Current Orthodoxy

As was the case under the Ptolemies and the pharaohs before them, relations between Rome and Kush focused on Lower Nubia. For centuries the kings of Kush had sought to extend their power over all of Lower Nubia. After failing to accomplish their goal in the late third and early second centuries BC, the decline of Ptolemaic power in the first century BC gave them a second opportunity. And, as Strabo's characterization of Pselchis - modern Dakka - as Aethiopian indicates ¹⁸, they were on the verge of finally achieving their goal, when Rome intervened in 30 BC.

Fresh from the suppression of a revolt in the Thebaid, C. Cornelius Gallus, Octavian's new Prefect of Egypt, crossed into Nubia in force, appointed a   rannos - a Roman client ruler - for the Triacontaschoenus, and forced local Meroitic officials to recognize Roman suzerainty and to agree to pay tribute to Rome ¹⁹. The sequel is too familiar for detailed treatment here. The middle and late 20s BC were marked by raids and counter­raids by Meroitic and Roman forces that finally ended in 20 BC with a meeting between Meroitic ambassadors and Augustus at Samos at which, according to Strabo, the Meroites "obtained all that they desired, and Caesar even remitted the tribute he had imposed" ²⁰.

This agreement is usually referred to as the "Treaty of Samos" and is assumed to have governed relations between Rome and Kush for almost three centuries until Meroe opportunistically took advantage of the weakening of Roman power in the eastern Mediterranean in the mid-third century AD to annex the Dodecaschoenus. Although no text of the "Treaty of Samos" exists, four principal provisions have been ascribed to it: (1) recognition of the independence of Kush; (2) cancellation of tribute; (3) withdrawal of all Roman forces to the Dodecaschoenus; and (4) guaranteed Meroitic access to the Temple of Isis on the island of Philae ²¹. Subsequent to the conclusion of the Treaty of Samos rival Roman and Kushite claims to the Dodecaschoenus are supposed to have been neutralized through reconfirmation of the status the region had enjoyed under the Ptolemies as the estate of Isis of Philae. Finally, the Dodecaschoenus is believed by some scholars to have been governed through a sort of condominium ²² in which Roman and Kushite officials were simultaneously active in the region.

Roman historians understandably lose interest in the story at this point. Cornelius Gallus' incursion into Lower Nubia and the subsequent campaigns of Gaius Petronius were, after all, only minor episodes in the grand story of Roman imperial expansion under Augustus; and, in any event, Roman suzerainty over Kush lasted only a few years before ending in 20 BC. For Roman historians, therefore, the most significant aspect of the story is Augustus' reaction to the failure of his policy in the upper Nile valley.

In their view Augustus' decision to limit Roman power to the Dodecaschoenus, foreshadowed his pragmatic reaction to other similar but much more serious imperial crises later in his reign such as the collapse of his German policy in 9 AD. Thus, Augustus, chastened by the unexpected intensity of Kushite resistance and discouraged by the poverty of the region, cut his losses, without, however, failing to secure important advantages for Rome, namely: control of the valuable gold mines in the Wadi Allaqi ²³ and the security of Egypt's southern frontier with, as Strabo noted ²⁴, minimal cost in manpower and treasure. So, according to Derek Williams ²⁵, "under these arrangements [sc. the Treaty of Samos]...the Nile frontier would remain inviolate for 250 years" while Erich Gruen ²⁶ succinctly observed in the new edition of the Cambridge Ancient History that "peaceful relations prevailed thereafter" between Rome and Kush.

Not surprisingly, those "peaceful relations" dominate discussion of subsequent relations between Rome and Kush. The evidence is primarily archaeological and includes: epigraphical allusions to the yearly visits of Kushite pilgrims to Philae ²⁷, and, of course, the numerous objects imported from Roman Egypt during the Principate that have been discovered on Meroitic sites ²⁸. Included in this substantial but little studied corpus of classical objects are a wide variety of small but high quality metal, glass, and ceramic objects including eating utensils; items related to personal adornment such as rings, jewelry, beads, and mirrors; and household furnishings and decorative objects. In addition, the existence of classical motifs and images in Roman period Meroitic art and architecture, the auloi from Tomb Beg N. 6 ²⁹, and the famous column drum inscribed with the Greek alphabet Sir John Garstang discovered in the Royal enclosure at Meroe ³⁰, may imply the presence also in Kush either of individuals skilled in their production and use or appropriate models. These items clearly are evidence for contact between Rome and Kush, and evidence for contact of a particular type. As their peculiar distribution indicates - essentially royal and elite residential and funerary sites at Meroe and Napata and elite tombs at major provincial centers such as Qasr Ibrim, Faras, Sedeinga, and Karanog ­ they most likely document not trade but the gift exchanges that accompanied diplomatic contacts between the two states.

A New View Of Roman Policy In Lower Nubia

The above summary with its emphasis on the importance of the "Treaty of Samos" and the consequent expansion of Greek influence in Nubia-­however one evaluates its significance--fairly represents, I believe, what I have called "current orthodoxy." In my opinion this interpretation is based on a fundamental misconception of the character of relations between Roman Egypt and Kush during the Principate. The problems are twofold: First, by assuming that political and military issues were definitively resolved in 20 BC and concentrating thereafter on culture-historical questions, relations between Rome and Kush are treated as though they were essentially static for almost three centuries. Second, and equally important, evidence is ignored that indicates that relations between the two states developed over time and that on occasion, beginning most likely sometime in the second half of the first century AD, those relations could be tense and sometimes even overtly hostile.

That evidence is threefold. First, unlike earlier Kushite kings who often modeled their titularies on those of their Egyptian contemporaries, first century AD Kushite kings modeled theirs on those of Napatan kings such as Anlamani, Aspelta and their successors ³¹, thereby reaffirming their ties to the founders of their kingdom and at the same time their independence from Rome. Second, as L. Török ³² pointed out a decade ago, Building M 292 in the Royal Enclosure at Meroe with its frescoes depicting Roman prisoners and the placement of the famous head of Augustus under its entryway, where it would be trod on whenever someone entered the shrine, must reflect hostility toward Rome. Third, a fragmentary and much discussed papyrus, P. Volgiano 40, whatever its original context, clearly refers to a battle that took place somewhere in lower Nubia between Roman cavalry units and Aethiopian and Trogodyte - that is, Kushite and probably Blemmye - forces in the late first or early second century AD ³³, and there is circumstantial evidence pointing to further tension thereafter ³⁴.

The underlying problem, of course, is the assumption on which the "current orthodoxy" rests, namely, the existence of a Treaty of Samos concluded in 20 BC in which Augustus formally abandoned Roman claims to suzerainty over Kush. In fact, however, the Treaty of Samos is a modern scholarly myth, without support in the ancient sources. Strabo--our only source for the meeting between Augustus and the Kushite envoys in 20 BC--refers instead only to generous concessions by Augustus, and Strabo's account is confirmed by Augustus' own version of his dealings with Kush in his Res Gestae.

Under the rubric: "I extended the territory of all those provinces of the Roman people on whose borders lay peoples not subject to our government" ³⁵. Augustus tersely records the Roman capture of Napata in his Res Gestae as follows:

At my command and under my auspices two armies were led almost at the same time into Aethiopia and Arabia Felix; vast enemy forces of both peoples were cut down in battle and many towns captured. Aethiopia was penetrated as far as the town of Nabata, which adjoins Meroe.

In Augustus' account there is no mention of abandoning claims to Kush, only of a victorious military penetration deep into Kushite territory. The reliability of the Roman accounts of Petronius' campaigns is not in question here ³⁶, only the Roman view of their significance. An illuminating parallel is provided by Augustus' public interpretation of his contemporary and similar treatment of Armenia, which, like Kush, had also been temporarily transformed into a province and then lost again.      In chapter 27 of the Res Gestae he described his treatment of Armenia in the following terms:    "Greater Armenia I might have made a province...but I preferred, following the model set by our ancestors, to hand over that kingdom to Tigranes...." In other words, Augustus rationalized his failure to maintain Roman control of Armenia as an example of self-restraint, not a formal recognition of Armenian independence, nor could he have done otherwise.

Research into the character of Roman imperial ideology has been an active area of contemporary classical scholarship ³⁷. One of the clearest results of that scholarship is the realization that Augustus and his contemporaries recognized no limits to the potential extension of Roman power and rule - the gods had granted the Romans "empire without limit (imperium sine fine)" in the poet Vergil's famous phrase. The only acceptable restriction possible was expedient restraint in the exercise of that power. In other words, once Roman arms had successfully penetrated Meroitic territory, ideology would not have allowed Augustus to formally and definitively recognize the independence of Kush at Samos in 20 BC. Rather, like Armenia, he graciously "preferred to hand over" Kush to its native rulers instead of transforming it into a province, without, of course, abandoning his right to do so at some time in the future. Any doubt that this was the official Roman view of these events is erased by the call for the conquest of Kush in later Augustan literature ³⁸ and the presence in late Roman provincial lists, which were based on Augustan sources, of a "pseudoprovince" of Aethiopia, that is, Kush ³⁹. Kush escaped renewed Roman aggression after the death of Augustus, therefore, not because its independence was guaranteed by the Treaty of Samos, but because most Roman emperors followed Augustus' precedent and chose to exercise restraint in dealing with Nubia ⁴⁰. The one exception was Nero - an emperor desperately in need of military laurels – to whom Pliny and Cassius Dio ascribe plans for a Nubian campaign ⁴¹. Pliny even claims that the famous expedition dispatched by Nero in search of the sources of the Nile was actually a reconnaissance mission for the projected campaign ⁴², and Michael Zach recently plausibly argued that Nero attempted to sow division in the Meroitic royal family in order to facilitate his planned conquest of Kush ⁴³. Any plans Nero may have had for an invasion of Kush, however, ended with his death in 68 AD. No similar restraint, however, limited Roman actions within the Dodecaschoenus.

Closely connected to the idea that Roman relations with Kush were regulated by treaty is the notion that governance of the Dodecaschoenus took the form of a condominium. Quite the contrary. Rome's actions in the first and centuries AD belie any such idea, suggesting instead a determination to assert total Roman sovereignty in the Dodecaschoenus. The most visible sign of this policy were the temples Augustus and his successors built in the Dodecaschoenus. These temples are usually viewed as representing an attempt by the Romans to appeal to the local population through support of their traditional gods. But while such an explanation might appear superficially plausible in the light of Roman patronage of a local deity like Mandulis at Kalabsha and Ajuala, it fails entirely to explain the construction of a temple in honor of Sarapis--a deity with no local ties but strong connections to the Roman state ⁴⁴ - at Hiera Sykaminos in the early first century AD ⁴⁵ nor the incorporation of the important temple of Thoth of Pnubs at Pselchis - a shrine with close ties to Meroitic royalty - within the walls of a Roman fort in the second century AD! Rather, like their Ptolemaic and Kushite predecessors, the first century emperors embarked on a program of temple building and renovation of unprecedented scale-­twenty temples were functioning in the Dodecaschoenus in the first two centuries of Roman rule ⁴⁶ - for the purpose of creating a sacred landscape in the Dodecaschoenus that would visibly proclaim divine sanction for Roman rule ⁴⁷.

As everywhere else in the empire, it was ultimately the Roman army that gave force to claims of Roman authority in the Dodecaschoenus. Historians have underestimated the extent of Roman military activity south of the First Cataract. Deceived by the poor recording of Roman military remains in the Dodecaschoenus - only the fort at Pselchis was systematically explored before the flooding of the region ⁴⁸ - they have far too readily assumed that Strabo's assertion ⁴⁹ that only three incomplete auxiliary cohorts based near Syene were required to secure Roman authority in Lower Nubia held true for the Principate as a whole. In so doing they have paid too little attention to Procopius' claim that Diocletian abandoned the Dodecaschoenus because of the expense of maintaining the large garrisons posted there ⁵⁰, which strongly suggests that Roman forces in the region were strengthened after Strabo wrote in the first century BC. The exact date is unknown, but archaeological evidence points to the reign of Trajan early in the second century AD ⁵¹. The extent of that reinforcement is made clear by the Antonine Itinerary, which reveals that a tight network of forts linked by roads covered the Dodecaschoenus ⁵². As to the size of the Roman garrisons in the area, explicit evidence is lacking, but one item is suggestive. The Roman fort at Pselchis - a rectangular structure approximately 175 meters by 152 meters on a side or ca. 26,600 square meters in area - was large enough to house a full auxiliary cavalry cohort ⁵³ and that fact, together with the other known forts and outposts, suggests that the Roman forces assigned to the Nubian frontier may have been significantly increased over what had been stationed there in Strabo's time. The same trend toward an expanded Roman presence in the region holds true for the governance of the Dodecaschoenus.

The theory that the Dodecaschoenus was organized as a condominium rests solely on the fact that during the first decades of Roman rule the governmental agents most visible to us in the sources are individuals of probable Meroitic origin who are identified as strategoi and agents of Isis or Thoth of Pnubs ⁵⁴. L. Török pointed out ⁵⁵, however, that, despite their Meroitic origin, the fact that their actions are dated by the years of the Roman emperors and taken in their name must mean that they were in Roman - not Meroitic - service. The implication is clear. As they did in Egypt and elsewhere, the Romans initially left traditional institutions intact, while enlisting the collaboration of local elites to consolidate their rule ⁵⁶. For similar reasons, Augustus confirmed the status as the estate of Isis of Philae that the Dodecaschoenus had enjoyed under the Ptolemies ⁵⁷. Villages within the region continued to be governed by their elders and disputes were settled in Egyptian fashion by oath before the deities of the region ⁵⁸. In the early years of Roman rule, therefore, public signs of Roman authority were probably limited to little more than the images of Augustus as Pharaoh on the walls of local temples, the use of Roman date formulas in official documents, and the presence of small detachments of soldiers that manned outposts in the region. As was also true elsewhere in the empire, however, manifestations of Roman authority in the Dodecaschoenus had become much more overt by the early second century AD.

The most obvious sign of the change was the termination of the status of the Dodecaschoenus as the estate of Isis of Philae. The last recorded confirmation of that status was by Augustus ⁵⁹, and by the second half of the first century AD inscriptions at Philae reveal the inhabitants of Philae and the Dodecaschoneus jointly swearing allegiance directly to the emperor ⁶⁰. Moreover, epigraphical evidence indicates that, just as in Egypt, temples in the Dodecaschoenus were subject to the authority of the High Priest of Alexandria and All Egypt ⁶¹. To be sure, local nobles with ties to Meroe continued to serve the Roman government; in the mid-second century AD the orator Aelius Aristides interviewed one whom he describes as one of the "powerful men in the area" who "had a certain responsibility" ⁶². Their judicial functions, however, seem to have been curtailed, since graffiti at Pselchis indicate that the Strategos of Ombos regularly visited the region to hold court ⁶³, the earliest dated visit is from the reign of Domitian at the end of the first century AD ⁶⁴. In addition, inscriptions from Delimit show Roman military officers in the reign of Trajan (110/111 AD) deciding border disputes between neighboring villages in accordance with an official land survey ⁶⁵. Even more significant for the life of the local population, however, was the introduction of the Roman tax system. Mention of a sitologos - "grain-collector" - of Pselchis and the upper topos of the Dodecaschoenus ⁶⁶ indicates that in the second century AD the land tax was levied and collected on the Egyptian model, while a reference to epiteretai eidon - tax overseers - of the Dodecaschoenus and Indian Ocean ⁶⁷ reveals that the same held true for collection of transit taxes for goods coming from Nubia and the Red Sea. More speculatively, it is possible that the opening of the quarry at Kertassi in the late second or early their century AD resulted in the increased use of forced labor in the area, if the quarry was worked on the Egyptian model ⁶⁸.