Chapter 6
The battle of adrianople wiped out two-thirds of the whole field army of the Roman East. It was the worst military disaster of the Roman imperial era, and one of the worst in Roman history. That it was inflicted by barbarians made it instantly controversial, as contemporaries struggled to understand the reasons for the loss. For them, little save divine displeasure could explain such a calamity, so debate centred on which god was angry and why. But from the perspective of the modern historian, the trail of events that led to Adrianople is dotted with human error at every step. The Goths who defeated the emperor Valens at Adrianople in 378 were not a horde of unstoppable invaders. They were, for the most part, the same Goths who had crossed the Danube just two years earlier, in 376, having done so with the full approval of the imperial government. The reception of barbarians into the empire was no unprecedented novelty, but a well-known procedure with centuries of success behind it. Of course, accidents can happen whenever large numbers of people move from one place to another. But the path to Adrianople was no accident. The orderly reception of the Goths broke down through mismanagement and thereafter the imperial government repeatedly exacerbated the problem in a toxic combination of venality and incompetence. And so the crisis marched inexorably onwards to the fatal 9th of August 378.
A modern narrative history is entirely at the mercy of the ancient sources that happen to survive, which condition both its depth and its detail. For the two years before Adrianople, our access to one particular stream of Gothic history grows tremendously larger. The Goths who entered the empire in 376 are better known to us than are any of their predecessors, or indeed any of their contemporaries who remained outside the empire. The pace and scale of our narrative can therefore change with the present chapter. Up till now, we have been able to look at Gothic history in only two ways: first, in a sort of static, analytical overview, based on the archaeological evidence; and second, in brief flashes of narrative when the Goths impinged heavily enough upon Roman imperial politics for our Graeco-Roman sources to leave a record of events. But beginning in 376, for the first time, we learn enough about both Roman and Gothic activity to write a detailed narrative history, one that permits some insight not just into what happened, but also into why and how it did so.
Huns, Alans and Goths
The sources do not reach this level of precision until the Goths arrive at the banks of the Danube in 376, and the train of events that brought them there is known in nothing like the same detail as are the two years that followed it. The basic source is Ammianus Marcellinus, supplemented only very rarely by the fragments of Eunapius or the later sources, like Zosimus, that drew on him. Ammianus gives us a satisfyingly linear account: the Huns, a mysterious and lethal new people, appear as if from nowhere, smash the only somewhat less savage Alans, and drive through the Greuthungian kingdom of Ermanaric, pressing a horde of Gothic refugees forward to the Danube where they clamour for entry into the empire. No one can deny the force of Ammianus’ account, but it has won rather more credence than it deserves from modern scholars. Ammianus always needs careful handling, but here even more so than elsewhere, because the events he describes took place so far from regions in which accurate knowlege was possible. His accountis highly schematic and telescopes what was a long, complicated, and dimly understood upheaval into an implausibly straightforward story of cause and effect.
The Huns of Ammianus appeared from the distant East. For him, they are bipedes bestias, ‘two-legged beasts’: they live on horseback and cannot walk normally as other men do, they scar their children’s faces and drink only mare’s milk, they never cook their food, but rather place raw meat between their thighs and the backs of their horses in order to warm it up.[126] Eunapius once reported something similar, for Zosimus tells us that Huns could not fight on foot because they even slept on horseback.[127] Whereas scholars once took this evidence very seriously, it is now generally agreed that almost every element in Ammianus’ description can be traced to older ethnographic traditions, often stretching back as far as Herodotus, 800 years earlier. Ammianus, we may be fairly certain, had never seen a Hun and nor had most of his readers, who would instead envisage the Huns as the historian intended them to – a patchwork of ethnic stereotypes stitched together to make a composite, but suitably barbarous, whole.
For all that we must distrust it, Ammianus’ account may simply be retailing the sort of rumours that were all most Romans ever heard of events beyond the frontier.[128] His own lack of certain knowledge must explain why his narrative of the Hunnic onslaught lacks any chronological markers. The Huns appear suddenly, at some unspecified time, and overcome the Alans who dwell between the Don and the Caspian. Unlike the Huns, these Alans had long been known to Graeco-Roman ethnography. They made periodic incursions into Roman territory, but were for the most part a greater threat to Persia than to Rome. As early as the second century, Arrian (c. 86–160), the Hadrianic governor of Cappadocia and famous historian of Alexander, had written a tactical manual, the Order of Battle against the Alans, explaining how a Roman army should be disposed in order to repel the charge of Alanic cavalry. Arrian was a keen observer, but even in his own time, they had been confounded with ethnographic stereotypes in existence since the time of Herodotus, and his sketch of their tactics is not very informative. By the fourth century, Ammianus’ sketch of the Alans does little more than nod towards the conventional Graeco-Roman image of the horse-nomad. Starting from these Huns and Alans, Ammianus narrates a simple chain reaction, one group of barbarians pushing against the next until eventually the massed Tervingi appear on the banks of the Danube.
The Defeats of Ermanaric and Athanaric
The Alans, so we are told, joined forces with the Huns after being defeated by them. In the company of their new Hunnic masters, they went on to assault the borders of the Gothic Greuthungi. These Greuthungi were led by the ‘most warlike’ king Ermanaric, whom we met briefly in the last chapter. Ermanaric determined to make a stand against his enemies, but to no avail. In the end, he committed suicide rather than face the coming horrors. A new Greuthungian king, Vithimir, succeeded him, and like his predecessor determined to make a stand on the battlefield. Unlike his predecessor, he lost his life in battle. Thereafter, his little son Videric was made king, but two duces – a generic term which Ammianus uses for subordinate commanders – acted as the new king’s guardians and seem to have taken Greuthungian affairs into their own hands. These duces, named Alatheus and Saphrax, led the Greuthungi of Videric westwards to the Dniester river. There, according to Ammianus, their plight came to the notice of the Tervingi and theiriudex Athanaric.[129]
Athanaric, Valens’ old enemy, advanced with an army to the banks of the Dniester, where he encamped at a safe distance from the Greuthungi. Sending an advance guard to observe and perhaps intercept the Huns, he waited at the Dniester, but was surprised by the Huns’ strategic skill. A party of Huns crossed the Dniester in the night, marched down it to Athanaric’s camp and forced him to withdraw into some unnamed mountains, perhaps the foothills of the Carpathians where he had previously sought refuge from Valens. What appears in Ammianus as a tactical retreat was in fact a massive withdrawal, for nearly 200 kilometres separate the Dniester from the line which Athanaric next determined to hold. This ran from north to south above the Danube, just beside the Carpathian foothills, and probably reconstituted the old Roman limes transalutanus, parallel to the river Olt and the frontier of the high imperial province of Dacia. Despite the efforts he put into throwing up earthworks and other defences, Athanaric’s new measures came too late. Though he repulsed a Hunnic attack somewhere in the region, many of his followers had already deserted him. The populi pars maior, ‘the larger part of his people’, left their stubborn leader to fight his own battles, themselves seeking refuge in the empire. The Greuthungi under Alatheus and Saphrax, for their part, disappear from view until 377, several months after many Tervingi were allowed to cross the Danube into the empire.[130]
The Chronology of Gothic Defeats
The foregoing narrative raises more questions than it answers, in large part because it derives exclusively from the last book of Ammianus’ history. His account is heavily telescoped: its stages may be well defined, but its actual chronology is almost totally invisible. Even if Ammianus’ account is substantially correct – and the very linear trajectory he suggests must be suspect – the series of conflicts among Huns, Alans and Goths will have taken much longer than the headlong rush implied by Ammianus. What is more, it is not easy to sustain his simple ‘domino-effect’ theory of causation, with the Huns toppling the Alans onto the Greuthungi onto the Tervingi onto the Romans. To be sure, the emergence of the Huns somewhere between the Caspian and the Black Sea probably did spark far-reaching changes in eastern and central Europe. But it is harder to make the case that the Huns were the proximate cause of Gothic collapse, rather than its catalyst. No named Hun appears on the frontiers of the empire until the very end of the 390s, two full decades after the disaster at Adrianople. Even then, it is another three decades before there is evidence for a Hunnic state, or even far-reaching Hunnic hegemony, in the barbarian lands near the empire where the Greuthungi and Tervingi had once held sway.
These facts suggest that, however much Ammianus may have envisaged Hunnic wolves snapping at the heels of the fleeing Goths, the process was altogether more gradual, not to mention more complex. The time frame of these events is unclear from Ammianus, and wholly beyond reconstruction from other sources, despite the best efforts of scholars. A reasonable guess might place the early confrontations between the Huns and Alans, and then between the Huns, Alans and Greuthungi, as far back as the 350s, but that can be no more than speculation. The only certainty is that the disruptions along the Don river and north of the Sea of Azov had not yet been felt along the lower Danube when Valens made his treaty with Athanaric in 369. As we have seen, the persecution in which Saba was martyred was a response to the campaigns of Valens and Saba’s Passion gives no hint at all of traumatic upheavals to the east. Even though that may be no more than a reflection of the hagiographic genre and its constraints, nothing else in the evidence for Valens’ campaigns shows the slightest awareness of trouble beyond Athanaric’s realm. Thus when all is said and done, our only firm chronological indicator is the arrival of a large number of Tervingi on the banks of the Danube in spring 376.
Tervingian Petition and Imperial Response
Early in the year, before the start of the campaigning season, masses of Tervingi occupied the northern bank of the river, begging for admission into the empire.[131] They offered to abide peacefully inside the imperial frontiers and to furnish auxiliaries for the Roman army if required to do so.[132] These Tervingi were divided into many different groups, without any overall leadership. The one leader who might have claimed some sort of supremacy, Athanaric himself, was certainly not among them, fearing that the breach between him and Valens was too great for him ever to be admitted to the empire. We hear of two Tervingian chieftains, Alavivus and Fritigern, in the context of the Danube crossing, and it is clear from later events that they led not all the Tervingi, but rather the most significant of several independent bands. Although Alavivus led the negotiations with the empire, Fritigern was perhaps the more powerful of the two chieftains. He was probably better known to the empire, if Socrates’ story of his conversion to Christianity in the earlier 370s is correct, and he was certainly the more competent general, for by 377 he was in overall command of the Goths’ military operations.[133] As to Alavivus, we can reject outright speculation that he was the father of the later general Alaric, a theory based on nothing more than the alliteration of their names.
Negotiations over terms of entry must have taken quite some time, certainly several months, given that messengers and ambassadors had to travel more than a thousand kilometres to Antioch in Syria before returning to Thrace with the imperial decision. Even if the senior negotiators moved very fast indeed, as a letter of Basil of Caesarea implies, agreements could not have been reached before high summer.[134] How order was maintained in the interim, we do not know, nor how the massed Tervingi were kept supplied with the necessities of life. But since we have no evidence for any disturbances during the ongoing negotiations, we must postulate a firm hand among the Goths, and an assumption on both sides that everyone involved was negotiating in good faith. In other words, the Tervingian leaders must have felt it likely that their petition would meet with success – success which was, to a degree, dependent upon their good behaviour.
From Valens’ point of view, the Tervingian offer was both opportune and welcome – every source tells us as much, and there is no reason to disbelieve that evidence or maintain that the Gothic entry to the empire was permitted only because it could not be repelled. The emperor was in the midst of preparing a substantial war against Persia, made necessary by complicated manoeuvres over who should control the kingdoms of Armenia and Iberia. Persian wars were always costly and sufficient manpower could be hard to come by. If the Gothic petitioners were indeed allowed into the empire, that would fulfill the promises that the Valens’ orator Themistius had made in 369 at the end of the last Gothic war.[135] In that speech, as we saw in the last chapter, Themistius had been forced to put a good face on a clearly compromised peace, arguing for public consumption that the empire benefited more from sparing its enemies and keeping them alive as potential soldiers than it did from their destruction. What was at the time an argument of necessity, and a weak one at that, could now be made into happy reality for all concerned. The Tervingi could be admitted as humble suppliants, and then formed up into units to be dispersed to the eastern frontier. Given that, it is no wonder that Valens seized the chance which fate had offered him, giving orders that the Tervingi should be allowed to cross the river, fed for a time, and thereafter offered lands to farm; the Goths, for their part, probably gave hostages to the imperial government to guarantee an orderly crossing and settlement.
The Crossing of the Danube
It took several days and nights to transport all the Goths across the river, and Ammianus gives the impression of people coming over in their thousands.[136] Where exactly the transfer took place is unclear, though Durostorum, on a straight road line south toMarcianople, seems the likeliest point. Numbers are likewise not to be had. Eunapius speaks of 200,000 Goths, but few have taken seriously a figure that high. Although it has recently been defended as plausible in light of the constant losses which the Goths suffered in the course of the next six years – to have lost so many, there must have been masses of them to start with – that position gives too little weight to significant re-enforcements which the band received in those same years. Questions of manpower in the ancient world never have clear answers, and the best we can say here is that the scale of later fighting implies that the Goths admitted to the empire numbered at least in the tens of thousands, and perhaps considerably more than that.
If Alavivus and Fritigern were the first to be received, there were other Gothic commanders as well. They came voluntarily, not in response to military defeat by the emperor, which may explain their relative strength. Certainly, few of them were disarmed. Standard imperial practice was to disarm barbarians before they were admitted to the empire, and only then to re-arm them out of imperial arsenals at times and places where they could not pose a threat. In this case, however, whether through corruption, neglect, or the sheer scale of the enterprise, many of the Goths retained the weapons they normally carried, despite the clear intention of the emperor that they be disarmed in the usual fashion.[137] When this oversight was combined with appalling abuse, the situation became volatile indeed. The officials put in charge of the crossing were Lupicinus and Maximus, the first a comes rei militaris, the second the dux of either Moesia or Scythia. For Ammianus they were homines maculosi, ‘men of tarnished repute’, but that seems like the judgement of hindsight, perhaps even the verdict of a later imperial inquiry into what had gone wrong in the lead-up to Adrianople.[138] Imperial officials were expected to profit from the offices they held, and we should not assume that the exploitation of the Goths by Lupicinus and his officials was in excess of the late Roman norm. Nor should we discount the possibility that limiting the Goths’ food supply was a deliberate way of controlling what was, after all, a large and potentially dangerous body of barbarians on imperial soil. By modern standards, however, the abuse was shocking. The food that ought to have been allocated to the Goths was diverted by the generals for sales that would line their own pockets. In its stead, the Goths were offered dogmeat at the price of one dog for one Gothic child enslaved. According to Ammianus, even the children of Gothic nobles could be rounded up and sold on to slavers.[139]
Alatheus and Saphrax
While this trouble was brewing, the Greuthungi of Alatheus and Saphrax – those Gothic duces who had taken custody of the child-king Videric – also arrived at the Danube, seeking entry into the empire. As Alavivus had done some months before, these two generals now sent envoys to Valens offering terms and asking for succour. Somewhere in the same vicinity, old Athanaric too had arrived, though it is not clear what finally drove him to seek refuge in the empire. We do not know why, but the request of Alatheus and Saphrax was refused. Some have argued that the emperor began to fear the consequences of letting in too many Goths at once, or that the Tervingi already inside the empire had become so restive that additional newcomers would impose too heavy a burden on an already overwhelmed officialdom. Perhaps, though, the treatment of the Greuthungi was simply a very public demonstration of the imperial power over barbarians: after all, the gesture proclaimed that the decision of whether or not to admit different Gothic groups was entirely in the hands of the emperor, who could pick and choose with total inscrutability. That, at least, was the lesson Athanaric learned: seeing the Greuthungian request rejected, he gave up on any prospect of accommodation and retreated to ‘Caucalanda’, perhaps the Transylvanian Alps, where he was to remain with his followers for half a decade. But if the deliberate arbitrariness of the imperial position was meant to intimidate the Greuthungi and cool their ardour, it did not do so. Instead, Alatheus and Saphrax bided their time.[140]
The Tervingi, for their part, were understandably dissatisfied, and Lupicinus began to fear unrest. He decided that the time had come to move them, and the coming of spring 377 made possible their dispersal out of winter quarters near the Danube. As Lupicinus and his officials began to organize the relocation of their charges, river patrols were neglected and the Greuthungi of Alatheus and Saphrax saw an opportunity to take for themselves what imperial orders had denied them. They crossed the river in makeshift boats and pitched camp at a great distance from where the Tervingi of Fritigern were being formed up for relocation at Marcianople (now Devnja in Bulgaria).[141] That substantial city, founded by Trajan during the Dacian wars, lay nearly 100 kilometres south of the Danube and was located at the junction of the east-west road to Nicopolis-ad-Istrum with the north-south road that led around the eastern edge of the Haemus mountains and down into open and populous Thrace. It was the ideal place from which to organize a major venture of this sort, and thus served as the headquarters of Lupicinus. At Marcianople, however, the disastrous train of events already underway became unstoppable.
A Treacherous Banquet
Lupicinus invited Fritigern and Alavivus to be entertained as his guests at Marcianople.[142] This was a perfectly normal gesture, for local commanders customarily invited officers in transit through their region to dine with them. If, as we must assume, Fritigern and Alavivus were being treated as the de facto commanders of Gothic units destined for inscription into the Roman army, then their reception and entertainment by Lupicinus makes perfect sense. At the same time, however, banquets were one of the usual venues for treachery in the Roman world. It was at banquets that usurpations were plotted and often set in motion, and it was at banquets that prominent barbarian hostages might be seized and spirited off to captivity.[143] Themselves quite innocent of treachery, Fritigern and Alavivus walked into a trap at Marcianople.
The Gothic leaders took up temporary residence within the city together with a small group of attendants, but Lupicinus kept their main following at a good distance from the town, interposing Roman troops between the Goths and the city walls. Before too long, confused brawling broke out between these two groups, prompted by the Romans’ steadfast refusal to allow Goths into the city to purchase supplies, and perhaps by the continued attentions of slave traders. In the riot, a number of Roman soldiers were killed and robbed by the frustrated Goths. News of this reached Lupicinus while he and his guests were taking their leisure and deep in their cups. Clearing his head and seeking to forestall a full-blown revolt, Lupicinus ordered the resident bodyguard of Alavivus and Fritigern to be executed. Though the order was carried out in secret, rumour of it spread rapidly, and the Goths outside the city prepared to storm the walls. Fritigern, conscious of his own danger, convinced Lupicinus that the only way to avert catastrophe was to demonstrate to his followers that he, at least, was still alive. Lupicinus immediately grasped the wisdom of this counsel. Fritigern and those of his attendants who were still living went out to their followers and were greeted rapturously; Alavivus, by contrast, is never heard from again, perhaps killed or retained as a hostage, perhaps even betrayed by Fritigern as a dangerous rival.
The Gothic Rebellion
Then, however, rather than attempt to retrieve the situation and carry on with his reception into the empire as planned, Fritigern made a momentous decision. In the face of constant harassment and sudden betrayal, he would reject the terms under which he had been received into the empire, lead his followers away from Marcianopole and into open revolt. He and his Tervingi therefore marched out into the province of Scythia, and as news of Lupicinus’ treachery spread, all of the Goths who had crossed the Danube the year before joined Fritigern. Why did things go so very badly wrong at Marcianopole? Modern scholars, influenced by the black colours in which Ammianus paints Lupicinus, tend to assume that he plotted treachery from the beginning. That seems unlikely given the normal habits of Roman officialdom. Exploiting the perquisites of office to get rich was one thing, deliberately provoking a rebellion another thing altogether. If Fritigern’s Goths were already destined for a secure place in the Roman army, as other Goths in Thrace certainly were at precisely this time, then Lupicinus had nothing to gain from eliminating Gothic commanders who had up to that point kept their following obedient and quiescent. Again, the banqueting at Marcianople, and indeed the separation of commanders and attendants from the main body of troops, was perfectly normal – it is exactly paralleled twenty years before, when the caesar Julian entertained his high commanders at Paris while their units were encamped well away from the city itself. Although Lupicinus must have seen that he had a chance of entrapping the Gothic leaders at Marcianople, it seems most unlikely that he actually planned to do so from the beginning. On the contrary, when riotous skirmishing flared up between Gothic and Roman troop, Lupicinus panicked. That panic, in turn, convinced Fritigern that his only safety lay in rebellion.
Retreating from Marcianople, Fritigern and his followers were pursued by Lupicinus and the army stationed there. Fourteen kilometres from the city, the two forces clashed and Lupicinus’ army went down to bloody defeat. The whole of its junior officer corps died on the field, the unit standards were lost, and Lupicinus himself only survived by escaping into Marcianople and shutting up the city behind him. Fritigern’s Goths equipped themselves with the weapons and armour of their fallen enemies and went on the offensive, raiding nearby regions, and then ranging further afield, as far south as Adrianople, about 320 kilometres to the south. We can be fairly certain that the rebellion would have been halted in its tracks had Lupicinus been victorious. Success, however, breeds confidence and Fritigern and his followers, tormented by Roman exploitation for long enough, were now in no mood to see reason. To their standards flocked not just the other Goths who had been admitted into the empire, but also the dissatisfied and oppressed of the provinces – slaves, some of whom were Gothic, miners, and prisoners of all stripes. These, in turn, made rebellion easier, for they knew their way around the provinces, knew the roads and imperial establishments, and thus made the task of supplying the rebels far less complicated than it would otherwise have been.[144]
The Rebellion Spreads
Gothic units in the Thracian army soon joined Fritigern as well. Two commanders named Sueridus and Colias, in winter quarters at Adrianople with their units, had observed with total unconcern the travails of the Tervingi admitted in 376. Nor did the revolt of Fritigern at Marcianople interest them. The fact that Sueridus and Colias demonstrably lacked any special feeling for fellow Goths is a salutary reminder that only extraordinary pressure of circumstances could turn different groups of Goths into ‘the Goths’. In this case, that pressure came from the managerial incompetence of local officials at Adrianople. Early in 377, Sueridus and Colias received their marching orders, detailing them to the eastern front where they were needed for Valens’ Persian campaign. When they asked the local authorities in Adrianople for money to equip their units with food for the journey, they were refused by the head of the local city council, the curia. Ammianus tells us that the councillor was angry with the followers of Sueridus and Colias for the damage they had done to his suburban property. Now it is true that the quartering of a Roman army – any Roman army, regardless of who composed it – was a severe burden on townsfolk, but the magistrate was not acting solely out of anger. While cities were obliged to house and feed Roman units quartered on them, the legal obligation of the curia to give troops supplies for a march was by no means clear. Indeed, on most readings of late Roman practice, imperial officials should have taken charge of equipping Sueridus and Colias’ troops for their journey, without involving the curia of Adrianople at all.
The curia armed and brought out the staff of the local imperial arms factory – the fabricenses – and with that force at their back ordered Sueridus and Colias to be on their way at once. Even if the legal right was on their side, for the curia to have refused the generals’ request with such brusqueness was political stupidity of the highest order. Sueridus and Colias were genuinely shocked by the unexpectedly heavy-handed treatment and made no move to go. At that point, no doubt egged on by their magistrates, the townsfolk and fabricenses began to harass the soldiers, pelting them with makeshift missiles and attempting to drive them off by force. Thus provoked, the soldiers of Sueridus and Colias fought back and, as usually happened when imperial troops were turned loose on civilians, massacred whomever they got their hands on. That done, and presumably well armed with the stores from the imperial fabrica, they marched their followers off to join Fritigern.[145]
As this one example shows, the Gothic rebellion in Thrace was not a single, planned affair, still less a barbarian migration. It was, on the contrary, a series of local revolts that in time converged into a mass uprising which threatened not just those regions in which the rebels were active, but the security of the Danubian provinces as a whole. There is no point in tracing in detail every skirmish mentioned in our sources.[146] They are too similar and we know too little about how they were connected to one another. However, one vital point is abundantly clear: the Goths under the overall command of Fritigern transformed themselves into a potent fighting force in a very short space of time. Equipped with Roman arms and armour, they also constructed a substantial supply train which allowed them to carry with them foodstuffs and other necessities gathered from the well-stocked regions through which they passed. This large force was made up of Goths from many different backgrounds, as well as all sorts of provincial malcontents. It was no longer the group of Tervingi that Fritigern and Alavivus had led across the Danube the year before, and Ammianus recognizes this fact by ceasing to speak of Tervingi and beginning to speak generically of Gothi, ‘Goths’. These Gothiroamed more or less at will in the land between the Haemus mountain chain and the Danube during 377 and most of 378. The rationale behind these movements is totally obscure, but it is striking that neither the Roman nor the Gothic side seem to have made any effort to negotiate throughout this period of more than a year. It is possible that Valens is to blame. If the Roman generals on the spot acted indecisively, it may be that they had received no guidance from an imperial court more interested in Persian affairs. The Goths, after all, were barbarians, and northern barbarians had always taken second place to Persia. In those circumstances, lacking direction from above and not wanting to take the wrong decision with so unpredictable an emperor as Valens, Rome’s Balkan commanders can hardly be faulted for trying to contain the Gothic threat rather than suppress it.
The Imperial Response
At some point in 377, however, Valens became convinced of the seriousness of the problem. He determined to patch up a truce with the Persians over Armenia, sending his longest-serving general, Victor, to negotiate it.[147] In preparation for his own eventual advance, he sent the generals Profuturus and Traianus to keep the Goths in Thrace under control. Meanwhile, Valens’ nephew Gratian likewise realized the gravity of the situation. He despatched two good generals, Frigeridus and the comes domesticorumRichomeres, to support the eastern troops, but also to ensure that the trouble was contained in Thrace and Moesia and did not spread westwards into Pannonia and the Latin provinces.[148] Gratian’s intervention demonstrates how worrisome the Gothic revolt had become during the course of 377. Western generals did not, as a rule, intervene in eastern affairs, nor junior emperors in those of their seniors, lest it look too much like provocation. As recently as 366, Valentinian had declined to help Valens face down the usurpation of Procopius, a far more direct threat to dynastic control than the Goths could hope to be. Only the prospect of chaos along the whole Danube frontier can have prompted Gratian’s intervention.
As it happened, Frigeridus fell ill and returned to the West for a time, leaving Richomeres to lead the western troops to their rendezvous with Valens’ generals Profuturus and Traianus. In late summer 377, they brought the Goths to battle near a site called Ad Salices (‘the Willows’). The precise location of this site remains unknown, though it probably lay somewhere between the coastal town of Tomi and the opening out of the Danube delta into its many channels, very near the imperial frontier rather than in the immediate vicinity of Marcianople. The battle of Ad Salices that followed was a major one, but a draw, for the Goths were secure within their well-guarded wagon train and could retreat into it as necessary. The Roman forces seem to have been smaller than the Gothic, and Profuturus himself fell in battle, but superior drill and training saved the army from total destruction. Having suffered too many losses to continue the assault, the Roman troops retreated south again, back to Marcianople, where the revolt had first begun in earnest.[149] At roughly the same time, Frigeridus returned to the East, fortified Beroe, and inflicted a major defeat on another Gothic noble, Farnobius, who had been raiding through Thrace. Frigeridus sent the survivors of the slaughter back to Italy, where they were settled as farmers, a useful reminder that barbarian settlement within the empire could work perfectly well when managed with a minimum of care.[150]
Despite Ad Salices, Richomeres and the other generals had inflicted serious damage on Fritigern’s Goths. Many of them withdrew into the safety of the Haemus mountains for the winter of 377–378. Richomeres went back to Gaul as autumn fell, planning to collect a larger force for the following year’s campaigning season. Valens, for his part, re-enforced his troops in Thrace with a more senior commander, the magister equitum Saturninus. He, with Traianus as his lieutenant, blockaded the Goths in the Haemus passes and deprived them of food. He hoped that by reducing them to desperate hunger and then removing the guards from the passes he could lure them into the open country and destroy them in pitched battle. The plan failed. Rather than moving north and standing to battle in the plains between the Haemus and the Danube, the Goths allied themselves with some unspecified Huns and Alans, and made their way south into Thrace. In that country’s wide open spaces, with their excellent roads, Fritigern could move freely, laying waste great stretches of land between the Haemus, the Rhodope, and the shores of the Hellespont and of the Bosporus near Constantinople.[151] So badly ravaged were the provinces of Moesia and Scythia that the emperor officially lowered their tax burden in 377.[152] Indeed, by early 378, much of Thrace itself was inaccessible to the outside world: Basil of Caesarea wrote to an exiled fellow-churchman, Eusebius of Samosata, then resident in Thrace, commenting on the unprecedented difficulty of communication and expressing surprise that Eusebius had managed to survive there at all.[153]
Valens Prepares for War
Valens’ generals Saturninus and Traianus may have had only limited success, but Gratian’s commanders managed to quarantine the revolt. By early 378, Frigeridus had fortified the Succi pass, the vital conduit between Thrace and the western Balkans.[154]Thereafter, Fritigern’s Goths were effectively confined to Thrace. In that same year, not just Richomeres but Gratian himself led a large portion of the western army into the eastern empire to assist his uncle. He had wanted to come sooner, but some Alamanni in the Rhineland detained him: hearing of the troubles in Thrace and Gratian’s plans to assist in their suppression, they seized a chance to raid into the western provinces.[155] Only in 378 could Gratian spare his main army for the Gothic war. By then, Valens had settled eastern affairs to the point where he felt able to march to Thrace. He arrived in Constantinople in spring 378, staying there for perhaps twelve days and facing down riots among a discontented populace, one no doubt frightened at the continuing Gothic presence on their doorstep.[156] Valens’ first move was to reorganize his officer corps, dissatisfied with their conduct up to this point, and not without good reason. In place of Traianus – whom Valens personally blamed for failing to stop the Goths at Ad Salices – the retired western general Sebastianus was made commander-in-chief and was perhaps given a strike force drawn from the emperor’s own seasoned palatine troops.[157] Certainly, he quickly won a couple of surprise victories over Gothic raiding parties.[158] But this welcome success brought an unexpected side effect: fearing lest his various followers be picked off piecemeal, Fritigern ordered them to form together and operate as a single unit. From their rendezvous point at Kabyle, a well-watered and easily defensible site in the plain between the Haemus and Rhodope mountain chains, the whole of the Gothic force began to make south for Adrianople. There Sebastianus was headquartered, sending back to Constantinople reports of his recent successes. On 11 June, Valens left Constantinople for what would prove to be his last journey.
The Battle of Adrianople
What actually happened on the battlefield of Adrianople is remarkably ill documented for so decisive a moment in Roman history, and one so comprehensively discussed in contemporary writings. Unfortunately for the modern historian, contemporary interest was chiefly concerned to explain why the disaster happened, not how it unfolded. Ammianus, as so often, gives us our only detailed account of the battle, but his outline of events includes substantial gaps – some of his own making, some the product of a faulty manuscript tradition – so that a tactical description of the battle is impossible. Nevertheless, Ammianus’ broad outline seems clear and is corroborated by other sources. In the first week of August, Valens marched his field army – between 30,000 and 40,000 men, in all likelihood – out from its staging post at Melanthias, just west of Constantinople. The emperor made for Adrianople with all haste, supposedly jealous of the successes that Sebastianus had won and wanting a share of his general’s glory. Fritigern’s Goths bypassed Adrianople and its substantial garrison, making instead for the road-station at Nike. There the Gothic army was observed by the imperial scouts who fanned out in advance of the emperor’s main force. The intelligence they brought back was misleading, suggesting that the Gothic forces numbered only 10,000 men, much less than their real number. This news gave Valens, eager for battle and a victory he could call his own, all the excuse he needed to attack at once.[159]
Advancing to Adrianople, he fortified a camp in the suburbs of the city and impatiently awaited the arrival of his nephew’s army. Perhaps on the 7th of August, the general Richomeres arrived with the western advance guard, advising Valens to wait the very short time it would take for Gratian’s main force to arrive.[160] Delay, however, did not suit Valens, and he called a meeting of his high command to debate the issue. The generals themselves were deeply divided, but which generals argued for which plan is unclear: in the aftermath of the disaster, contemporaries strove to shield their favourites from blame and shift it onto others, a task made easier by the death of almost all those who had witnessed the debate. Thus Ammianus claims that Sebastianus led the group which argued for an immediate assault on the Goths, while the magister equitum Victor led those who argued for the delay that would guarantee victory. Eunapius, by contrast, defended Sebastianus, as is clear even from the very confused Eunapian chronology preserved by Zosimus.[161] Regardless, the council decided on the course of swift action. Valens favoured it, and his civilian officials played upon his natural jealousy, suggesting that he ought not to share with Gratian the glory of an inevitable victory.
Roman victory was, after all, expected by everyone, not least the Gothic leader Fritigern. At Adrianople, within striking distance of the imperial army, he showed himself more eager for a peaceful settlement than at any time since the very first crossing of the Danube. Perhaps he feared risking battle in the continuing absence of the Greuthungi under Alatheus and Saphrax, whom he had long since sent for. Perhaps, on the other hand, he worried that the Goths could not defeat a proper imperial field army when their victories had thus far come only against smaller, provincial commands. Be that as it may, on 8 August he sent a Christian priest and some provincials of humble status to offer terms to the emperor: he and his followers, poor exiles driven from their own lands and with no place else to go, wanted only Thrace with its crops and its lands. In exchange for that, he could offer the emperor lasting peace. Thus ran Fritigern’s public message. With it came a private message for Valens himself, in which the Goth assured the emperor that he really did want peace, but that for him to enforce himself upon his followers, the emperor would have to keep his army mustered and active as a visible threat to the Goths. Valens distrusted these overtures, and at any rate wanted very much to fight a battle he was convinced he could win.[162]
Thus on the morning of the 9th, leaving his civilian court officials and his treasury safely inside the walls of Adrianople, he marched his troops northeastward out of their encampment into the rolling plain where Fritigern and his army were based. We cannot really be sure how many men either side fielded, but tens of thousands of men went into battle on that August morning. Not long before noon, the Romans spotted the Gothic camp, probably near the modern village of Muratgali. Massing on a low ridge line in front of their wagon circle, the Gothic warriors were well rested and eager for battle. Valens began to dispose his troops in line of battle, cavalry units on each wing, and the mass of his infantry in the centre. Neither side was as prepared for a pitched battle as they might have been: the left wing of the Roman army was still scattered in columns-of-march, while the Greuthungi under Alatheus and Saphrax had not yet arrived. Fritigern therefore played for time, sending envoys to beg for peace while the imperial forces roasted in the blazing sun, and choked on the fires which he had lit to punish them further. Watching as the condition of his troops deteriorated, Valens thought better of his refusal to negotiate – possibly he even decided wait for Gratian – and made ready to send higher-ranking officials to meet representatives of the Goths.[163] This was a mistake, and one cannot imagine Valentinian or Constantius Ⅱ opening protracted negotiations with the enemy while their soldiers’ readiness withered away in the wake of a forced march. Yet as so often happened in ancient battles, fighting began by accident, before either side was ready.
Two units of the elite scholae palatinae, the Scutarii under Cassio and the Sagitarii under Bacurius, probably on the right wing and near to the emperor where scholae were usually posted, advanced prematurely and engaged the enemy.[164] Their move disrupted the imperial line of battle, which was then disordered still further by the sudden appearance of Alatheus and Saphrax and their followers, in company with a unit of Alans. What followed was a military disaster, described by all our sources in lurid colours. The Roman left wing drove too far beyond the Gothic line and was cut off, surrounded and slaughtered. With the main infantry’s left flank thus exposed, the Roman line was compressed in on itself, hampering the ability of the soldiers to fight and causing many to die from wounds inflicted by their own side. Towards late afternoon, the Roman infantry line broke and the rout began. The imperial bodyguard and the scholae palatinae must have been almost totally destroyed, for Valens was forced to take cover with theMattiarii, a unit of the regular field army rather than an imperial schola, but seemingly one of the few Roman units to have stood its ground. Some of the generals attempted to rally the auxiliaries who had been held in reserve, but these had already melted away off the battlefield. Seeing that further attempts at rallying the disintegrating army were useless, the generals Victor, Richomeres and Saturninus fled the field. There, the butchery continued until nightfall.[165]
The fate of Valens was uncertain even at the time. Some said that towards evening he was struck by an arrow and fell dead amongst the common soldiers. Others claimed that, mortally wounded, he was carried off the field by a few loyal bodyguards and eunuchs, and hidden in a farmhouse; there, as the emperor lay dying, Goths surrounded the farmhouse and, rather than waste time breaking in, set the house ablaze and burned to death the emperor and his attendants. Only one man escaped through a window and explained that the Gothic firebrands had just deprived themselves of the glory of capturing a Roman emperor on the field of battle. Whichever story – if either of them – was true, Valens’ body was never recovered.[166] With him at Adrianople fell the generals Traianus and Sebastianus, the tribune and Valens’ relative Aequitius, thirty-five senior officers, and fully two-thirds of the army that had taken the field on the morning of 9 August 378.[167] As Themistius would put it five years later: ‘Thrace was overrun, Illyricum was overrun, armies vanished altogether, like shadows’.[168]