Before we look at Roman frontiers in a more theoretical way, it is important to remember that, in practice, countries that are expanding have little interest in the limits to their power. The aim of this chapter is to trace very broadly the story of how the Roman Empire grew and to see to what extent consciousness of frontiers impinged on that experience when the main phase of conquest slowed down or came to a halt in some parts of the world.
Maps and Imperialism
In Rome as in Greece, maps were a rarity. Yet where they are mentioned they were almost always associated with conquest. On the eve of the great Sicilian expedition in the fifth century we are told the Athenians were so dazzled by the glittering prospects that all over the city young and old alike began drawing maps of the countries they hoped to conquer (Plut. Nic. 12). The first map we know about in Rome, which was probably a kind of picture map of Sardinia-what Livy called a simulacra picta (41 . 28. 10)-was to commemorate the victories of Tiberius Gracchus the Elder over the Carthaginians, and it was put up in a temple in Rome in 174 B.C.
The notion of taking the Roman imperium to "the ends of the earth" (ultimos terrarum fines) had gained currency in the second century B.C. as Rome's empire grew; but it reached its climax in the age of the dynasts of the first century. Pompey, we are told, 'wanted to reach with his victories Oceanus which flows round the world" (Plut. Pomp. 38.2), and he set up an inscription recording his "deeds" with a full list of the gentes he had subdued, saying, "He has taken the boundaries of the empire to the limits of the earth" (Diodorus 40.4). No doubt this was the model for the later Res Gestae of Augustus. Pompey's propaganda was accompanied by trophies and statues bearing representations of the oikoumene, the globe of the whole world (Dio 37.21.2).
Soon after this Julius Caesar commissioned the first known world map, probably as part of his great triumphal monument on the Capitol, portraying himself in a chariot with the globe of the world at his feet (Dio 43.14.6, 43.21.2). Latent in this imagery is the ideology that there were no limits to his conquests; and indeed, Ovid says Caesar was planning to "add the last part of the orbis" by a campaign 2 against the Parthians when he died in 44 B.C. (Ars am. 1. 177).
To the idea of ultimos fines of the periphery, which are contained in these accounts, we must add the geographic notion of an imperial center. The historian and antiquarian Varro, who lived at the same time as Caesar and Pompey, describes in his treatise on agriculture (RR 1.2. 1) how he met some friends at the temple of Tellus, where they were examining a map of Italy painted on the wall. That led to a discussion of the natural advantages Italy had overall other nations of the world because of its geographic position. It was only a short jump in imagination to start thinking that the geographic position of Italy was divinely ordained to ensure that Rome, which was in the center of Italy, could control the world. As the architect Vitruvius wrote at the time of Augustus, "The Roman people have the perfect territory [veros fines] in the middle of the heavens [mundus], with the whole world and its countries on each side. " Therefore, Vitruvius concluded, "The divine spirit has allotted to the Roman state an excellent and temperate region so that they may rule the whole world" (De arch. 6. 1. 10-11).
Frontiers and Imperialism
With the emperor Augustus Roman concepts of space and geographic measurement took a momentous leap forward. It is not surprising that as a single, autocratic emperor took control of the administration we should wish to ask about a frontier "policy" as such. Strabo, who wrote the finest work we possess on the political geography of the Roman Empire, inspired by Augustus's organization of the empire, says, "It would be difficult to govern in any other way that by entrusting it to one person" (6.4.2.). But there is a sense in which the Augustan empire marked a change in intellectual outlook too, much as was later to take place with the rise of the absolutist state. In both cases administrative efforts were made to coordinate scattered citizens, accompanied-as in Colbert's France of the seventeenth century-by an intense interest in geography and cartography with the aim of fiscal efficiency. The center and the periphery were being drawn together.
Augustus's personal interest in the organization of space was explicit. His attention to cadastration is referred to in the work of the agrimensores, who talk of "distances given by limites according to the law and decision of Augustus" (Hyg. Grom. Cons. limit., p. 157 Thulin). The emperor himself is said to have been "the first to display the world by chorography" (Div. orbis 1), and at his death he left a breviarium totius imperii-a detailed account of the empire's resources Jac. Ann. 1. 11).
But Augustus was a child of the republican conquistadores. Never at any stage of his life is it easy to prove that he lost the dream of world conquest or that he recognized permanent limits to growth. In the list of his achievements, the Res Gestae, which he left to be inscribed on his mausoleum and in various centers of the empire, his proudest boast was omnium provinciarum ... fines auxi -- "I have extended the boundaries of all the provinces" (RG 26). The boast was echoed by poets, artists, and architects.' On the celebrated Vienna cameo, the Gemma Augustea, the emperor is flanked by figures personifying Tellus (Earth) and Oceanus, while being crowned by a figure who is thought to represent Oikoumene. Above him are a globe and his personal astrological sign, Capricorn (fig. 6).
It was Augustus, too, who was responsible for the display of Agrippa's map in 7 B.C., the first world map we know very much about, although it was obviously in the tradition of Caesar's earlier map. Since it was set up in the Porticus Vipsania alongside other great public buildings of Augustus and Agrippa in the Campus Martius, it was very much a public monument to Augustus's own conquest of space or, as Pliny the Elder put it, "The world for the city to see" (orbis terrarum urbi spectandus; NH 3. 17) (Dilke 1985, 41 ff.). Besides a detailed "chorographic" commentary which Pliny the Elder was able to consult, and which may be the same work that was later said to have been written by the emperor, the map itself would have depicted only the major landmarks, like the great rivers and mountains. This is perhaps why later authors, such as Josephus, Tacitus, and Herodian, so often described the Roman Empire with frontiers "hedged about by the sea of Oceanus and remote rivers" (Hdn. 2.11.5; cf. Tac. Ann. 1.9- Jos. BJ 2.371 ff.). But there is no reason to think Augustus or Agrippa thought in terms of such "natural" limits.
At Augustus's death he left three documents along with his will; two of them, the Res Gestae and the breviarium totius imperii, have just been mentioned. The third was a book of instructions, a kind of codicil to the breviarium. Contained in this, probably, was what Tacitus says was the famous consilium I discussed in chapter 1. Tiberius, we are told, took it as a command, that the empire should be confined within limits." The exact context and significance of this codicil are matters for debate, but it is generally agreed that it was written by Augustus at the end of his life as he contemplated the result of the rebellion of the Pannonian auxiliaries in A.D. 6 and the appalling military losses by Varus in the Teutoberg Forest of free Germany in A D 9. it constitutes the first evidence of an imperial frontier policy-if that is what it was.
But was it really a "policy"? At the most it was surely temporary advice to Tiberius that the provinces of the empire were not yet sufficiently Romanized to allow further expansion. Roman historiography contained a deeply pessimistic consciousness of the fragil ity of imperial rule if allowed to outgrow its own resources, and this may well have been reflected in Augustus's practical counsel to consolidate. it would have been truly astonishing if Augustus had really intended all his successors to forgo the Roman virtue of winning military glory. Strabo, who wrote during Tiberius's reign, gave no hint of any ideological restraint in what he wrote. Tacitus in the next century made no comment on the fact that 35 Gaius and Claudius apparently ignored the advice when planning the invasion of Britain. On the contrary, Tacitus lamented the long delay in conquering Germany.
In fact, the ideology of imperium sine fine, an empire without limit, remained central to the Roman stereotype of a good emperor, proclaimed on coins or inscriptions and inserted in panegyrics. Pliny the Younger in A.D. 100, for instance, imagined Trajan on the banks of the Danube dealing with some insolent barbarian king, whom "nothing will protect from our very territory taking him over" (Pan. 16)-as indeed happened a few years later when Dacia was annexed. Cassius Dio's analysis of Trajan's motives for the conquest of Mesopotamia boiled down to "his desire for glory" (Dio [Xiph. ] 68.17. 1). Dio attributed the same motive to Septimius Severus a hundred years later (Dio [Xiph.] 75.1.1).
Trajan was not a pathological megalomaniac; in Roman upperclass mythology he was the most popular emperor after Augustus, with whom he was often associated. It was true praise thereafter to applaud an emperor as "braver than Augustus and better than Trajan" (fortior Augusto melior Traiano). Hadrian obviously realized this and anticipated, but was unable to prevent, the unpopularity of his withdrawal from Mesopotamia by claiming a republican precedent. Antoninus Pius, the most unwarlike of Roman emperors, also appreciated the force of the ideology and gave great publicity to the fairly ordinary military achievement of his reign when the frontier in Britain was advanced. Marcus Aurelius, we are told, was contemplating annexation of two new provinces north of the Danube when he died. His son Commodus was proposing to carry on the policy of going on "as far as Oceanus" until he discarded his father's senatorial advisers. With Septimius Severus the slogan propagatio imperii rose to new heights.
So if there was a discordant note, it appears to have been confined to Greek authors, whose tradition reflected more strongly than that of Roman nobles the sense of moral "otherness" and the barrier between urban order and the disorder beyond the city walls. Appian and Aelius Aristides, both Greeks writing in the reign of Antoninus Pius---the first an imperial official and the 36 second a professional orator-are often cited as examples of a new mentality of defensive imperialism and frontier fortifications that came to the fore in the mid-second century. This is what they said:
Beyond the outermost ring of the civilized world you drew a second line, quite as one does in walling a town.... An encamped army, like a rampart, encloses the civilized world in a ring. (Ael. Arist. Ad Rom. 81-82)
The Romans have aimed to preserve their empire by the exercise of prudence rather than to extend their sway indefinitely over povertystricken and profitless tribes of barbarians. (App., pr. 7)
But there are good reasons to reject these as typical views of the Roman Empire. Both may have been expedient ways by which those dependent on imperial favor interpreted the quiescent roles of Hadrian and his adopted heir, Pius. Even Aelius Aristides, however, went on to praise Rome because "you recognize no fixed boundaries, nor does another dictate to what point your control reaches" (Ad Rom. 10). That does not support the theory some have suggested, that Hadrian-who, it is true, was disliked by the Roman elites for his love of Hellenism-had adopted the Greek ideal of the cosmopolis. Although there was a kind of "moral" ideological barrier between Rome and barbarians, in practice Romans were far less rigid and had always accepted their civilizing mission. Roman authors who lived under Hadrian still spoke in the Roman tradition. Tacitus, although rarely explicit, was undoubtedly sympathetic to imperial expansion. Suetonius, his contemporary, was circumspect in praising Augustus for not seeking imperial glory at any cost, but he admired the wars of Germanicus that Tiberius had halted. Florus openly regretted the folly of Varus that had halted Roman expansion into Germany, while recognizing that there were gentes who could be left beyond direct rule provided they respected the majesty of Rome.
It remains to be seen how far the ideology of Roman frontiers corresponded to the reality. The account is arranged in chronological order, beginning with Augustus, but is divided between West and East. The object is to discover whether there were any fundamental strategic differences either between emperors or between regions of the empire.
The Western Frontiers
The places mentioned in the text can be found on the maps in figure 7 (Rhine), figure 8 (Danube), figure 9 (Africa), and figure 10 (Britain).
Not surprisingly, after the civil wars Augustus found himself immediately faced with a series of urgent military problems that had to be solved without reference to deep-seated policy studies. Spain, where he was involved in fighting until 19 B.C. as a result of the [actions of the civil war, led on naturally to Gaul, where Caesar's conquests had not halted the progress of Germans crossing the Rhine. Not content with crossing the Rhine, Augustus's commanders had by 9 or 8 B.C. reduced the region between the Rhine and Elbe to "practically a tributary province," says Velleius Paterculus (2-97), who was a serving officer under Tiberius. What our sources do not ever say is what modem writers have tried to infer, that Augustus was looking for a strategic riverine frontier or a shorter line of communication. indeed, the Elbe had been crossed in 9 B.C., and in about 2 B.C. the Roman army had set up an altar on the north bank.
The significance of this religious act, as illustrated later in A.D. 16 by Germanicus's consecration of a trophy between the Rhine and the Elbe, was to bring the territory within the spatial cosmos of Roman discipline." In fact, it seems plausible that Augustus genuinely intended to reach Oceanus. He boasts of sailing in Oceanus as far as the Cimbri of Jutland, and in about A.D. 4 the fleet made an attempt to sail around from Jutland to find the passage to Scythia, to the Caspian Sea, and to India but was deterred by bad weather. If Strabo is correct that the emperor forbade crossing the Elbe in order not to antagonize the tribes beyond, it was only for reasons of temporary prudence.
In Dalmatia and the Balkans much the same situation existed as in Gaul and Germany. Even before the defeat of Antony the territory of what was vaguely called Illyricum had required military 38 action, and a peace had been patched up with a diplomatic return of captured Roman standards. The urgency lay in the threat to the heart of Italy through the key passes of Raetia and Noricum (roughly Switzerland and Austria), which was rapidly resolved in 15 B.C. by the coordinated movement of two armies from the Rhine and northern Italy. The upper Rhine and the upper Danube were now connected. By 9 B.C. the middle Danube had been reached in tandem with the trans-Rhine campaigns launched from Gaul. Italy was henceforward firmly linked to the Balkans.
In both these cases it is easy to see the value of the single commander-in-chief, Augustus, and the possibilities of a strategic policy. That in turn has led modem scholars to wonder whether further pincer movements between 7 B.C. and A.D. 6 were ultimately aimed at providing a shorter line of communication between Germany and the Balkans, specifically along the river Elbe and its tributaries. Plausible as the strategy may appear, it is unfortunate that almost all our sources are defective for the period. Reconstructions based on inscriptions to some extent clarify the movements, which certainly included penetration into southern Slovakia. It may well be, too, that the campaign to the north bank of the Elbe had begun from the Danube. But nothing is said about a frontier.
If ancient sources do not attribute this activity to the search for a frontier, we may well ask, what was the objective? Augustus himself said that he believed all lands from Illyricum "to the banks of the Danube" were part of the Roman Empire (RG 30), which sounds as if he really did think the Danube was a frontier. But he did not confine his activities to the west bank of the river. "My armies," he says, "crossed the Danube and compelled the Dacian gentes to submit to the Roman people". The campaign was designed, much like that against the Germans beyond the Rhine, to keep back the penetration of new peoples pushing southward and took the route along the river Murq, crossing the Danube somewhere south of Budapest into the Carpathians. But there is no question that Augustus believed he had the right to rule them. The Danube was clearly not the limit of Roman power but is a good illustration of the dividing line between internal and external control.
Nothing illustrates better the difference in Roman thinking between the limits of organized space and their claims to imperial control beyond these limits than the progress of the army in Africa. By A.D. 14 the main legion of the province had advanced its base camp as far as Ammaedara (modem Haidra) on the high Tunisian upland plain. A boundary road was constructed linking Tacape (modem Gabes) and the camp, followed soon after by a massive work of centuriation of the land in the southeast. In a sense, therefore, we are right to regard this as a frontier line. But the Romans did not. The extraordinary southern expedition of Cornelius Balbus in 19 B.C. took Romans deep into the interior as far as the Fezzan. Despite the absence of any known posts or direct administration, Pliny the Elder-and no doubt Agrippa's mapsignified this action as evidence that "we have subjugated the gens of the Fezzani" (NH 5.35).
After Augustus it is often argued that, apart from Roman Britain, there was no substantial territorial addition to the Roman Empire in the West until Trajan's annexation of Dacia in the early second century. The rest is put down to retrenchment, adjustment, and administrative organization (for example, Nicolet 1983). This view gains its validity by stressing the point Suetonius made in his biography of Augustus, that allied native kings were from the start regarded as membra partisque imperii (Aug. 48)-an integral part of the metaphorical body of the Roman Empire. So when Claudius changed the government of Raetia and Noricum, Mauretania and Thrace from that of native alliance to provincial rule, he merely reclaimed what had originally been "the gift of the Roman people" and was not making any great new advance.
On the other hand, it was on the appeals from native British kings, who had also shown their submission to Augustus, that Claudius invaded Britain. In a sense they too were allied kings and therefore parts of the empire. But it is surely stretching the imagination to see the invasion as merely an administrative adjustment, when Claudius was so obviously an emperor who desperately needed military glory, as the twenty-seven military salutations in his reign prove. Nor was Britain, once invaded, left again to its kings, and although Suetonius claims that Nero considered abandoning Britain (Nero 18), since it was beyond terra cognita, there was also no sign in his reign of any search for a frontier.
It does not appear that Claudius felt any restraint in crossing the Rhine and Danube either, although the legionary camps now began to acquire stone buildings and an air of permanence along the river line. More and more archaeological evidence shows a Claudian and Neronian presence in the Taunus-Wetterau salient and in the Black Forest-Neckar region of Germany (Schoenberger 1969; cf. Tac. Ann. 12.28, 13.56). Much propaganda was made about the return of the standards and some of the prisoners cap- tured by the German Chatti from Varus in A.D. 9, which carried the same symbolic significance of submission as the return of the Parthian standards in Augustus's day. One reason for thinking that first-century emperors were involved in more than mere consolidation of Augustan termini is that, as some allied territories were annexed as provinces, further allied kings were brought into play as the "limbs" of empire: the Cherusci in northern Holland, the Vannian kings of the Suebi on the middle Danube, and the Quadi and Marcomanni in Bohemia and Slovakia. A famous inscription records how a Roman governor of Moesia under Nero fought the Sarmatians beyond the Danube and transported thousands of people, with their princes and kings, across the Danube into the province. "The might and power of the kings," says Tacitus, "depends on the authority of Rome" (Germ. 42).
The Flavian emperors more or less carried on from where the Julio-Claudians had left off, despite the destabilizing effect of the civil wars that had set alight the Germans of the lower Rhine. It is not easy to detect any real change in policy or the establishment of a "scientific frontier" by radical new measures, as has been claimed. In Britain Tacitus is explicit that in A.D. 81 Agricola ignored what might have been considered a scientific frontier at the Forth- Clyde, which was, after all, the shortest distance between two seas. It was not adopted, significantly, because the Roman commanders were driven on "by the courage of our armies and the glory of Rome" (Agr. 23). When a halt was finally called, it was not for reasons of local strategy but because the military manpower was needed in Germany.
In Africa the walls, forts, and ditch of the fossatum in southern Algeria and Tunisia are-in origin at least-rightly attributed to Hadrian. But they are intermittent and quite unlike the continuous wall in Britain. In Tunisia the clausurae walls are also associated with a series of tribal allotments, marked by boundary stones, which were made by order of the emperor Trajan and which lay to the south of the wall system. In other words, here is a vivid and visible example of the care one must take not to think that walls were the outer fines of Roman territory (CIL 8.22782- 88; Trousset 1978).
Whatever we may think about the debut of a new policy or strategy or defensive frontiers in the mid-second century A.D., it is obvious that on the middle Danube the Romans felt no restraint about crossing the river line. In southern Slovakia and Moravia a remarkable number of Roman-style buildings and legionary building tiles have been found on sites in the northern valleys of the Danube's tributaries (fig. 32) . Across the Hungarian plain Trajan's annexation of Dacia led to the construction of a west-east road from the Danube to the mouth of the river Mures, where a Roman vehiculatio or posting station is recorded on an inscription. Marcus Aurelius's proposal, after defeating the Marcomanni and Quadi, to annex two great new provinces north of the Danube would have carried Roman administered territory as far as Bohemia and must surely indicate that there was no agreed new defensive frontier policy. Although he did not carry out the project, there is no reason to think Romans gave up believing their sovereignty extended that far. In the fourth century the emperor Valentinian died of apoplexy when the kings of the Quadi behaved as though they were independent.
Likewise in Britain. After the curious advances and withdrawals between the occupation of Hadrian's Wall and the Antonine Wall, which are becoming clearer as more epigraphic evidence is unearthed, the Scottish line was finally abandoned in the reign of Marcus Aurelius. But then in 208 Septimius Severus launched a 48 new campaign in the footsteps of Agricola and worked his way right up the east coast of Scotland, only to return to the line of Hadrian's Wall. Severus's aim was explicitly to win glory and victories in Britain, according to his contemporary Herodian, and he probably first thought of extending the provincial boundary. As we know from a fragmentary inscription, the large camp at Carpow on the Tay and the fort at Cramond on the Forth remained in commission well into the rule of Caracalla. Although the visible line of Hadrian's Wall therefore was consciously and expensively refurbished, Severus intended that the Roman writ should run in Scotland.
The Eastern Frontiers
The names mentioned in the text can be found on the maps in figure 11 (Syria and Mesopotamia), figure 12 (Armenia and the northern Euphrates), and figure 13 (Palestine and Arabia).
The theory of a single grand strategy of the Roman Empire is a priori implausible, since geographic and political conditions were so different between various parts of the empire, particularly the eastern and western halves. Much of the western borderlands consisted of unformed states, loosely organized gentes and federations that were poor and little urbanized. in the East Greek culture and trade had penetrated most of the states Rome came into contact with, quite apart from the fact that over long periods the Persian, Greek, and Macedonian empires had imposed forms of centralized controls and urban organization.
In the West the driving force of Roman imperialism was primarily the glory of conquest, although tinged with some expectation of profit. in the East it was almost the reverse. The motive for an expedition under Augustus to Arabia Felix was "the report that they were very wealthy" (Strabo 16.4.22). The myths of El Dorado were fueled by the more immediate and sometimes tangible reality of the luxury trade and by tales from enterprising travelers, from the Caucasus Mountains, where it was said the rivers ran with gold, to the shores of the Yemen with their caravans from the East.
But above all other differences was the existence of the Parthian empire, which was taken over after the third century A.D, by the revived Persian empire. The Parthians had their own ambitionsor so the Romans persuaded themselves- of recovering the lands of their Persian predecessors with an imperial administration and army that, though not perhaps the equal of the Romans', was centralized, Hellenized, and familiar with the Roman legal concept of negotiated borders. By the time of Augustus they had also demonstrated their ability to inflict severe damage upon incautious Roman generals, such as Crassus at Carrhae.
This was a different world from that of the West. The Parthian king Orodes, it was said, was watching a performance of Euripides' Bacchae when the head of Crassus was brought to him. Ten thousand Roman prisoners were settled in Parthia and married local women after the battle. In A.D. 36 the Roman governor of Syria, L. Vitellius, negotiated with a Parthian noble who was also a Roman citizen. It comes as something of a shock to realize that C. Julius Antiochus Philopappus, Roman consul Of A.D. 109, who has left his visible mark on the skyline of Athens, was the grandson of a Parthian king.
If the two worlds of East and West were so very different, then, the critical question must be asked, Did the Romans have a different policy and a different concept of frontier in the East from that in the West? In the East we would expect, perhaps, an administrative boundary, legally negotiated between equals and determined by political rather than military conditions. For this idea of power sharing there is some evidence in the sources and in several historical examples of treaties negotiated around the assumption that the river Euphrates was the legal boundary, the most obvious being that of Augustus in 20 B.C., when the standards of Crassus were returned by Parthia.
But we need to take some care in assuming that this is the whole story. When Pompey was asked to recognize the Euphrates as a treaty line, he gave the unsatisfactory reply, "Rome would adopt a boundary that was just" (Plut. Pomp. 33.6). And Augustus publicized on coins the return of the standards as submission to Roman rule .31 The eastern front was, in any case, more than just Syria and the Euphrates. There were Palestine and Nabataean Arabia to the south, while to the north were Armenia and the shores of Pontus as far as the Caucasus. The Euphrates was perhaps a plausible dividing line to negotiate in northern Syria, but in the southern desert part it was the main route of east-west movement." The contrast between deserts, steppes, and mountains was as great on the eastern front as between Hadrian's Wall and the African fossatum in the West.
There is another difficulty in answering the question: the relatively undeveloped state of archaeology in the East compared with the West, despite important new work in the past decade that has radically changed some old concepts and ideas about the eastern frontiers. One simple fact that militates against the notion of a fixed, linear frontier in the East is that in the first and second centuries of the Empire more Roman military inscriptions have been found beyond the Euphrates on the Cappadocian front than on the immediate line of what was once called the eastern frontier (Crow 1986b). Whether this is an accident of research or a reflection of reality is impossible yet to judge.
The irony about our ignorance is that the eastern front was occupying proportionately greater and greater Roman attention and manpower than the West from the first century to the fourth. Whereas Pompey's settlement considered three to four legions in Syria sufficient as the sole garrison for the whole Middle East, by the second century there were eight legions and as many auxiliaries stretched out from the Black Sea to the Gulf of Aqaba, a force of some hundred thousand men, to which Septimius Severus added 34 two more legions in Mesopotamia.
Between Pompey's settlement in the late Republic and the emperor Nero there were no fundamental territorial changes in the area directly administered by the Romans, and even Corbulo's campaigns into Armenia under Nero in the 60s did not annex any new province. But this certainly did not mean there was an aversion to military action or advance. Pompey's answer to Parthian demands for a treaty recognizing the Euphrates as the common frontier was in effect a refusal. Romans knew that at his death Caesar was planning a campaign against the Parthians in which Octavian was to participate (Suet. Aug. 6). Augustan poets reflect the dreams of eastern victories taking Roman arms to India. According to the contemporary poet Propertius, "the ends of the earth are preparing triumphs for you, Tiber, and the Euphrates will flow subject to your jurisdiction" (Carm. 3.4).
Even after the diplomatic settlement of 20 B.C., we are informed that the emperor's grandson Gaius, when he died in 2 B.C., was planning campaigns "extending the termini beyond the Rhine, Euphrates, and Danube" (Sen. Brev. vit. 4.5). After Augustus had claimed that he could have made Armenia a province but had preferred to leave it a dependent kingdom, all future emperors regarded the country as a Roman "gift," revocable at any time. It was a Roman right to install and control the kings of Armenia, a right they were prepared to enforce with arms.
Farther south, on the Syrian front, it is completely ambiguous whether Palmyra, the caravan city, was inside or outside the Roman Empire, and Appian calls the residents "border people" who skillfully played off Rome against Parthia (BC 5.1.9). But that was hardly the way the Romans would have seen it. They could claim to have some authority over the city by A.D. 18, when we find a Roman tax law published there. Theoretically, therefore, that extended Roman military rule, through Palmyrene militia, along the desert routes and watering points as far as the lower Euphrates and perhaps beyond. But it is a perfect example of control without rule or of ideology before reality. Parthia too had interests there, and it is a measure of the fluidity of the frontiers that Palmyrene militia were permitted to operate from Dura-Europos and Ana on the Euphrates even while they were still Parthian cities in the first century A.D. Palmyra remained an Arab city with considerable support among the tribes of the Hauran and northern Arabia, which allowed it to act as an intermediary between Parthia and Rome while marshaling the desert trade routes. Possibly Hegra in the Hejaz played a similar role in the second century A.D. in controlling the routes from Petra to Saudi Arabia.
Claudius was probably the first to construct a fort on the line of the Euphrates; but he also seems to have reinforced the Armenian king beyond the Euphrates with a Roman garrison (Dobrawa 1986, 96-97; Mitford 1980, 1174). Indeed, the campaigns of Corbulo under Nero were conducted primarily to "hold on to Armenia"that is the way Tacitus described it (Ann. 13.8)-and already in A.D. 62 the territory was being regarded by the governor of Syria as ready for direct provincial rule (Tac. Ann. 15.6). There were also Roman fortified sites in Mesopotamia at this date. Although in the settlement after Corbulo's wars the Parthians demanded Roman withdrawal to the Euphrates line "as before" and the Romans are said to have set up a line of forts "against the Armenians" (whatever that means), it is not said that Nero formally accepted the Euphrates as the Roman frontier. Strabo, who wrote under Tiberius, probably captured the ambiguity of the situation when he said that, although the Parthians regarded the Euphrates as the 'limit" (horion) of their territory, "Parts within are held by the Romans and the phylarchs of the Arabs as far as Babylonia; some of them adhere more to the Parthians and others more to the Romans" (16.1.28).
As in the West, the Flavian achievement in the East has to be judged almost exclusively by archaeology. Here too theories have been put forward of a new, scientific approach, the start of a new policy of fixed frontiers. But many are now inclined to revise earlier claims of a Flavian master strategy and to regard as a myth the idea of a fortified frontier along the Euphrates that supposedly anticipated and resembled that of the fourth century.
The most striking change was in the disposition of the legions. Cappadocia was developed as a first-class military province, and the two legionary bases were built at Satala and Melitene, roughly along the lateral line of the upper Euphrates and its extension road to the Black Sea. In Syria-Commagene there was probably a parallel pair of legionary bases at Samosata and Zeugma, although the precise archaeological date for their foundation is lacking. Both were important crossing points of the Euphrates, certainly occupied earlier and probably customs posts on the provincial border.
But these sites did not so much form a military front as provide a base for invasion routes into Armenia or Mesopotamia and control of caravan routes coming out of the East. Legionary bases were not normally placed on the leading edge of a military front. Rivers, as I will stress again and again in this book, generally were not natural frontiers but lines of communications and supply. The legionary camps on the rivers were therefore the rear base for attacks.
In fact there is more evidence of the Flavian army beyond the upper Euphrates and Pontic road than on it. In Azerbaijan a rock inscription records a legionary detachment overlooking the Caspian Sea one thousand miles from its base at Melitene. Another detachment was placed in the Caucasian Gates (modem Darial Pass), where a wall also was built for the Iberian king at the southern end of the pass. A fort was rebuilt at Gomea near the Armenian capital of Artaxata, and another inscription (now lost) 41 records a Roman detachment on the Araxes.
The evidence gives substance to the poem of Statius (Silv. 4.1.41-43) suggesting that Domitian was planning an expedition to India and the Far East. Another possibility is that Colchis and the Caucasus were being brought under control to prevent piracy of grain supplies for the legions from the Black Sea, which was arriving at the port of Trapezus (modem Trebizond). Colchis is another example of an ambiguous frontier state, which Procopius later described as "subject to the Romans without paying tribute" (BPers. 2.15.2-3).
What seems sure is that Romans did not think they had reached the limits of empire in the East when Trajan came to power in the early second century. His annexation of Armenia and Mesopotamia between A.D. 114 and 117 provides illustrations of the sort of frontier thinking that existed in the East, But it is worth noting that as Armenia was organized, so too were the military garrisons on the Euphrates, now far to the rear, strengthened. An obvious example is the new fort at Zimara, where the upper Euphrates makes a sharp bend southward as it flows out of Armenia. There is no question, therefore, of concluding from the archaeological evidence that a fortified river front line was intended. Armenia itself came under the governor of Cappadocia, as far as we can see, but no attempt was made to define the eastern borders.
Although Hadrian reverted to an Augustan type of Roman control of Armenia through allied kings and abandoned Mesopotamia, there is no reason to think he renounced all claims to control east of the Euphrates. if we can trust the story that Hadrian's excuse for pulling back Roman troops was the example of Cato's withdrawal from Macedonia in the second century B.C. (HA Hadr. 5), the analogy was one of control by indirect rule. Both in the far north (Iberia, Armenia) and beyond the middle Euphrates (Edessa, Osrhoene, Bactria, Hyrcania) we are told of the submission of allied kings, not all of it empty propaganda. There is no warrant to conclude from this evidence that Hadrian had adopted a purely defensive policy.
Whatever we may think of Hadrian's intentions, however, it is obvious that during the rule of Marcus Aurelius and Verus (A.D. 161-69) Trajan's policy of direct rule and administration of Mesopotamia as a province was thought both possible and desirable. Although the final coup de grace of annexation had to come from Septimius Severus, probably in A.D. 198, there were already Roman garrisons, forts, and some sort of imperial administration (at Nisibis) in the northern part of the interriverine territory before it became a province: "occupation without annexation," it has been termed (Duncan-Jones 1979). It is a good example of how indeterminate the frontier remained and how little the Romans felt constrained by the differences between directly administered provinces and indirectly controlled territory beyond. In Armenia Marcus Aurelius deliberately left the king in place but bolstered up by a Roman garrison in the new capital of Kainopolis (Dio 71.3.1'[Suda]).
Severus did not then actually change the frontiers of Mesopotamia. He only converted the area of indirect control into one of direct provincial administration. The cities of Mesopotamia continued to be the keys to Roman power without there being in any real sense a new frontier. The city of Hatra, which Severus unsuccessfully assaulted, became a base for Roman troops soon after, but we look in vain for any sign of a frontier. On the middle Euphrates we know more about Severan forts below Dura-Europos as a result of recent discoveries and Dura itself had become the base of a dux ripae under the Antonine emperors. That must mean the Romans controlled both banks of the river and the trade traffic that used the river, but it is difficult to see what territorial significance this had (Isaac 1990, 147-53).
Trajan's attempt to add Mesopotamia to the list of provinces was paralleled by his organization, to the south of Syria, of the new province of Arabia, where recent study has clarified his work. What emerges is that the Arabian frontier was in reality no more than a road-the via nova Traiana-studded with fortified posts, which ran from the Gulf of Aqaba to Bostra. The road followed the old King's Road that had been worked by the Nabataean Arabs to control the trade routes from the Red Sea and the Yemen to Petra; and in many ways it resembled the earlier construction.
The road was supplemented by long-distance outposts at Hegra in the Hejaz (Saudi Arabia) and along the route of the Wadi Sirhan, although the precise dates for these last developments are uncertain. The military evidence from the Azraq oasis in the Wadi Sirhan is Severan in date, and the organization of the Arab tribes in the Hejaz by the Roman governor is first attested by a famous inscription at Ruwwafa establishing an emperor cult in the time of Marcus Aurelius. But these dates are only post quem, and it looks at least possible that from the start the Romans established patrols of native troops paying allegiance to Rome on the southern and eastern flanks of the Arabian road to serve in much the same way as Palmyra did at the northern end. They were meant not so much for defense in depth as for surveillance of nomadic movements and to provide caravan escorts.
In other words, the Arabian frontier was a true limes, a road for movement and not a blocking, defensive system. In the Palestine sector it has been suggested that the road forts were posted more for internal security against the bandits of Judaea than to protect the territory against external, nomadic raiders. But this perhaps underestimates the need to monitor even peaceful transhumants. The fort at Khan Kosseir, for instance, which lay on the inner steppe of Syria along the road from Damascus to Emesa, recorded its construction under the Severan emperors "for the sake of public security and against the terror of the Scaenitae Arabs" (CIL 3.128). But there is no way of deciding whether these raiders came from within or from outside the limes.
The important point, however, is that we do not find in this sector anything that could be called a frontier "system." Here particularly we have confirmation that the eastern frontier, as it is traditionally described, from the Pontic shore to the Red Sea was in essence a line of communication and supply, the base from which the Romans extended their control without any sense of boundaries.
This does not mean that such lines were never defensive when external attacks came. We have already seen how the concept of a praetentura of forts "stretched out" like a battle line had been applied by Tacitus to the Armenian front. But it is not proved that fines imperii were perceived as having a permanently defensive role. Indeed, it has been calculated that recorded Parthian attacks on Roman provinces over three centuries were only half the number of those launched by Romans on Parthia; and it is never proved that the Parthians took the initiative.
In short, although strategy obviously differed between East and West, the concept of frontiers was remarkably similar. That is the subject of the next chapter.