5

Barbarians

`... wars are decided by courage rather than numbers. We have no mean force. Two thousand of us have gathered in all, and we have this deserted spot as a base from which to damage the enemy by attacking him in small groups and ambushing him. ... Let our watchword in battle be our children and all that is dearest to us, and to save these let us set out together for the conflict, calling on the gods who watch over and aid us.' - The historian Dexippus'version ofhis rallying speech to the Athenians after the fall of their city to raiders in 267/8.'

In June 251 Emperor Decius attacked a group of barbarians who had crossed the Danube to plunder the Roman provinces. They were led by a chieftain named Cniva, and the emperor had been hunting them and other similar bands for more than a year. At the beginning of his reign he took the name Trajan, in memory of the conqueror of Dacia. Doubtless the idea was to promise new victories on the Danube, but so far there had been few enough. Trajan had been an experienced commander, with a powerful and disciplined army drawn from all over the empire. Both army and empire had been free of civil war for a generation. Decius was less fortunate, a good deal less talented and led forces far weaker than his namesake. Cniva had already beaten him in the previous summer. Afterwards, when Philippopolis in Thrace was besieged, Decius was either unable or unwilling to march to its relief. Abandoned to his fate, the commander of the city betrayed it to the enemy and promptly declared himself emperor. Other usurpers emerged at this time on the Rhine and in Rome itself. All were swiftly killed by loyal officers, but Decius knew that his prestige was at a low ebb. He needed a big victory.

Cniva was on his way home with the spoils of his raid when the emperor caught up with him near Abrittus - an insignificant settlement today near the border between Romania and Bulgaria. Most of the details of what then happened are now lost. There may have been one battle or a series of smaller skirmishes. Initial success turned to disaster when the Romans were ambushed in boggy terrain. Decius was killed, as was his son and co-emperor. One source claims that their horses became stuck in the mud and they were then killed by missiles.

We do not know how many soldiers fell with Decius and his son, although the accounts of the campaign do not suggest that either army was exceptionally big, probably numbering thousands rather than tens of thousands. Roman armies had been beaten by barbarian tribes in the past. Famously, Augustus had lost three legions and their auxiliaries - a force of perhaps 15,000-20,000 men - in an ambush in the Teutoberg Forest in AD 9. The losses were probably lighter at Abrittus, but it was the first time an emperor had been killed by a foreign enemy - and was made worse because his body was never found for proper burial. Ironically enough, the disaster occurred not far from Tropaeum Traiani, where Trajan had built a huge drum-shaped monument to commemorate his Dacian victory, as well as a memorial to honour the war dead.

Cniva's raid was just one of many to burst into the settled provinces of the empire during the middle decades of the third century. Regions that had been peaceful for generations - Gaul, Italy, Spain, Greece, Asia Minor and North Africa - fell prey to marauding bands from beyond the frontiers. Most of the raiders spoke a Germanic language, but the Germans were divided into many separate tribes and some wider groupings. Cniva was from a Gothic tribe, a people who had only fairly recently come into close contact with the Romans. By the end of the third century the Goths, along with other apparently new and powerful peoples like the Franks and Alamanni, posed serious threats to the frontiers on the Rhine and Danube. The balance of power seemed to have shifted profoundly.2

The Germans

Julius Caesar claimed that the Rhine marked the boundary between the Germans and Gauls, conveniently providing him with a `natural' stopping point for his conquests. The restless, aggressive and very numerous Germans were always trying to push westwards into the rich lands of Gaul and beyond. Caesar portrayed the Germanic tribes as semi-nomadic and posing a threat to Rome's allies and even to Italy itself. He dealt with this `threat' ruthlessly, but it is clear from his own account that the situation was a good deal more complicated than this. Some German tribes were already well established west of the Rhine and were incorporated into the new Roman provinces without difficulty. Caesar also recruited many German mercenaries from east of the Rhine to fight alongside his legions. From the very beginning, the Germans were a valuable source of military manpower as well as a threat.

Augustus tried to annex the lands from the Rhine to the Elbe, but his plan for a great German province died with the legionaries massacred in the Teutoberg Forest. Although Roman armies marched out to punish the tribes for this defeat, the project was never revived and the frontier settled on the Rhine. Neither this nor the Danube was a strict limit to Roman territory, for a strong military presence was maintained on the far banks. Near the end of the first century, the gap between the two rivers was linked by a fortified line, bringing a substantial area of territory, known as the Agri Decumates, under direct Roman rule. A little later, Trajan conquered Dacia, but this was the last major change to the frontier line in Europe for a century and a half. Along it was stationed over half of the entire Roman army - almost two-thirds if the garrison of Britain was included.'

To the east were tribal peoples, not all of them Germanic. On the Danube there were the Carpi, relatives of the Dacians who had escaped the conquest of that kingdom, and Sarmatians, originally nomads from the steppes who had moved to the Hungarian Plain. Another tribe, the Bastarnae, may or may not have been Germanic. Yet the overwhelming bulk of peoples who lived next to the frontier were Germans as far as Roman observers were concerned. However, it is doubtful that they saw themselves this way. Some tribes had relations of kinship and shared cults with their neighbours, but there seems to have been no real sense of `German-ness'. Although their languages had similar roots, it is quite possible that people would have had trouble communicating with anyone from a distant tribe. The important ties were to a tribe, and perhaps even more to smaller groups of clan and family.4

Caesar describes the Germans as essentially pastoralists who did not till the soil. Even in his day this was a huge exaggeration. In some areas it does seem to have been common for villages to be simple and shortlived, the people moving on after a few years to a fresh site when they had exhausted the closest fields. Yet the overall picture from the Roman period is one of continuity and stability, and the creation of the static frontier line may well have encouraged this. Several excavated villages were occupied for three or four centuries. Most were small, but some consisted of a dozen or more - in one case thirty - sturdily built rectangular timber houses. The population of such a community most likely numbered a few hundred people. There were no towns comparable to those that had grown up in Gaul and parts of Germany in the Late Iron Age. Instead the picture is of many villages and isolated farms.5

Some tribes had kings, but in the first century it was noted that war leaders were elected from a broader aristocracy. A chieftain's power was measured by the number of warriors he maintained in his personal following - his colnitatus, to use the Latin word. These men were pledged to fight beside him in battle, and in return he feasted them and rewarded their valour with weapons and gold. Centuries later similar attitudes would be celebrated in poems like Beowulf. One powerful fourth-century king had 200 warriors in his band and similar numbers are suggested by spectacular collections of weapons excavated in Scandinavia. These were spoils taken from defeated enemies and then offered to the gods by the victors, who threw them in a sacred lake. At Illerup in Denmark weapons - spear and javelin heads, shield bosses and some swords - sufficient to equip some 300-350 men were dedicated sometime around the year 200. A century later enough weaponry for around 200 warriors was thrown into another Danish lake near Ejsbol.

Finds like these confirm the impression given by the literary sources about the equipment used by Germanic warriors. Only a few were mounted and armour was rare. The majority of warriors employed a spear, a javelin for throwing and carried a shield for protection. A substantial minority also had swords - by the third century almost invariably Roman-made. More Roman swords have now been found outside the empire than within it and it is striking just how many turned up in the possession of tribal armies in Scandinavia, so far away from the frontier. Some had been captured in raids, but more were probably acquired through trade - much of it illicit - and perhaps in gifts to loyal chieftains.'

The semi-professional warriors who made up chieftains' warbands were not especially numerous. At times, larger armies could be formed when they were joined by all those free tribesmen able to equip themselves, but such forces could not stay in the field for long. Most inter-tribal warfare was small scale and mainly consisted of raiding. Occasionally the stakes and scale became higher. At the end of the first century Tacitus wrote gleefully of barbarian tribes wiping each other out without any Roman participation. With just a couple of hundred warriors in his band, a charismatic and successful chieftain could rise to dominate his tribe, and sometimes its neighbours as well, but his power was always precarious. Arminius, the man who had destroyed the legions in the Teutoberg Forest and resisted Roman attacks in the following years, came to lead a confederation including his own Cherusci and a number of other tribes. When the Roman threat receded he was murdered by his own chieftains because they feared he was aiming at permanent rule. Periodically, a chieftain would win great power amongst the tribes, but his status remained personal and tended to die with him. No one managed to establish more permanent authority that could be passed on to an heir.7

It was much easier for the Romans to deal with a few kings or chieftains than large numbers of individualistic tribesmen. From the beginning leaders perceived as friendly to Rome were supported with subsidies and even occasionally direct military aid. Many finds of ornate gold and silver vessels from beyond the frontiers are most likely to represent prestigious gifts given to such men. They were not the only ones to benefit from the arrival of the Romans. The new frontiers were very densely populated with crowded army bases and the settlements that inevitably grew up around them, as well as larger towns and cities. Tribesmen farming the lands beyond the frontier found a ready market for any surplus produce.'

Trade flourished. Grain and animals from east of the two rivers helped to feed the army and the civilians living on the frontiers. In turn the tribesmen had ready access to many luxury goods only available in quantity from the empire. It was a serious restriction when the Roman authorities barred any group from coming to market in the communities along the frontier and a great privilege to be readily admitted. There was also an impact on barbarian communities further away from the empire. We read of Roman merchants going to the Baltic to buy amber, which was prized as jewellery. Other trades - for instance, in fine furs - seem likely, but are harder to prove. A village discovered on one of the Danish islands shows how a community grew rich through trade, acting as a staging post for markets further afield in Scandinavia. There was also the slave trade, for once the Romans stopped fighting regular wars of expansion there were fewer war captives to be sold. Weapons become more common as grave goods for peoples in eastern and central Europe from around the time the Roman frontier was established. Probably this meant far more frequent predatory warfare, as chieftains raided their neighbours for slaves to sell to Roman merchants in exchange for luxuries.'

Chieftains profited most from the slave trade. This was probably also true of agriculture, simply because they were able to amass greater quantities of surplus grain than a single farmer. Already established leaders were also far more likely to attract Roman subsidies. Therefore, although many prospered through the long-term presence of the frontier zone, the impact of Rome also increased divisions within society. In excavated villages all of the houses tend to be of much the same size at the beginning of the Roman period. Later on, it was common for one to be substantially larger and perhaps also fenced off from the rest. Communities prospered through supplying the needs of the densely populated Roman frontier, but it is clear that some individuals benefited far more than others.'°

Peaceful trade was the most common form of contact between Romans, both civilians and soldiers, living on the frontier and the peoples living beyond. Violent encounters were rarer, but that does not mean that they were less important. Raiding was endemic in most of the societies of Iron Age Europe. It was seen as entirely natural whenever the opportunity occurred and a neighbour seemed vulnerable. For chieftains, successful raids brought glory and the plunder with which to reward their warriors. Caesar noted that German tribes took pride in keeping a wide strip of depopulated land around their territory. This showed that they were a warlike people and so acted as a warning to potential attackers."

The arrival of Rome may well have increased the frequency and perhaps the scale of warfare beyond the frontiers. Certainly, from the middle of the first century BC, weapons appear far more frequently as grave goods in Germanic burials. The new economic conditions meant that previously very rare items such as swords were now available in greater quantities. The slave trade encouraged raiding. Even more importantly, Roman subsidies allowed the chosen chieftains to support larger bands of warriors. This rise in their status and power was often resisted by rivals within the tribes. Competition for power amongst leaders was given a new edge and intensity. Wealth was not enough to guarantee long-term success. Some of the rulers Rome supported were killed by rivals and others fled across the frontier to a life of comfortable exile."

Many tribesmen chose to join the Roman army, presumably viewing it in much the same way as joining the band of a chieftain from another tribe. Some chieftains also took service with Rome, bringing with them the warriors of their household. Raiding the Roman provinces was also a tempting prospect. Although it was more dangerous than attacking another tribe, there was the possibility of far more plunder and glory. Most raids were probably small scale, but if these proved successful then they invited bigger attacks. An inscription from Commodus' reign records the construction of small outpost forts along the Danube to stop `secret crossings of the river by bandits'." The pattern had been the same on many of Rome's frontiers since the Republic - when the provinces were seen as vulnerable they were liable to be attacked. The size and apparent readiness of the frontier garrison acted as deterrents, but if Roman strength was seen to be an illusion, then it took hard campaigning to restore Rome's position.14

It was difficult to catch every band of swift-moving raiders - although easier when they withdrew burdened down with booty - and often the Roman response would be a punitive expedition against those held responsible. Villages were burned, crops destroyed, herds driven off and the people either massacred or enslaved. The aim was short term, to instil fear, but such ruthlessness also sowed the seeds of future hatred. Diplomacy aimed to keep peace more permanently, and tribal leaders were threatened or bribed to refrain from hostilities. The communities closest to the frontiers were generally more inclined to be peaceful, since to them the Romans were a valuable market. As importantly, they were also easily within reach of Roman retribution. Leaders and peoples further afield were harder to control. Maintaining Roman dominance on the frontiers was an on-going task, influenced by the shifting politics of the tribes, as well as events in other parts of the empire.

Frontiers in Crisis

In the middle decades of the third century the frontier defences on the Rhine and Danube proved utterly inadequate as successive bands of raiders broke through into the unprotected provinces beyond. Almost every scholar sees this as a sign that the threat from outside had become greater. Most connect this with the appearance of new confederations of tribes, seen as far more dangerous than the Germanic peoples who had lived next to the frontier in the first century. Opinion has divided between those who saw the confederations as new arrivals and others who believed that they evolved from the already existing tribal groups. These days, archaeologists are far less inclined to resort to migrations to explain cultural change, so that most accept the latter view. Even so, the evidence does suggest that the Goths moved from the Baltic coast to the region of the Black Sea and southern Danube between the first century and the start of the third. They were not a single united people, but a loose grouping of distinct tribes speaking a related language and with many cultural similarities. The same was also true of the Franks, who appear on the Rhineland, and the Alamanni, who emerge to the south of them. Both Franks and Alamanni were certainly important groups by the end of the third century, but it is much harder to say when they first appeared."

The tough wars fought by Marcus Aurelius against the Marcomanni and Quadi are generally portrayed as the first warning signs of this shift in German society. The threat from the barbarians was now greater and it revealed fundamental weaknesses in the defences of the Roman frontiers. The army was dispersed around the perimeter of the empire, so that once an enemy broke through there was no central reserve to cope with it. Furthermore, the difficulties encountered in dealing with barbarian attacks on the Danube so soon after the invasion of Parthia are seen as indications that the Romans had great difficulty fighting two major wars in quick succession.

All of this is dubious. The Marcomanni had been seen as a major threat by Augustus when they and neighbouring tribes had united under a strong king. It may well be that one or more similarly charismatic leaders had appeared amongst them again. A similar pattern is observable with the Dacians, who were perceived as a great threat in Julius Caesar's time, but then disappear until the later first century when another strong king emerged. It is true that during the Marcomannic Wars one raiding army got as far as Italy, but this was never repeated. A rather more important factor in the weakness of Rome's defences at this time was not the slow return of troops from the eastern war, but rather the plague that came with them. The impact of this on the army was not simply a question of the men who died, although sources do suggest that there were very many of these in the crowded barracks. As important were the many who must have fallen sick and the extreme difficulty of carrying on normal training in the midst of an epidemic. Some attacks penetrated the frontiers and, as usual, their success encouraged other chieftains to emulate them. The Roman army was in a poor state and struggled to deal with the problem. Yet in time it did so, although at the cost of a great effort and considerable resources. No territory was lost, and there was even talk of creating new provinces.''

There is no record of major fighting along this frontier for well over a generation after the end of these wars. Caracalla spent some time on the Rhine, and Alexander was there when he was murdered. Maximinus spent much of his reign campaigning there and on the Danubian frontier, but also recruited large numbers of German warriors to strengthen his army when he marched on Italy in 238. Soon afterwards Goths and Carpi launched raids across the Danube. For a while the former were bribed to keep the peace, but this payment was stopped either under Gordian or Philip. Predictably, this provoked a renewed burst of raiding from 243 onwards. Philip had to go in person to the region to restore the situation. In 248 the Quadi and Sarmatian Iazyges - familiar names from Marcus Aurelius' campaigns - attacked Pannonia, and their success encouraged a renewed onslaught from the Goths. Decius was sent to deal with the problem and instead made himself emperor. He soon had to return to the Danube, fighting the campaigns that eventually led to his death at Abrittus.''

The new emperor Gallus bought the Goths off, promising them an annual subsidy and allowing them to withdraw, taking their plunder and captives with them. He was far more worried about internal rivals and hastened to Italy. In 253 the man he left as governor of Moesia, Aemilianus, seems to have attacked some Goths and won a victory. The success prompted him to lead off another great chunk of the frontier army in a bid for the throne. Both Gallus and Aemilianus were dead within a matter of months, but the weakness of the frontier defences prompted a new surge of raiding. Cniva's band of Goths were again involved and may have been one of the groups that reached as far as Macedonia."

Around the same time a new threat emerged from the Black Sea. Several groups, including a people called the Borani and several Gothic tribes, began to launch plundering expeditions by sea. At first the targets were local, mainly the few remaining Greco-Roman communities along the northern coast of the Black Sea. By 255 some raiders even harassed the northern coast of Asia Minor. The next year they returned in far greater numbers. The situation was so serious that Valerian had to go to the area, permitting Shapur's Persians to strike at a weakened frontier."

In the meantime there had been more attacks on the European frontiers, on the Danube and southern Rhine. In 254 some Marcomanni reached Ravenna, and in 26o raiders from another tribe again broke into Italy. They were checked by Gallienus near Milan and some of them were defeated on their way home. Recently, an inscription was found at the city of Augusta Vindelicum (modern Augsburg) in Raetia thanking the goddess Victory for this success. It tells of `the barbarian peoples Semnones or luthungi killed or routed on the 24th and 25th April by the troops of the province of Raetia and from Germany and also local militia', and `the rescue of many thousands of Italian prisoners'. This party of raiders - it is interesting that even the Roman victors seemed uncertain just who they were - had got close to the Rhine before being defeated. A spectacular find of wagons loaded with gold and other valuables dumped in the river itself later in the third century almost certainly represents abandoned plunder.20

Some raiding bands were caught and defeated, even if only during their return home, but as many or more were successful. The Roman provinces seemed vulnerable and so inevitably further attacks came. A little later some Germans - in later sources they are said to have been Franks, but this may be an anachronism - marauded through Gaul and into Spain, sacking the city of Tarraco (modern Tarragona). More than a century later a local historian claimed that the scars of this attack were still visible, although no significant trace has been found in archaeology on the site.2'

In 267 there was a new wave of seaborne attacks from the Black Sea, which spread devastation all along the shores of Greece and Asia Minor. One source claims that the warriors, who included Goths as well as other tribes such as the Heruli, numbered no fewer than 320,000 men sailing in 6,ooo boats. The figures are wildly exaggerated, but are an indication of the panic caused by fast-moving attackers who could strike at widely separated targets in a short time. In a later age the Vikings would provoke similar terror. One group attacked Ephesus and burned the great Temple of Artemis, one of the seven wonders of the world. Athens was sacked by a band of Heruli, whose retreat was harassed by Athenians led by a local aristocrat, P. Herennius Dexippus. Sadly the history he wrote of these wars only survives in fragments.22

It is difficult to measure the amount of damage done during these raids from the archaeological evidence. Some sites, particularly those along the Rhine frontier, have revealed layers of burning and destruction. Yet it is not always clear whether this was the result of warfare or accident. Dating such a layer can be equally problematic, and often in the past they have been too readily assumed to have resulted from one of the recorded raids. There are also problems in understanding the very large rise in the number of coin hoards buried during the second half of the third century. Some almost certainly were hidden by people afraid of barbarian attack, who were subsequently killed, taken captive or otherwise unable to recover their treasure. Yet there may have been other reasons for hiding money at a time when the quality of coinage was varying hugely, as silver coins in particular contained higher and higher proportions of base metals. Some hoards may simply not have been worth the trouble of recovering.23

The impact of a raid can only have been dreadful for those communities and individuals actually attacked. Hardly any cities within the empire possessed modern fortifications. Athens made some effort to repair its ancient walls after the first incursions into Macedonia, but this was clearly not enough to stop the Heruli. Most cities were unwalled and scarcely any had a garrison to man whatever defences they did possess. They were vulnerable, and the news of attacks on other communities can only have increased nervousness. The provinces nearest to the Rhine and Danubian frontiers inevitably suffered the worst. This was especially true of cities and villages along the main communication routes, which were likely to be attacked more often. Northern Gaul suffered a good deal. Many villa farms and small settlements appear to vanish in the second half of the third century, although as usual we are dealing with only a small sample even of the known sites.

By the end of the third century every sizeable city within the empire had acquired a wall. There was no standard pattern, but almost all were very thick and strengthened by towers that projected out from them and so allowed defenders to throw or shoot missiles into the flanks of any attackers. In larger cities these towers were often designed to house artillery. Occasionally the defences looked stronger than they actually were, but the aim was clearly to deter any attack. Almost all of these new walls enclosed an area smaller than the full extent of the city in earlier centuries. Many cities in Gaul shrank dramatically in size, and presumably their populations had fallen as well. Amiens seems to have been attacked several times and in the second half of the third century became much reduced in size. After the sack by the Heruli, the Athenians built a new wall cutting through the old marketplace and excluding a number of great monuments. Much of the stone used was taken from older buildings that had presumably fallen into ruin or were now deliberately demolished. This plundering of old monuments for material to construct new defences was common in many cities.14

It was not just urban communities who fortified themselves, for the same inclination is visible in rural areas. The grander villas had often been built with towers, but these were essentially ornamental, increasing the visible presence of a great house and also providing an impressive view. In parts of North Africa during the second century some villas had already taken a more defensive form as a response to the threat from bandits and raiders. Now this became more common in other areas close to exposed frontiers. In Gaul before the arrival of Rome it had been common for settlements to be placed on hilltops. Roman peace brought a move down into the plains as communities grew in size and had no fear of enemy attack. In the later third century the trend was reversed and more and more walled villages appear on high ground as places of refuge during attacks.

The Gallic Empire

The Roman doctrine had always been that the best way of dealing with attacks was to defeat the enemy in open warfare. Ideally, the army presented a facade of overwhelming strength so that potential enemies were deterred from aggression. Every defeat weakened this impression, as did the frequent withdrawal of troops from the frontiers to fight each other in civil wars. Valerian's capture by the Persians was another humiliation at a time when there were already plenty of cracks in the facade. His son Gallienus was later vilified in most of our sources as indolent and far too fond of the luxuries of life in Rome. This was more than a little unfair, as he spent a good deal of time on campaign on the frontiers in Europe. In 268 he was in Greece chasing the bands that had plundered Athens and many of the other famous cities of the classical past. He is said to have won a victory, but the terms extended to the defeated Goths were very generous. Their king was taken into Roman service and given senatorial rank. Gossips also declared that the emperor was infatuated with a Gothic princess whom he took as a mistress."

Soon after his father's defeat, Gallienus lost real control over many of the western provinces as usurpers emerged. In 26o the governor of Germania Inferior, Postumus (fully, Marcus Cassianus Latinius Postumus) declared himself emperor. He had already arranged the murder of Gallienus' infant son and the latter's guardian who had been left in overall charge on the Rhine. Both the German provinces and all of Gaul soon rallied to Postumus, who may well have come from the Gaulish aristocracy. In time they were joined by Britain and much of Spain. Almost uniquely amongst usurpers who lasted for any length of time, Postumus made no effort to march to Italy and defeat Gallienus. Instead, he was content to remain on the defensive, fighting against Roman rivals only when they attacked him. Gallienus' armies were driven back twice. In 269 a challenger emerged in Mainz. Within a few months Postumus had defeated him, but his refusal to let his troops plunder the city led to his murder. The man the army appointed as emperor lasted twelve weeks before he too was killed. His successor Victorinus remained in power for the best part of two years, his murder allegedly being caused by an attempt to seduce one of his officers' wives.

Historians conventionally refer to this regime as the Gallic Empire. There is no historical basis for this, although one fourth-century historian does talk of Postumus `assuming power over the Gallic provinces'. As far as he and his successors were concerned they were the legitimate emperors of the whole empire. Consuls for each year were appointed - ignoring the fact that Gallienus continued to appoint them in Rome as well. It is uncertain whether a second Senate was formed. Aristocrats from Gaul filled many of the jobs in the imperial service, but this had more to do with the fact that it was difficult to draw men from further afield and provinces that did not acknowledge these emperors. Culturally, there was nothing particularly Gallic or `western' about the new regime. The titles and iconography - and also the law - employed by these emperors were entirely conventional. The only unusual thing was their reluctance to seek control of the remaining provinces of the empire.26

For much of his reign, Gallienus' rule extended over only Egypt, North Africa, the southernmost parts of Spain, Italy and the provinces behind the Danube. He is counted as the legitimate emperor because he came to power before his various rivals, and ultimately it was the men who succeeded him who would reunite the empire. Rather large claims have been made about his achievements, but these are based on hindsight and tend to ignore the peculiar conditions of his reign. One is a question of strategy. Defeats on all the frontiers are supposed to have shown the need for a central strategic reserve, which could move to confront any enemy who broke through the outer perimeter. Gallienus kept an important part of his army in or near Milan. Coins attest the presence of detachments from at least thirteen different legions - several of them whose parent unit was stationed in provinces no longer loyal to Gallienus. Especially prominent are the cavalry, who seem to have been grouped together under their own distinct commander. Septimius Severus had greatly increased the troops at his immediate disposal by raising II Parthica and augmenting the guard. Now, Gallienus had taken this much further by forming a complete army as his reserve and placing a new emphasis on the importance of mounted troops.

None of this makes much sense. We have no real idea of the numbers of troops involved or the proportion of cavalry to give an indication of how revolutionary this force may have been. Cavalry are faster than infantry over short distances, but on longer marches the advantage markedly decreases. Horses are also a lot more difficult to feed and keep in good condition than men. The idea of using a cavalry force as a mobile reserve only makes practical sense if the forces involved were relatively small. Also, northern Italy may at first sight seem like the heart of the empire, but for Gallienus, Milan was actually only just behind the frontier with the `Gallic Empire'. The circumstances were exceptional, but the deployment of these troops, and as far as we can tell their actual use on campaign, was entirely conventional.17

Gallienus' reign was long by the standards of the period, but his ultimate fate was reminiscent of many other usurpers whose power was briefer. In 268 the commander of his cavalry - we do not know the precise title and it may simply have been a senior rank commanding both horse and foot - rebelled against him. Gallienus came back from Greece to attack him, and after a victory in the field began to besiege the usurper in Milan itself. However, he had clearly alienated his other officers and these conspired to murder him. A false alarm was sounded and when the emperor rushed out of his tent to deal with the supposed enemy attack, he was stabbed to death.Z"

Gallienus was about fifty when he was killed. His successor Claudius II (fully, Marcus Aurelius Claudius) was a few years older and was an equestrian from one of the Balkan provinces. It is difficult to know how far he was involved in the conspiracy, but it is clear that a number of officers from this same region formed a powerful group in this period. The usurper in Milan was swiftly dealt with, and then Claudius spent the next two years fighting against barbarian raiders, first in Italy and then in the Balkans. In 269 he won a victory over the Goths and took the names Gothicus and Maximus in celebration. Early in the next year he was one of the victims of an outbreak of plague, achieving the distinction of being the first emperor to die of natural causes since Septimius Severus. His brother Quintillus was proclaimed emperor, but a few months later faced a challenge by one of the senior generals, Aurelian (fully, Lucius Domitius Aurelianus). When the rival armies moved to confront each other, Quintillus' men quickly decided to change sides. He either was killed or committed suicide when he realised what was happening.29

Aurelian was another equestrian from Illyria and had been heavily involved in the plot against Gallienus. He was a tough and experienced commander who within five years had reunited the empire. Occupied first on the frontiers and then dealing with the disorder in the eastern provinces, in 274 Aurelian attacked the `Gallic Empire'. Victorinus' successor Tetricus seems to have had little enthusiasm for the struggle and had serious problems maintaining the loyalty of his own troops. It is even claimed that he betrayed them, sending them out to fight in a hopeless position against Aurelian. Tetricus himself surrendered. Unusually, he was spared by the victor and even appointed to an administrative post in Italy. Similarly, many men who had held office in the army and administration within the `Gallic Empire' continued their careers in imperial service afterwards. None of the Gallic emperors had their names formally damned and wiped from the record.3°

The Barbarian Hordes

It is possible that in the third century some stretches of the empire's frontiers faced an increased threat from the peoples who lived outside. The very existence of the empire, as well as its diplomacy, encouraged the rise of powerful leaders within the tribes. There may have been other factors, too. The archaeological record suggests that the population may have been rising amongst the tribes beyond the Rhine and Danube at this time. It is possible that there were also problems caused by climate change and the exhaustion of soils through farming, although as yet there is not enough evidence to understand this in detail. Sea levels on parts of the North Sea coast do appear to have been rising, so that some parts of the coastline were flooded and in other places the soil became too salty to cultivate."

The scale of the threat to the frontiers had always fluctuated, but the biggest difference between this and earlier periods was the frequency of civil war within the empire. Troops were drawn off from the frontiers time and again to support the ambitions of their commanders. With so many changes in the senior ranks of the army - presumably often followed by considerable alterations at lower levels - it can only have been difficult for the army to train properly. The frontiers were thus weakened. It became easier to mount a successful raid, and each success only encouraged more attacks. It is notable that emperors and usurpers alike invariably saw foreign enemies as less of a threat than internal rivals. Time and again they settled with foreign attackers - including the Persians - or granted them generous peace terms so that they could deal with a challenge for the throne.

There was another side to civil wars. When one Roman army met another in battle it could not count on any superiority in discipline, tactics or equipment. This made numbers crucial, but it was difficult to raise and train new soldiers quickly. Frequent warfare also thinned the ranks of the regular troops and disrupted normal recruitment and training. Hiring the services of a barbarian chieftain and his followers was an attractive option chosen by many Roman leaders. These warriors might lack the discipline of professional Roman troops, but they were certainly more effective than hastily raised conscripts or volunteers. Yet when their Roman paymaster was beaten or murdered, such contingents could not be sure of welcome and employment from the next emperor. It is more than likely that some of the groups to maraud through the provinces had initially been invited into the empire. The desire of Roman leaders to recruit warriors in this way was another factor encouraging the emergence of powerful chieftains. These men were even more dependent on continuous warfare than was normally the case for leaders amongst the tribes. If they were no longer able to find a Roman willing to pay them, then the only other options were fighting against other tribes or attacking the empire. Even usually peaceful farming communities may have found trade disrupted so badly by civil wars that raiding became an attractive alternative. Others no doubt found themselves under attack by emperors keen for quick and clean glory won over foreign enemies and not other Romans.

The victims of a marauding band are unlikely to have been too concerned over the reasons that had prompted the warriors to go on the warpath. The local impact of a raid could be appalling, especially if other raids came in subsequent years. It was probably not much worse to be caught up in a civil war, since even fortifications were not always enough to hold back a Roman army. Civil war and barbarian invasion alike smashed communities and ruined livelihoods, adding to the ranks of the desperate and hopeless. The barbarians who swept across the Black Sea were said to have learned how to make boats and then sail them from the survivors of the cities they had overrun. It is clear that many army deserters and runaway slaves joined the raiding bands, while others set up as bandits on their own. The Marcomannic Wars had been followed by a so-called `Deserters War' in Gaul, and in the third century there were similar outbreaks of violence.31

Some areas - parts of North Africa, southern Italy, Sicily and most of Spain - escaped harm during these disturbed decades and other regions were only lightly touched. Gallia Belgica, the region nearest to the Rhine, suffered badly in spite of the best efforts of the Gallic emperors. Most Roman outposts on the far bank of the river were permanently abandoned and so was the Agri Decumates, the patch of land between the Rhine and Danube. Several forts along the frontier in this area show signs of violent destruction. At the auxiliary fort in Pfunz in Raetia three human jawbones were found inside its south-east tower. The nineteenth-century excavator guessed that these were the remains of sentries. It looked very much as if the soldiers had been taken by surprise, for they did not have their shields - traces of the bindings of these were found outside. If anything, the finds at the fort at Niederbieder in Upper Germany were even more dramatic. The skeleton of a soldier still wearing his hob-nailed army boots was found in the headquarters.

Such finds indicate the difficulty of interpreting some of the archaeological evidence for this period - particularly when it was acquired in early excavations using unsophisticated techniques. These military bases were clearly destroyed by enemy action, but it is harder to identify that enemy. The original excavators assumed that the attackers were Germanic tribesmen. More recently it has been suggested that they were other Roman soldiers, hence the degree of surprise at Pfiinz. At Niederbieder a crushed plaque decorated with the head of a youthful emperor was found in the remains of the headquarters building. The young Caesar has been identified as the son of Gallienus who was killed early on in Postumus' rebellion. Hence the garrison may have been attacked because it stayed loyal to the old regime. These interpretations are attractive, but remain conjectural. In the end, we do not know who stormed these bases. It may have been Roman soldiers or barbarian warriors, and if the latter then these could equally have been acting for themselves or as mercenaries and allies hired by one faction in a Roman civil war.

The archaeology does not suggest that the Romans were ejected by a barbarian onslaught from these frontier regions. In time tribes settled in the area, but it does seem to have been a slow and cautious process. After the initial struggles, the Agri Decumates effectively fell in the border region between the territory controlled by Gallienus and the regions loyal to the Gallic emperors. It may have made sense for each side to fall back on the line of the Rhine and Danube, respectively, in case the other attacked. Yet, whatever the cause for abandoning the advanced frontier line and the territory behind it, it is clear that the Romans were either unable or unwilling to reoccupy this region after the Gallic emperors had been defeated and some stability returned to the empire."

On the Danube the situation was similar and there was an even greater loss. Dacia was one of the more recent additions to the empire, but it was rich in mineral resources and for a century and a half had been extremely prosperous. Although it lay beyond the Danube, the natural barrier of the Carpathians protected much of the province from attack. During Marcus Aurelius' wars it had suffered a number of Sarmatian raids, but its major cities had been provided with walls from the beginning and damage was limited to the structures outside. The bulk of the raiding in the third century bypassed the province, yet there are signs of serious problems. From the middle of the century the archaeological record suggests a huge drop in the circulation of coinage. The provincial mint shut down and virtually no new coins seem to have been brought in from outside. Perhaps this was a sign of much of its legionary garrison being posted elsewhere. There is no trace of the auxiliary units moving anywhere else, they just seem to vanish from the record. Government control seems simply to have stopped, perhaps through lack of funds. Aurelian formally abandoned the province, although a new Dacian province was formed west of the Danube. Some of the population from the real Dacia may have moved there or elsewhere within the empire. Others remained behind. There was no rapid inrush of barbarians and for a while a form of Roman lifestyle seems to have continued, perhaps for some time. At the old provincial capital Sarmizegethusa, someone was able to convert the amphitheatre into a defensive redoubt sometime in the fourth century.34