`Now the sun climbed higher ... [and] the Romans were weakened by hunger and thirst, and burdened by the weight of their equipment. In the end the great force of the barbarian onslaught shattered our battleline ... Some fell without seeing who struck them, or were knocked down by the sheer weight of the attackers, or even killed by comrades... . In amongst the ordinary soldiers the emperor was struck by an arrow, and soon breathed his last - or so it was believed for no one ever claimed to have seen him or been beside him - and his body was never found subsequently.' - Ammianus'account of the disaster atAdrianople.'

On 17 November 375, Emperor Valentinian was on the Upper Danube, receiving a delegation of chieftains from the Quadi - Marcus Aurelius' old adversaries, who had recently raided into Rome's Pannonian provinces. Valentinian's campaigns were always as much, if not more, about diplomacy than actual use of force. He was known as a quick-tempered man, something that goes beyond the well-entrenched fourth-century stereotype of the irascible and ill-educated Illyrian. When the chieftains claimed that the raids had been launched without their consent by bands of foreigners, and that in fact recent Roman building of new fortifications was provocative, the emperor flew into a rage at such insolence. In the middle of his violent harangue, Valentinian had a stroke and died. He was fifty-four.'

Some years before, Valentinian had named his older son Gratian as Augustus. The lad was now sixteen and had been left at Trier by his father. His younger brother Valentinian I I was only four, but was immediately also proclaimed as Augustus by troops and bureaucrats on the Danube. Neither Valens nor Gratian had sanctioned this, but they did not feel able to reject the child's elevation. Valentinian, and through him Valens, were both selected as emperor by an influential group of senior bureaucrats and army officers. Throughout their reigns the brothers needed to be careful to keep these men happy. It is notable that a number of the most prominent officials remained in office for many years, far longer than was typical in the past. Julian's reign in particular had been marked by a rapid turnover of men holding senior posts. Distinct cliques of senior officials dominated the administration of the territories controlled by each of the emperors. Such men had no desire for reuniting into a single administration under a sole emperor, unless they could be sure that they alone would monopolise the senior positions. Valentinian and Valens knew that their dynasty was too recent to be fully secure and that they must respect the views of their senior officials. In 375 enough senior men decided that they wanted a separate court and administration under the nominal control of Valentinian II to force the hands of Valens and Gratian.3

The empire was once again divided into three. Valens remained in control of the eastern provinces, while Valentinian II was given Italy and North Africa. Illyricum was also technically part of his territory, but in practice this and the remainder of the west was controlled by Gratian. In spite of his age, the latter was active on the frontiers from the beginning of his reign, continuing his father's round of punitive expeditions and forceful diplomacy. For the moment the groups of officials dominating the imperial courts were content with and able to maintain the rule of a youthful and an infant emperor.4


In 376 a large group of Goths massed on the far bank of the Danube. This was not a raiding party, but an entire people on the move, their women and children riding in wagons. They were called the Tervingi, although they were not all of the people who called themselves by that name. There was another major group of Tervingi, and altogether at least half a dozen distinct groups of Goths are known from our sources - more may well have existed, but are simply not recorded. The Goths, just like the Almanni, Franks and others, remained a deeply divided people, split into tribes and other groups, loyal to many different kings, chieftains and magistrates. In the fifth century kingdoms were carved out from Roman territory by the Ostrogoths and Visigoths. There is no evidence that these groups already existed in Valens' day, under these or any other names. Although they figure in older accounts of the 370s, the Visigoths and Ostrogoths did not come into being for another generation.5

The Tervingi sent envoys on ahead seeking permission from Valens to cross into the empire. They asked to be settled on land, preferably in Thrace, and promised in return to provide soldiers for his army. At the time the emperor was in Antioch, for there was continued friction with the Persians over the control of Armenia, so inevitably there was a delay of a month or more before his reply arrived. His past relations with the Gothic tribes had not always been happy. From the beginning of his reign there was tension. In 365 a contingent of 3,000 Gothic warriors answered the summons of Procopius. They arrived too late to make a difference and excused their action by saying that they felt obliged to honour their old treaty with Constantine by supporting any member of his house, however distant. It is hard to tell whether this was genuine. Roman civil wars must often have been confusing to war leaders from outside the empire. They were also great opportunities to profit and the Goths who rallied to Procopius may simply have felt that a usurper was more likely to be generous if they aided his victory.'

Valens was unimpressed by these excuses and spent the next three summers campaigning on the Danube. There was little fighting, for the Goths avoided battle and took refuge in the mountains. Nevertheless, the Romans' display of force was sufficient to prompt negotiation. Valens held a meeting with the Tervingian King Athanaric on a barge moored midway across the River Danube, honouring the latter's solemn oath never to set foot on Roman soil. Both sides paraded their troops on the opposing banks. Valentinian had once similarly conducted talks on a river boat, but in each case the willingness to do so granted a measure of equality to the barbarian leaders involved. Traditionally, representatives of Rome had negotiated in a way that made clear the empire's overwhelming superiority, making the enemy come to them and bow down before a tribunal and the serried ranks of the legions. By the later fourth century it was often more important to gain a quick peace than to insist upon such displays.7

Valens needed the Goths to be quiet so that he could deal with the escalating tension with Persia. They agreed to keep the peace and would no longer receive subsidies. At first sight this was clearly a penalty for the Gothic chieftains, but it may well be that in a gift-giving society, receiving anything from an outside power was a clear mark of dependence. They may have seen this as a considerable gain. Similarly, the restriction on any trade across the border except at two nominated posts probably reinforced the power of the Gothic leaders best placed to control access to these points. Athanaric was probably well satisfied by the treaty of 369. Like Valens, he and the Goths were also facing other problems-.'

The origin of the Huns is shrouded in mystery and discussion of this is best left until later, when looking at their direct attack on the empire (see page 315). In 376 the Romans were only dimly aware of their presence. Wild stories circulated of their savagery and barely human behaviour. They were ugly and misshapen, with shaven heads and beardless faces. Superb horsemen, they could barely walk on their own two feet. Growing no crops they lived on milk and raw meat, which they heated by placing beneath the saddle cloths of their horses.

They are all without any fixed home, without household gods or laws, or stable way of life, and they always wander from place to place, almost like fugitives, with the wagons in which they live... . Like unthinking animals, they have no concept of right and wrong, are deceitful and evasive in speech, not bothered by religion or belief.

All the old stereotypes of barbarism were revived and repeated, but the spread of such stories gives some idea of the fear inspired by these nomads from the Steppes. Once again, we should not think of the Huns as a single, united people. They were divided into many sub-groups and answered to different leaders. The power of a few kings may have been growing at this time, but the Hunnic attacks in the second half of the fourth century should not be seen as a concerted and organised invasion. Instead, there was an increase in scale and frequency of the raids launched against their neighbours.9

The arrival of the Huns added a new factor to the struggles for power within and between the tribes of the region, presenting opportunities as well as a threat. Local chieftains were faced with a choice between opposing Hunnic raiding parties or seeking to ally with Hun leaders to gain the support of their bands. In this way some Gothic chieftains were able to defeat their rivals and expand their own power. Others suffered and were killed, driven from their homes or forced to accept subordination to their enemies. The impact of the appearance of the Huns in the lands around the Black Sea was to make warfare in the region more decisive. The Alans, themselves originally another nomadic people from the Steppes, were the first to feel the brunt and in due course all of their leaders either fled or accepted the overlordship of Hunnic kings. The Goths were next, and the same pattern of resistance and alliance was repeated. At times Huns were hired by both sides in struggles between different Gothic groups. It was not just a case of hopelessly heroic resistance by Gothic kings against the merciless horsemen from the Steppes. Some Goths quickly came to terms with their new aggressive neighbours and fought with them against other Goths.'°

Athanaric fought against the Huns and was beaten, retreating into the mountains just as he had done to escape Valens. For the moment at least he was resolute in his refusal to break his oath and seek sanctuary within the empire. It was another group of Tervingi who approached the Danube and asked to be admitted. Two chieftains, Alavivus and Fritigern, are named in our sources, but it is clear that their power was not absolute - they were simply the two strongest and most influential leaders of the warrior bands with the migrants. It is also wrong to imagine a single great caravan rolling towards the empire. For practical reasons of supply as much as anything else, many distinct parties travelled in the same general direction and only massed together when they reached the crossing point on the Danube. They were a loose group, some fugitives from the Huns or enemies within their own people, and others most likely simply eager to enjoy the more comfortable life within the empire. Service in the Roman army was an attractive prospect to many warriors, and the chieftains in particular could look forward to rewarding careers in imperial service."

We do not know how many people there were altogether. One late and unreliable source claims that there were 200,000, but this is likely to be vastly exaggerated. Ammianus simply says that there were too many for the Roman troops on the frontier to count. One modern estimate suggests some io,ooo warriors, along with four or even five times as many women, children and elderly. This is plausible enough, but still no more than conjecture. It remains perfectly possible that the group was larger or smaller than this. Similarly, the ratio of adult males to the rest is very hard to estimate. Clearly, an entire community fleeing from aggression would have contained a higher proportion of non-combatants than bands seeking military service. Soon after the Tervingi approached the frontier, the Romans became aware of another large party of Goths advancing with similar purpose. These were the Greuthungi - although again they were only one section of the people going by this name."

After a round trip of well over i,ooo miles, the Tervingian ambassadors returned from Antioch and their audience with Valens with the news that he had granted their request. Ammianus tells us that his advisers had easily convinced the emperor that the migrants would prove an asset. They would provide a steady supply of recruits for the army. This would mean that the levy of conscripts from other provinces could be commuted into a payment in gold. Thus the empire would have both soldiers and money. There is no support in the ancient sources for the modern suggestion that the ongoing tension with Persia meant that Valens could not have refused the Tervingi entry even if he wanted to do so. Soon afterwards he rejected a similar appeal made by the Greuthungi. Quite why the two groups were treated differently is unknown. Suggestions have ranged from the inability of the authorities to process so many people to a display of strength to emphasise to the Tervingi that the Romans had not been forced to admit them. Just as likely is the possibility that there were differences in the past relationship of their leaders with Rome."

There was nothing new about settling tribesmen from outside the empire within the provinces. Diocletian and Constantine were amongst the many emperors who had chosen to do this. Previously hostile peoples were transplanted to more productive land, so that they ceased to be a threat and in time provided tax revenue and/or soldiers for the army. Precedents for similar behaviour by the Roman authorities went back a long way, to the development of frontiers further and further away under the Republic. In the first century AD a senatorial governor had proudly recorded that `he brought over more than ioo,ooo of the people who live across the Danube to pay tribute to Rome, along with their wives and families'. As always the number may be exaggerated, but it was clearly a substantial group of people and was included as one of his greatest achievements.14

Yet not all migrants were admitted. Julius Caesar began his Gallic campaigns by refusing to permit a tribe called the Helvetii to move through his province on their way to settle in Gaul. He not only repulsed them when they tried to force their way through, but - claiming that they were plundering Rome's allies - chased after the Helvetii, defeated them in battle and sent them back to their homes. This was an especially robust response from an ambitious general who needed the glory of major victories. Yet it was not that unusual, and there are plenty of other cases where migrating groups were refused entry or driven back by force. The choice was always supposed to lie with the Roman authorities, who would ruthlessly suppress any refusal to accept their decision. In most cases the peoples involved had already been clearly defeated by the Roman army. At other times the submission was more symbolic and a display of Roman might was accompanied by gestures of subservience from the barbarian leaders. In essence, the migrants had first to surrender to Rome. Then they were settled, usually in small groups over a wide area on land that had fallen out of cultivation or was part of an imperial estate. The majority of settlements proved highly successful. The precise legal status of the barbarian colonists varied - descendants of those who had been defeated were one of the few groups not included in Caracalla's grant of citizenship. '

The Tervingi had not been defeated, but since they came as suppliants, Valens' decision to grant their request was neither unprecedented nor unreasonable. The details of the treaty elude us, as do the precise terms on which the migrants were to be settled. One of the conditions seems to have been that the Goths convert to Christianity. The Goths certainly did this, adopting the Arian form favoured by Valens himself. A later source also claims that the tribesmen were to be disarmed, although the contemporary Ammianus does not mention this. It is possible that this was part of the agreement, although even if it was, then the gesture of handing over a few weapons may well have been mainly symbolic. In the event, the Tervingi retained a good number of weapons. It took a considerable time to carry the Goths across the Danube - normally there was little need for so many ferry boats. The naval squadron that patrolled the river assisted, but their craft were not especially numerous and certainly not designed to carry large numbers of people or bulky wagons. Many of the Goths crossed in rafts built for the purpose, but a few are said to have tried swimming and drowned in the process.'6

The Road to Disaster

There were well-established mechanisms for accepting groups of barbarians into the empire and settling them within the provinces. Yet from the very beginning things did not run smoothly for the Tervingi. Possibly there was negligence on the part of Roman officials over the question of disarming the tribe. Certainly there was sloth, incompetence and corruption in almost every other aspect of the affair. Ammianus blamed the two army officers in command on the spot - Lupicinus, the comes in charge of the comitatenses in Thrace, and the dux Maximus, who controlled the limitanei. The most basic problem was one of food. The Tervingi may well have used many of their own supplies while they waited for the response from Valens and then during the long process of crossing the river. The Romans were supposed to feed them, but what the Goths were given proved barely adequate. The supplies may simply not have been available. The Tervingi were equivalent in numbers to a very large Roman expeditionary army and it usually took a couple of years to mass the grain and other supplies needed by such a force. The officials on the Danube had had no more than a few months to prepare. Even so, the state received a considerable amount of taxation in the form of agricultural produce and was supposed to store the surplus in granaries within walled cities and army bases ready for use by troops, the court or officials. If the resources were not there to meet the needs of the Tervingi, then the decision to admit them would seem extremely unwise. Perhaps the supplies existed, but had not been moved to the right place. It is hard to believe the suggestion that the emperor ordered his officials to restrict the amounts given to the Goths in order to keep them dependent, since this was bound to be a very risky strategy. The officials on the spot may have decided on such a dangerous course of action. Certainly, they chose to profit from it. Ammianus tells us that once Lupicinus had prised much of the barbarians' wealth from them in exchange for black market food, he began an even more sinister trade. The Goths were desperate enough to sell their children for paltry amounts of dog meat. The going rate was one child for one dog - Lupicinus' men were organised enough to have gathered up stray dogs from a wide area.'7

Slowly the Tervingi were moved to the city of Marcianopolis where Lupicinus seems to have had his headquarters. They were not admitted to the city or its market, but made to camp some distance outside. To supervise the march most of the Roman troops were drawn away from the frontier leaving it seriously depleted. At some point the Greuthungi, who had been refused entry, crossed into the empire - the concentration of the Roman patrol boats to assist the Tervingi meant that large stretches of the Danube were not being watched. The Roman authorities were rapidly losing control of the situation. Either there were insufficient troops or they were very poorly deployed. The situation at Marcianopolis was already tense when Lupicinus invited the Tervingian leaders to a banquet. Such meetings were a regular feature of Roman frontier diplomacy and also, as we have seen, opportunities for treachery. We do not know whether or not Lupicinus planned to imprison or kill the chieftains. Given the considerable time lag, it is unlikely that he was acting under explicit orders from the emperor.

The trouble began outside the city when an argument between soldiers, townsfolk and the Goths escalated into a small battle. A party of troops was routed. At this point Lupicinus - Ammianus notes that it was late in the evening and he was already more than half drunk - ordered the execution of the chieftains' attendants and the Gothic chieftains were also arrested. When news of this spread outside the city to the Gothic encampments, there was an uproar and more and more warriors arrived to join those who had been involved in the fighting. Fritigern managed to talk his way out, convincing Lupicinus that only he could calm the angry mood of his countrymen. He was released. Nothing more is ever heard of Alavivus.'8

Lupicinus gathered all of the troops he could and decided to march against the Tervingian camp some 9 miles away from the city. The Goths were waiting and routed the column. Lupicinus himself escaped, allegedly because he was one of the first to gallop for safety. A war had started and quickly began to spread. A group of Goths who had been accepted some time earlier into the empire were waiting at Adrianople to move to the east, presumably to serve with the army. There had already been some friction with a local magistrate, who now raised a force from the city, including the workers from a state arms' factory. These Goths cut this hastily armed militia to pieces, plundering them of their newly made weapons before joining up with Fritigern. Together the combined army tried to besiege Adrianople, but failed dismally. As they withdrew, Fritigern sullenly reminded them that he `kept peace with walls'!

The Goths lacked skill at siege craft, but far more importantly they lacked the ability to stay in one place for long enough to capture a defended and fortified town. They were now free to plunder the countryside, burning villages and villas, gathering animals for meat and as much grain as they could find. Yet the biggest stores of food were always kept in walled towns and the Goths could not capture such places. So many hungry mouths quickly consumed whatever supplies were available in any stretch of countryside, forcing them to keep moving. The Goths' numbers had grown considerably, as Fritigern's band was joined by the Greuthungi and groups like the one from Adrianople. More came as individuals. Some were Goths recently sold into slavery in exchange for food, or captured years earlier by slave traders or in imperial campaigns. As news spread of the rich plunder for the taking in Thrace, other warbands crossed the Danube to join them. Whatever troops the Romans still had stationed on the frontier were clearly incapable of preventing this. In 377 Fritigern even hired some bands of Huns and Alans with the promise of a generous share of the spoils.Z°

The Goths rapidly became more numerous and powerful, especially since the recent arrivals consisted mainly of eager warriors, not migrants with their families. Many were now equipped with good-quality Roman weaponry and most likely more of them wore mail or other body armour and helmets than was normal for a tribal army. The pressure of the situation, surviving in enemy territory month after month, cemented the authority of their leaders and their ability to work together. None of this altered the basic fact that they were fighting a war they could not win. The main groups of Tervingi and Greuthungi were migrants without homes. Unlike the raiders who had joined them, they could not retire back across the Danube. However much their numbers grew, their military might and resources were dwarfed by those of the empire. The best that they could hope for was to be granted lands by the Roman authorities. The worst outcomes were annihilation, slavery or to make peace with Rome and again suffer mistreatment. Fritigern and the other leaders may have understood this and realised that their best hope was to negotiate from a position of strength. They had no clear military objective. This, combined with the never-ending problem of supply, shaped the apparently purposeless meanderings of the Goths during the next years. They did not remain as one concentrated army, but continually split up into many small parties to forage and plunder. When threatened by Roman forces the different groups would try to re-form as quickly as they could.

In the long run the Romans could not lose this war, but that did not mean that it was easy for them to win it. Lupicinus had lost heavily in his rash and unprepared attack outside Marcianopolis. Other troops were dispersed as garrisons, dotted around the walled cities of the region to defend these against the enemy. At first, all the local forces were capable of doing was holding the vital mountain passes, which kept the Goths bottled up in just one part of the Thracian plain. In 377 a field army was put together from a mixture of units sent by Gratian from the west along with eastern troops. Several successes were scored when isolated groups of Goths were attacked and killed or captured. Rapid movement and surprise attacks were once again the most effective Roman tactics. However, the Goths were sometimes capable of doing the same things, and several Roman units were cut to pieces outside the city of Dibaltum - properly, Deultum - modern Debelt in Bulgaria."

Much of the time the Romans operated in small detachments, harassing the scattered enemy. Only once did they concentrate to attack a substantial number of Goths - we are not sure which particular group this was - who had laagered their wagons in a great circle near the town of Ad Salices. As the Romans gathered, the barbarians also had time to call in many of their dispersed bands of raiders. When the Romans finally launched an attack, there was very heavy fighting around the wagon line. The Roman left wing was broken by a Gothic charge and the situation was only stabilised by the units in reserve. The battle ended in a costly stalemate, but it was the Romans who withdrew some days later. After this they returned to their harassing strategy.

Valens made peace with the Persians in 377 and was back in Constantinople the following year to deal with the problem of the Goths. He had gathered a field force and the plan was to combine with an army brought east by Gratian before confronting the main strength of the enemy. Unfortunately the western army was delayed. An Alamannic soldier from Gratian's guards went on leave and happened to mention the planned move eastwards to his kinsmen. The latter decided to take advantage of the absence of the bulk of the Roman forces by raiding the provinces. Gratian fought a short campaign to punish the tribe involved and only after this was complete was he able to begin the march to join his uncle."

Adrianople and After

It was now late summer and Valens had already decided to advance on his own. At the beginning of August he was at Adrianople, closing with a substantial group of Goths led by Fritigern. His patrols reported that the enemy numbered about io,ooo. This proved to be a serious underestimate, but Ammianus fails to tell us how large the Gothic army actually was. At the same time messengers arrived telling him that Gratian was only a few days' march away. Some of his senior officers advised caution, arguing that it was only prudent to wait for Gratian and so make their victory absolutely certain. Others suggested that there was an opportunity for a quick victory over this one section of the enemy. Valens is said to have wanted to win glory for himself, so that he would not be overshadowed by the recent achievements of his nephew and the minor victories won by his own army.

Valens' confidence was boosted when a Christian clergyman acting as Fritigern's envoy arrived. Publicly he asked that the Goths be granted Thrace to settle in, but he also carried a private message in which Fritigern assured the emperor of his goodwill. He asked Valens to mount an impressive display of force so that it would be easier for him to calm his fellow tribesmen and persuade them to accept peace. No answer was given to the envoy, but this seemed a clear sign of the Goths' nervousness. On 9 August 378 Valens marched his army out of Adrianople and moved against the Gothic camp, which once again took the form of a great circle of wagons. The Roman army did not begin to arrive until early afternoon and the men were tired from marching under the hot sun.

The column began to deploy by wheeling to the right, so that the units at the head of the column would form the right of the battle line. Ahead of them were bands of Goths in front of the wagon laager, chanting their battle cries. Others lit bush fires, the wind carrying the heat and the smoke into the Roman lines. As well as adding to their discomfort, it reduced visibility. This was important as Fritigern was expecting to be joined by a large group of Greuthungi, including many cavalry."

The Gothic leader may have been playing for time to allow these extra forces to arrive when he once again asked for a parley. Alternatively, he may genuinely have understood that he had absolutely nothing to gain and a huge amount to lose from fighting a battle against the emperor. Valens refused to speak to the first group of envoys because they were too obscure, but the Romans responded to a second approach asking them to send over a senior officer as a hostage. Valens may also have simply been prevaricating, wanting to allow the rest of his army to arrive and form a battle line. On the other hand, he would certainly have been content with a bloodless victory, where the enemy submitted in response to a Roman display of force. Whatever the rival leaders' real intentions, it was not to be.

Fighting began when the two units of Roman cavalry on the far right flank attacked without orders. There was always a danger of such things when two rival armies stood for hours on end facing each other across a short distance. They were quickly driven back, but it seems to have produced a general attack all along the Roman line. On the left flank the units had scarcely arrived and were not properly formed up when they joined the attack. The cavalry units that should have protected the flank of the infantry were not in place, leaving them very exposed. They were completely unprepared to meet the sudden attack of the Greuthungi, along with the Gothic cavalry and a band of Alans. The Roman attack lost momentum, but there was still a long period of savage combat before the battle was over. Some of the infantry were surrounded, the units too confused and densely crowded to form a proper fighting line, but they continued to resist for some time. In the previous year the Romans had been able to deal with the collapse of one flank by sending in reserves. This time the army was not properly formed or under control. A regiment that should have been stationed in reserve could not be found when it was needed, most probably because it had already become drawn into the attack.

Tactically, the Roman army was in a hopeless situation, and in the end the soldiers broke and fled. The Goths pursued with enthusiasm and, as was usual in the battles of this period, the losing side suffered badly. Around two-thirds of the soldiers were killed, along with no fewer than thirty-five tribunes - some commanding regiments and others unattached but with the imperial staff - and two more senior officers. Valens was also among the dead and, just like Decius over a century before, his body was never found. One story circulated that he and his attendants had holed up in a farmhouse. When the Goths were unable to break in, they set fire to the building, killing all save one attendant, who told them how close they had come to capturing an emperor.14

We do not know how big Valens' army was at Adrianople and therefore cannot calculate the total loss. Most modern estimates put both the Roman and Gothic armies around the 15,000 mark, so that some 10,000 Roman soldiers are thought to have died. Once again, the figures are plausible but entirely conjectural. We do not know how many of the tribunes who died commanded units - but then, since we do not know how big such regiments were, let alone whether they were present in their entirety or merely as detachments, this would not tell us anything definite. Nor do we know how many tribunes commanding units survived the battle. Clearly, Valens felt confident that his army could deal with a force of 10,000 Goths - presumably all warriors, although Ammianus is not specific. Once again, we are left to guess at whether this would mean having parity or a numerical advantage. Julian was supposed to have beaten an army of Alamanni almost three times larger than his own force at Strasbourg.15

Adrianople was a major disaster. Whatever the precise figure, the critical point was that the greater part of the soldiers immediately available to the eastern emperor for active campaigning had been killed. Ammianus compared the defeat to Cannae in 216 BC, where Hannibal had slaughtered some 50,000 Roman soldiers and captured 20,000 more. Adrianople was much smaller in scale, but it was the worst defeat at the hands of a foreign enemy since the third century. Luck played a part, but Valens had been overconfident in closing with the enemy, then indecisive in considering negotiation at the last minute, and had utterly failed to control the attack itself.

The Goths won a great victory, but in the long run it did nothing to improve their situation - they needed to negotiate with an emperor, not to kill one. They followed up by assaulting Adrianople, hoping to capture the supplies there as well as the imperial treasury. Enough troops had been left behind by Valens to repulse every onslaught and an attempt by some turncoats in the army to betray the city was also thwarted. After a while Fritigern and his warriors moved on to threaten Constantinople itself. He had been joined by more warbands, including groups of Alans and Huns. Even so, his army was overawed by the massive size of the city, already bigger than anywhere else save Rome, Antioch and Alexandria. As disturbing were the aggressive sallies of a unit of Saracen cavalry. Ammianus says that one of these riders rode half naked into battle and after killing a Goth by slitting his throat, seemed to drink his blood. Suitably impressed, Fritigern and his men returned to their practice of keeping peace by the use of walls. The Goths had co-operated to fight the battle - although it is more than probable that a number of bands were not present - but remained divided into many separate groups under different chiefs. This and the familiar problems of supply meant that they soon broke up into many fragments, marauding through the region in search of food and plunder.26

Fear spread rapidly as the news of the disaster at Adrianople reached the other provinces. In a particularly ruthless series of massacres, groups of Goths throughout the empire were disarmed and slaughtered by the authorities in case they, too, chose to rebel. For a while Gratian was effectively emperor of the entire world, since his younger brother was still too young to assert his own power. Within a few months - probably early in 379 - he acknowledged a recently appointed Master of Soldiers named Theodosius as Augustus of the eastern provinces. Theodosius' father and namesake came from Spain and had enjoyed a distinguished military career under Valentinian, winning victories in Britain and Africa. He was the man who exposed the misdeeds of Romanus. However, in 375 the older Theodosius had been condemned by the emperor and executed. This may have been posthumous revenge from associates of the discredited men or simply the result of the habitual back-biting of the court and paranoia of the emperor. The son was dismissed from service, but may well have been recalled by the time of Adrianople. Soon afterwards he was given a command and won a minor victory on the Danube. He was probably backed by significant figures at the eastern court. Whether or not he was Gratian's choice, the two men did show that they could work together.27

Almost the first task was to rebuild an army. Theodosius displayed little of his father's talent as a soldier, but was certainly a great organiser. Men were found from a whole range of sources and a series of strict laws passed against draft dodging, self-mutilation to avoid service and desertion. The army grew in size, but many of the new recruits were not yet properly trained and the confidence of the rest was at a low ebb. The Romans continued their previous strategy of harassing the individual groups of Goths, blocking their movements and depriving them of food. One attempt at more direct attack ended in failure when a column led by Theodosius himself was badly mauled.Z"

Over the next few years the Goths were gradually worn down by ambushes and surprise attacks. Individual groups surrendered and Gratian settled some in Italy. The details of these campaigns elude us - sadly Ammianus' account stopped a few months after Adrianople and there is no comparable narrative history until the sixth century. In the end all of the Goths who remained within the empire capitulated in 382. Fritigern is not mentioned, and it may well be that he was already dead or killed as part of the settlement. Ultimately, the Goths got much of what they originally asked for - they were settled on land in Thrace or in the adjacent border areas along the Danube. The precise details of the treaty are hotly debated and need not concern us here. It is safe to say that their fierce resistance meant that the conditions under which they were settled were far more generous than was usual. Their own chieftains seem to have retained considerable authority and may, in practice if not in theory, have enjoyed a degree of local autonomy.''

There is nothing surprising about the eventual defeat of the Goths, for they simply could not compete against the resources and organisation of the empire. What is startling is that it took six years to force their surrender, and that even then the Roman victory was not as complete as they would normally have expected. This is a serious problem for those who emphasise the strength and efficiency of the empire in the later fourth century. The allegedly enlarged and highly efficient army seems in practice to have struggled to find enough men to deal with the migrating tribes. Yet this was scarcely a new problem or one on a massive scale. Of the major battles of these campaigns the Romans were clearly beaten three times - disastrously, in the case of Adrianople, but badly enough even if on a smaller scale under Lupicinus and Theodosius - and at best managed a hard-fought draw at Ad Sauces. This is scarcely an impressive record and again confirms the impression that in this period the army operated best on a small scale, using surprise, speed and ambush rather than direct force. The empire still commanded huge resources, but it does seem to have been difficult to apply these to any problem. There was clearly a shortage of readily available and willing military manpower - the Tervingi were admitted specifically to help meet this need. In these six years of warfare the Roman Empire won not because it was efficient, but simply because it was big. In 386 another group of Goths attempting to cross the Danube were efficiently blocked by a Roman army. It is more than probable that the best course of action in 376 would have been to refuse entry, since the authorities proved so incapable of effectively processing the migrants. Even if some Goths had broken through by force, they are unlikely to have caused as much damage and disorder .31

Just a year after the treaty with the Goths a familiar problem reared its head once more. The local commander Magnus Maximus was proclaimed emperor by the troops in Britain. He was another Spaniard, probably known to - perhaps even a relation of - Theodosius. Gratian refused to recognise the usurper and massed an army to confront him when Maximus crossed into Gaul. There was some skirmishing near Paris, but after several days Gratian's army went over en masse to his opponent. He fled, but was caught at Lugdunum (modern Lyons) and executed. It was clearly a well-orchestrated coup and Maximus had secured the backing of many senior officers and court officials. Some senior members of the court were executed, but the majority switched sides. Less clear is why Gratian had lost their support. His military record was quite good, but he was accused of granting excessive favour to one regiment of Alan cavalry and starting to indulge too much in his pleasures rather than working.31

Maximus controlled the European provinces north of the Alps and clearly hoped for recognition as a colleague by Theodosius. He invited the twelve-year-old Valentinian II to move from Milan and join him at his court in Trier, so that they could rule `as father and son'. Skilfully created delays in the negotiations gave time for troops loyal to the boy emperor to secure the Alpine passes. Maximus still hoped for reconciliation and made no attempt to use force at this stage. For the moment Theodosius recognised the usurper and his name appeared in official documents. He also elevated his son Arcadius - who was no more than five or six years' old - to the rank of Augustus. However, a few years later Maximus launched a sudden attack on Italy, and by 387 he was in Milan and in full control of Valentinian's territories. The latter, along with most of his court, escaped safely to Theodosius.j2

The extension of Maximus' ambitions signalled a permanent break with the eastern emperor. Valentinian's mother Justina was a formidable woman and had clearly had considerable sway over her son's decisions. Now she is said to have exploited the beauty of the boy's sister to fascinate Theodosius. The two were married soon afterwards and allegedly the bride price was a promise to recover the lost territory from Maximus. Whatever the precise reasons, in the summer of 388 a fast-moving expeditionary force caught the western usurper at Aquileia. Maximus was stripped of the imperial robes and beheaded. There was some more fighting against forces that remained loyal to his family before the west was fully recovered .31

Formally the empire now had three Augusti - Theodosius, Valentinian and Arcadius - although it was abundantly clear that real power rested with Theodosius himself. Valentinian remained little more than a cipher. After his mother died actual control rested with some senior officers appointed by Theodosius. The most important of these was Arbogast. Like many senior officers in the army he was of barbarian - in his case Frankish - descent. As time went on, he became more and more contemptuous of the Augustus he was supposed to serve. He assumed the rank of Master of Soldiers without bothering to consult Valentinian. When the emperor dismissed him, Arbogast calmly told him to his face that he did not have the power to do this. The twenty-one-year-old Valentinian II was a pathetic figure and on 15 May 392 he was found dead in his bedroom. It may have been suicide.

Arbogast obviously felt his background barred him from becoming emperor himself, and so named a certain Flavius Eugenius as Augustus. Once a teacher of grammar and rhetoric, his prized literary education had won him a post at Valentinian's court. From the beginning he was obviously a figurehead. At least nominally a Christian - unlike Arbogast who was openly pagan - Eugenius began to cultivate the support of pagans. This probably grew stronger after Theodosius named his younger son Honorius as Augustus at the beginning of 393, making clear that he would not accept the usurper. The civil war that followed was once again decided near Aquileia. In September 394 the two armies met beside the River Frigidus. After very heavy and costly fighting lasting for two days, Theodosius' army was triumphant. Eugenius was captured when his camp was stormed and promptly executed. Arbogast committed suicide before he was taken.34

A very sizeable contingent in Theodosius' army was formed by Gothic warriors raised from the peoples settled in 382. On the first day of fighting they bore both the brunt of the fighting and the losses - later some Romans would claim that this made it a double victory for the empire. At Adrianople the Goths had killed Valens, although probably more by accident than design. They never had any prospect of inflicting a permanent defeat on the empire, and Theodosius had been able to wear them down over the next few years. Yet ultimately they represented a valuable resource of military manpower. This was why they had been admitted in the first place, and why the Romans would probably not have chosen to destroy them even if they had been able to do so. Just over a decade later their warriors greatly strengthened Theodosius' army and may even have given him a decisive advantage over Eugenius. Barbarian incursions were a nuisance, but it was always internal enemies who threatened an emperor's rule and very life.35

If you find an error please notify us in the comments. Thank you!