Sabbah in his translation.

⁴² Barceló (n. 2), 68–9, notes other cogent motives for Constantius' moderation: the need to deal with Persia as soon as possible, with a minimum of casualties and the possibility of fresh recruits.

⁴³ Against any drastic undermining here of the official report (as suggested by Sabbah [n. 15], 189–90), see R. Seager, PLILS 9 (1996), 194–5

⁴⁴ That the Limigantes were not merely deceitful but in breach of their treaty is recalled by the label rebelles applied to them in 15.

⁴⁵ Noted, though not explained, by Szidat (n. 15), 719. For a succinct but sound estimate of Constantius' strategy on the northern frontiers, see Barceló (n. 2), 70.

⁴⁶ This justifies both punitive expeditions and pre-emptive strikes across the Rhine, which should not therefore be seen as instances of Julian exceeding his brief or as a new strategic departure affer Strasburg. (Thus Barceló [n. 2], 35–6,40.) This view can be traced back to Lib. Or. 18.68. See Crump, G. A., Ammianus Marcellinus as a Military Historian (Wiesbaden, 1975), 64–5.Google Scholar For barbarian arrogance, see e.g. 15.8.7, 16.4.16ff., 16.12.2ff., 17.13.5, 8–9, 31.7.2–3, 9, 31.8.4, 9, 31.9.1, 31.12.6, 31.15.2ff.

⁴⁷ Cf. 15.8.13: tutelam … Galliarum; 28: tutelaque peruigili Galliarum; 21.13.13: tuendis praefecimus Galliis. A similar conception of Julian's mission is to be found in Lib. Or. 12.44,48ff., 18.36. When writing what Constantius might want to hear, Julian himself espoused comparable views (Or. 1.7C, 9D); see also Ep. Ath. 279A-D.

⁴⁸ On Ammianus' exaggeration of Julian's role in this year, see Barceló (n. 2), 34; Crump (n. 46), 17. That Julian was at first only a figurehead was well known to Libanius (Or. 18.42–3), Eunapius (fr. 16.2), and of course Julian himself (Ep. Ath. 277D-8A).

⁴⁹ See Lib. Or. 18.47. For the likelihood that Ammianus exaggerated the significance of the Frankish occupation of Cologne and for the possibly undistinguished role of Ursicinus in these events, see Drinkwater, J. F., Francia 24/'1 (1997), 2ff.Google Scholar

⁵⁰ The narrative pays as little attention as possible to Constantius' readiness to listen to and act upon Julian's complaints about Marcellus. See Seager, R. in Hillard, T. W., Kearsley, R. A., Nixon, C. E. V., and Nobbs, A. M., Ancient History in a Modern University, vol. II (Grand Rapids, 1998), 282.Google Scholar

⁵¹ On this campaign, the divergent accounts of Ammianus and Libanius, and Julian's feud with Barbatio, see Barceló (n. 2), 34–5; Crump (n. 46), 18ff.; Sabbah (n. 15), 266–7, 466ff.

⁵² That this Alamannic coalition was a response to Julian's aggression is plausibly argued by Barceló (n. 2), 35. Lib. Or. 18.52 presents a very different version of the German ultimatum, in which they claim to have settled on lands granted by Rome and produce letters from Constantius to prove it. (In Or. 13.24, however, his language had been much closer to that of Ammianus.) The two versions are not necessarily incompatible. The Germans might have claimed that they had won the lands by force of arms and that Constantius, unable to expel them, had accepted the fait accompli and granted them title. Ammianus would then be emphasizing the theme of barbarian presumption, Libanius that of Constantius' weakness and inertia.

⁵³ For similar vocabulary, see Julian in his speech before Strasburg (31) and Ammianus' retrospective description of the envoys' message (17.1.1).

⁵⁴ Bitter (n. ll), 96ff.

⁵⁵ Ammianus is obviously concerned to ensure that the blame for such a risky engagement is distributed as widely as possible, while leaving Julian in the clear. Thus Sabbah (n. 15), 170, though he perhaps goes too far in claiming that the battle was fought in defiance of Constantius' orders. There is a parallel with Julian's supposed innocence and passivity at the time of his usurpation, though there Ammianus makes it clear between the lines that he is well aware of the truth. (For references and brief discussion, see Seager [n. 11], 114–15; id. [n. 50], 282–3.)

⁵⁶ That these dissensions were fostered, if not actually instigated, by Constantius is highly likely; see Barceló (n. 2), 30.

⁵⁷ Crump (n. 46), 18.

⁵⁸ See 31 on the long-awaited day. For Julian's aggressiveness from 357 on, see Drinkwater (n. 49), 8.

⁵⁹ For the theme of restoring honour, see 32.

⁶⁰ Lib. Or. 18.69.

⁶¹ Barceló (n.2), 37.

⁶² See De Jonge, Sabbah ad loc. For Ammianus' minimisation of the threat and Barbatio's achievement, see Barceló (n. 2), 60ff.

⁶³ For Julian's own views on a similar situation, see Or. 1.7C. But he claims a further motive, which Ammianus omits: the Salii were refusing entry to ships from Britain unless the Romans paid a toll (Ep. Ath. 280B-D). Ammianus also plays down Julian's aggressive arrogance (as does Zos. 3.6).

⁶⁴ See above, n. 35.

⁶⁵ On these campaigns, see Barceló (n. 2), 38–9.

⁶⁶ See Lib. Or. 13.30–1, 18.75ff.

⁶⁷ The text is uncertain, but the general sense fairly clear.

⁶⁸ On clauses concerning prisoners, supplies, and assistance in rebuilding, see Schulz (n. 19), 29–30.

⁶⁹ See in general Lib. Or. 18.81.

⁷⁰ For other such assassination attempts against barbarian leaders, see 27.10.3–4 (Vithicabius), 29.6.5 (Gabinius), 31.5.4R”. (Fritigern and Alavivus).

⁷¹ 18.2.19: tamquam ipsi haec deliquissent in nostros. Mommsen added non before ipsi. But this change produces a somewhat illogical train of thought, and does not exculpate the Romans, since deliquissent would still refer to their conduct as well as that of the Germans. (See De Jonge ad loc.)

⁷² It is unclear why Barceló (n. 2), 45, thinks they were the innocent victims of Julian's unprovoked aggression in a campaign undertaken solely to boost the morale of his troops. There is no obvious reason why the derelictions ascribed to them should not have been genuine.

⁷³ V has el dignum. But Mueller's emendation of et to si must surely be adopted to save Ammianus from contradicting himself in the space of four lines. Other sources are less coy about Constantius' guilt: see Lib. Or. 13.35, 18.107, Eunap. fr. 14.1, Jul. Ep. Ath. 286A, 287C, Mamertinus Pan. Lat. 3.6.1.

⁷⁴ Barceló (n. 2), 46–7, is highly critical of Julian, on the grounds that the raids were directed against Rhaetia and were therefore not Julian's concern, and that Vadomarius himself probably knew nothing about them, since he is hardly likely to have invaded a region over which his patron Constantius held sway. However, Ammianus does not say that the target was Rhaetia, but regions bordering on Rhaetia (21.3.1), so this argument fails. That Julian was eager to seize on any pretext to lay hands on Vadomarius cannot be doubted, but equally he can hardly be blamed, once he had decided on civil war, for not wanting to leave at large in his rear such an enthusiastic supporter of Constantius. That he (and Ammianus) should then present this fundamentally selfish motive in terms of the welfare of the provinces is only to be expected.

⁷⁵ As noted by Barceló (n. 2), 48.

⁷⁶ On giffs to barbarians, see Heather, P., Goths and Romans 332–489 (Oxford, 1991), 114Google Scholar; Heather (n. 9), 70; Heather, P. and Matthews, J., The Goths in the Fourth Century (Liverpool, 1991), 23–4Google Scholar; Schulz (n. 19), 38–9. For possible exaggeration of the Alamannic threat, see Drinkwater (n.49), 11.

⁷⁷ For the campaign of 27.10, see Seager (n. 43), 191ff. The treatment there is defective in that it fails to note the close relationship between 27.10 and 27.2, on which attention is therefore concentrated here. Sabbah (n. 15), 205–6, comments on Ammianus' favourable attitude to Jovinus, but makes no connection with 27.10.

⁷⁸ Some specific verbal parallels between 2.1 and 10.6ff. deserve notice. Careful preparation: 2.1: 10.6. Cautious progress: 2.1: 10.6; 2.2: 10.7. Reliable scouts: 2.2: 10.8.

⁷⁹ All this contrasts sharply with Jovinus' complete control of the situation at 2.5–6.

⁸⁰ Again contrast Jovinus' freedom from such dangers at 2.8.

⁸¹ See Crump (n. 46), 114ff. He suggests (124) that Valentinian's aims were not solely defensive, but ‘future Roman punitive expeditions’ (125) are a part of defensive strategy.

⁸² Ammianus' account is probably derived from Syagrius' report, see Sabbah (n. 15), 157ff.

⁸³ V offers forticen regendus uerius milite barbaris quam pellendis. Heraeus' foret in for forticen should be accepted. The minimum change that would then produce acceptable grammar is regendis for regendus, but ‘controlling the barbarians with troops rather than expelling them’ makes little sense. The general sense is clear enough: something was better than driving the barbarians off. Common sense suggests that this was keeping them out. (Compare the repeated emphasis on Valentinian's work in strengthening the frontier defences.) The most attractive reading is therefore in coercendis uerius limite (Clark), i.e. ‘in restricting them by a frontier’, but it is hard to justify elevating any conjecture to the text.

⁸⁴ For Ammianus' somewhat complex and very cautiously expressed attitude to the elder Theodosius, see Seager, R., Histos 1 (1997)Google Scholar forthcoming (already available in the Internet version).

⁸⁵ It is clear from Ammianus that Macrianus remained on shore. This is correctly understood by Schulz (n. 19), 46; contra: Heather (n. 76), 120; Heather and Matthews (n. 76), 26.

⁸⁶ Crump (n. 46), 126–7.

⁸⁷ Seager(n. 11), 68.

⁸⁸ The further brief notice at 30.3.2 presents these events in conventional fashion as just another typical frontier violation with overtones of barbarian presumption, with no mention of the extenuating circumstances.

⁸⁹ Drinkwater (n. 49) is too simplistic. His own assessment of Valentinian's achievement (13) is very hostile.

⁹⁰ On the treaty of 332 and subsequent relations between the Theruingi and Rome, see E. A. Thompson, Hermes 84 (1956), 377ff.; Heather (n. 76), 108ff., 116–17; Schulz (n. 19), 32, 51.

⁹¹ For instance, of the scholars cited in n. 90, Heather accepts a clause permitting free trade, inferred from Them. Or. 10.135C-D, but this is rejected by both Thompson and Schulz, though their alternative explanations are quite different from one another.

⁹² Themistius had already prepared the ground in 368. Or. 8.114C implies unmistakably that although the Goths were behaving in an arrogant manner, they were going to be let off lightly. See Heather (n. 76), 117; Heather and Matthews (n. 76), 14ff.

⁹³ For discussion, see Sabbah (n. 15), 352ff.; Heather (n. 76), 115ff.; R. C. Blockley, East Roman Foreign Policy (Leeds, 1992), 32–3; Schulz (n. 19), 50ff

⁹⁴ Heather (n. 76), 120, suggests that Athanaricus' oath may have been invented to save Valens' face.

⁹⁵ The implication of both Amm. 27.5.9 and Them. Or. 10.133C is that Valens and Athanaricus were on separate ships. (Thus Schulz [n. 19], 46; contra: Heather [n. 76], 118; Heather and Matthews [n. 76], 25; Blockley [n. 93], 33.) This is only to be expected; neither security nor dignity would have allowed either ruler to set foot on the other's vessel.

⁹⁶ See above, section 1 and n. 6. Constantius' speech on that occasion has much in common with the elaborate justification offered by Them. Or. 10.131D-2A (already put in a nutshell at 130D). Cf. Heather (n. 76), 118; Heather and Matthews (n. 76), 14ff.

⁹⁷ Themistius on the other hand works overtime (Or. 10.132A, 133C-D, 135C-D, 140A) to insist that Valens dictated the peace, which he granted only as an act of charity when he could, had he wished, have annihilated his terrified, grovelling opponents. His protesting too much is perhaps the strongest argument in favour of the view of Heather (n. 76), 116ff., and Heather and Matthews (n. 76), 20–1, that the terms favoured the Theruingi more than the Romans. But even if this is so it is hard to believe that Themistius actually admits that the peace was a compromise, as is implied by the translation of Or. 10.135A offered at Heather and Matthews (n. 77), 43: ‘cast down in spirit by those aspects of the treaty in which he had lost his case but exulting in those in which success had fallen to him’. The sense required for Ṣιελέγϰθη ‘had lost his case’, is somewhat unnatural, as is the assumption that τῶν σπονṢῶν, ‘of the treaty’, is partitive, rather than being governed directly by ἒτυϰε, ‘he secured’. More importantly perhaps, why should Valens' tame publicist suddenly reveal that everything he had said previously about Valens' crushing victory in the debate was in fact false? We should expect Themistius to be saying ‘Athanaricus was downcast because he had been defeated in debate but elated because he had secured peace’; in other words, the fundamental antithesis should be between Ṣιελέγϰθη, ‘he was worsted in argument’, and ἒτυϰε τῶν σπονṢῶν, ‘he secured the peace’. This would at least be consistent with what has gone before, and would preserve the parallel between Valens' achievement in negotiation and the familiar pattern of Roman success in the field: victory followed by clemency in the granting of terms. Whether the text as it stands can bear this interpretation is questionable; it might be worth considering emending at least the second οἷς, ‘in respect of those things which’, to ώς, ‘because’.

⁹⁸ Crump (n. 46), 117.

⁹⁹ Them. Or. 136A, C. The conception of defensive strategy presented by the orator here, with peace guaranteed by a readiness for war, since in the long term only fear keeps barbarians in check (138B, D), is virtually identical with that of Ammianus.

¹⁰⁰ On problems connected with the nature of this settlement, see Heather (n. 76), 122ff., who rightly stresses that the terms were quite favourable to the Theruingi. It is possible that their conversion to Christianity occurred at this time, perhaps as a condition of the agreement, cf. P. Heather, GRBS 27 (1986), 289ff.

¹⁰¹ Perhaps the scapegoat of an official enquiry, see Heather (n. 76), 132–3, 141.

¹⁰² For events leading up to Edirne as a catalogue of Roman errors, see Bitter (n. 11), 107.

¹⁰³ That Lupicinus is unlikely to have acted on his own initiative is plausibly suggested by Heather (n. 76), 132–3.

¹⁰⁴ Bitter (n. 11), 105.

¹⁰⁵ See Heather (n. 76), 128ff., for a detailed presentation of this view.

¹⁰⁶ Heather (n. 76), 135ff.

¹⁰⁷ This is compatible with Ammianus' estimate of the victory: it tamed the western tribes; that is, the most urgent objective had been achieved.

¹⁰⁸ That the debellation of the arrogant was an essential element even in a defensive strategy is most clearly shown by Constantius' formulation at 15.8.7.

¹⁰⁹ Perhaps significantly, the vision of lasting peace is ascribed to Constantius (15.8.14).

¹¹⁰ For peace through fear, see the passages cited above, n. 4. For breaches of the peace, see 17.6.1, 19.11.5, 15, 26.5.7, 27.5.1, 31.10.2 and esp. 27.10.5 and 28.2.6ff. (See also 17.1.3, 18.2.7, 18.) Contrast 30.3.6.