Byllis is a well-preserved greenfield ancient site in central southern Albania at 524 masl. Known locally as the fortress of Hekal (the nearest hamlet), it occupies a flat hilltop overlooking the River Vjose (the ancient Aôos), which runs through a deep valley to the west. The ancient city occupies a triangular plateau about 30 ha in area. It almost certainly lay on a route from ancient Apollonia (situated on the plains to the north) leading to central and southern Epirus.

The town was created in the 4th c. BCE and remained occupied until the 6th c. CE. Initially it was one of several fortified Illyrian hilltops controlling the high hills of Mallakastër. By the 3rd c. BCE it had become the center of a koinon of the Bylliones tribe mentioned in an inscription at the (southern Epirote) sacred site of Dodona. From this period the town was walled (Byllis I), the 30 ha of urban buildings being arranged on an orthogonal grid of streets. Within the grid were raised terraces, on which were constructed a large theater, a stoa, and a gymnasium, besides other civic monuments. Outside the Hellenistic walls, remains of three cemeteries from this period have been discovered. In the 2nd c. BCE, it boasted an active mint.

Byllis is briefly mentioned in the civil war between Julius Caesar and Pompey, as the former fled south after suffering defeat in a battle near Dyrrachium. Pliny the Elder in his Natural History (4.10) notes that the town lost its autonomy and was designated a colony, Colonia Julia Augusta Byllidensis. Almost four centuries later, its bishop attended the Council of Ephesus in 431 CE. By the 6th c., Byllis had become one of eight towns in the province of New Epirus. Three inscriptions in Greek record the reconstruction of its fortifications by Viktôrinos on behalf of the emperor Justinian, now enclosing a much-reduced area of only 11 ha. Soon afterward, in the later 6th c., the town was abandoned. It is thought that its citizens moved to Ballsh, 8 km to the west, where a church – perhaps a monastery – had existed since the 5th c. CE. The ancient site of Byllis was never reoccupied, although some of its stone was robbed in Ottoman times for construction elsewhere in the region (20).

The site was first described by the early 19th-c. antiquarians F. C. H. L. Pouqueville and W. M. Leake, before any investigations were made. The first survey and trial excavations at Byllis (of tower V in the Late Antique fortifications and the orchestra of the theater) were made during the winter of 1917–18 by C. Praschniker.Footnote ² Fifty years later, in 1974–76, the Albanian archaeologists N. Ceka and L. Papajani began archaeological investigations here, and then the former pursued large excavations with S. Muçaj from the summer of 1978.

This is the first volume of reports of the excavations made by Ceka and Muçaj, which were pursued after 1991 in collaboration with a French mission led by J.-P. Sodini. Since 2003 the collaborative excavations have been led by P. Chevalier and S. Muçaj. Future volumes are promised that will describe the excavations of the cathedral (Basilica B), the bishop's quarter, Basilica E, and many smaller excavations within the Byzantine town, as well as providing full details of the diverse finds. A full discussion of the archaeology is also signaled to appear before long.

This volume begins with a description of Byllis's location and the history of research at the site by the Albanian and Franco-Albanian teams. This chapter includes an introduction to four neighboring Illyrian fortresses, at Rabie, Cfir, Margëlliç, and Gurzezë. It continues with a historical outline of the city of Byllis. Chapter 2 describes the town's system of fortifications. The first survey by Praschniker has been elegantly updated and phased into five iterations of construction. These Hellenistic, Roman, and Late Antique phases – some including flanking towers – are described in significant detail with plentiful plans and photographs. The third phase, belonging to Late Antiquity, amounts to a significant reduction of the town, with a new inner and eastern wall inserted to defend the old (Hellenistic and Roman) civic center, which was concentrated in the western part of the enclosed triangular area. These fortifications include four major foundation inscriptions, which are published and discussed in detail. The authors ascribe this wall to Viktôrinos on the orders of Justinian and, on the basis of coins from levels with burning within the town, to the early 550s, following a Slavic incursion.

Chapter 3 is a full description of the archaeology and architecture of Basilica A. This large basilica was located on the west end of the Hellenistic stoa. It had two substantive phases: the first with a narthex, to which was added an exonarthex in the second phase. Traces of a fire were discovered below a fall of tiles. Within this level were three folles of Justinian minted in the 540s. Basilica A is notable for its polychrome mosaics in the chancel, nave, north aisle, and north transept. The many mosaic images include birds, fowl, and exotic and wild animals. One striking frontal image is of a peacock, similar to those known in churches of this period at Corfu and Nicopolis, as well as Bulla Regia. Apart from a rich range of stone church furniture, the finds included 173 Later Roman coins from Carthage, Ravenna, Rome, and Ticinum, as well as from Antioch, Constantinople, and Thessalonica.

Chapter 4 is devoted to Basilica C, which lies southwest of Basilica A. This was built over the atrium of an earlier house. Like Basilica A at its zenith, it had a narthex and exonarthex, as well as ancillary buildings for either accommodation or storage. The basilica was paved with mosaics in the chancel sanctuary, in a room to the south of the chancel, and in a room to the south of the narthex. The mosaic pavement in the sanctuary includes a line of figures, presumably donors. Praisos, Bishop of Byllis in the Justinianic period, is cited in a Greek inscription in this pavement. The main nave was floored with paving slabs, while the floors of the aisles and narthex were made of beaten earth. The excavations produced 25 Late Antique coins from Carthage and Constantinople.

Chapter 5 is a report on Basilica D, outside the north wall of the Late Antique fortifications. This basilica possessed a large atrium in its final phase. There was also a range of buildings flanking its south side, which were not excavated. The church had a marble floor in the sanctuary, with mosaic pavements in the nave, north aisle, narthex, and the north annex to the narthex, as well as possibly the south annex to the narthex. Bishop Praisos is again referred to in a prominent inscription in front of the prostôon. The excavations produced seven Late Antique coins.

Chapter 6 amounts to a short study of the contextualized ceramic finds (with ample illustrations) discovered in the excavations by M. Bonifay and Y. Cerova.

Finally, two and a half pages of conclusions complete this handsome volume. These conclusions focus entirely upon the Late Antique town. The Justinianic fortifications (termed Byllis III) erected by his architect, Viktôrinos, were strengthened with six rectangular towers. The walls and flanking towers contain spolia from earlier Hellenistic buildings, notably from the theater and stoa. As mentioned above, at 11 ha in area with five narrow gates, the town had shrunk to a third of the fortified area from its zenith in Hellenistic times (Byllis I). (The refurbishment in the Early Roman period is defined as Byllis II.)

The authors conclude that the churches are relatively homogeneous in form. Any architectural differences between them were largely determined by their individual topographical circumstances. Basilica A, for example, is the largest of the three churches considered in this report. Its northern extent was delimited by the ancient decumanus. It had no atrium to the west, but an exonarthex. By contrast, Basilica C was much smaller, with initially a small narthex. Basilica D, outside the north wall, had space for a substantial atrium to the west. According to the authors, the construction techniques closely follow those employed elsewhere in the Balkans.

Thanks to the many fragments of local limestone furnishings, the interior spaces of the basilicas have been reconstructed. The arrangements are consistent with other Epirote churches, with the nave separated from the aisles. The pavements dedicated by Bishop Praisos appear to be the work of an atelier in the region of Ohrid (Lychnidos). There is a predilection for geometrical frames and multiple panels colorfully and vividly depicting animals and birds. The carpet of images was clearly intended to serve the ceremonies.

The ceramics from the excavations confirm both long-distance and local networks. Of these, the most interesting discovery is of pots that had held locally sourced bitumen. On the basis of the numismatic and ceramic evidence, the Late Antique town flourished briefly and then was abandoned during the second quarter of the 6th c.

The volume is a commendable report, with outstanding architectural reconstructions of each of the basilicas, excellent color photographs of the mosaic floors, and useful inclusion of the objects found associated with each of the churches. The only elements missing are excavated elevations (sections). It is a small criticism of an otherwise admirable publication.

When all five basilicas have been fully published, it will be the occasion to review their micro-histories as monuments, the intricacies of their floor decorations, and the prodigious quantity of carved limestone furniture. Two issues, however, stand out already from this first report. First, Byllis in Late Antiquity was clearly dominated by its constellation of churches. The exceptional role of the Church in this reduced fortified area in Late Antiquity merits discussion. Second, the chronology of Late Antique monumental apogee and abandonment is strikingly concentrated in a very short period. With this publication it is now clear that these two discoveries are not exceptional in Epirus. As a result, a larger historical picture – first described by W. Bowden nearly two decades ago – is cogently supported by a great deal of recently published archaeological evidence.Footnote ³

The Late Antique reduction of the fortified area of Byllis to a third of the enclosed area in the Hellenistic and Roman periods is far from unique. The Late Antique fortifications of Nicopolis, Augustus's victory city, in southern Epirus likewise enclosed a much-reduced area that was also dominated by a constellation of sprawling churches, including an episcopal complex.Footnote ⁴ Phoinike, too, was reduced to a small fraction of its immense Republican to Mid-Roman size.Footnote ⁵ By contrast, the Late Antique fortifications of Butrint broadly followed the line of the Hellenistic defenses. All the evidence points to a port that retained much the same footprint that it had possessed at its apogee in the Early Imperial period.Footnote ⁶ At nearby Saranda (Onchesmos) the Late Antique fortifications similarly followed those of the Earlier Roman period. At this somewhat smaller town, though, there was a double line of walls (a proteichisma), surely an indication of the town's projection of its status as a portal for the nearby sanctuary of Santi Quaranta.Footnote ⁷

Of all these Epirote Late Antique fortifications, Byllis's are the only ones dated by inscriptions – commemorating their construction by a certain Viktôrinos on behalf of the emperor Justinian. All the other fortifications are ascribed on archaeological evidence – largely ceramics – to the earlier 6th c.Footnote ⁸ The excavators of Byllis attribute the third iteration of fortifications to the growing 6th-c. threat posed by extra-regional forces (18). However, who posed this threat in the first half of the 6th c., a generation before the Slavic attacks, is a matter of speculation. Like the earlier Hellenistic and Roman defenses at Byllis and other Epirote towns, these powerful Late Antique walls may be as readily explained as expressions of local martial display, as well as indices of status.

What is especially striking at both Byllis and Butrint – the former in the Epirote uplands, the latter on the coast – is the exceptional presence of the Church. Almost half the walled area of Byllis was filled by churches and their associated buildings – intended, it is presumed, for accommodation and storage. Much the same is true at 6th-c. Butrint, where in the course of extensive excavations inside the walled area, as well as in the suburb and beyond, the presence of the Church is strikingly conspicuous. Besides Butrint's Great Basilica, several smaller churches have been identified inside the walls and at the postern gate, as well as in the suburb and on the shore of Lake Butrint.Footnote ⁹ Probable secular buildings have been only identified within the area once occupied by the old Roman forum and within the ruins of the 5th-c. Triconch Palace.Footnote ¹⁰ Importation of goods, especially from the eastern Mediterranean, occurred on a notable scale, while small-scale productive activities – to judge from the workshops in the Triconch ruins – were practiced by households.

In short, these Epirote towns now accommodated conspicuous ecclesiastical communities and less visible numbers of lay citizens occupying dwellings that, to judge from the Butrint examples, were far from eye-catching in architectural terms.Footnote ¹¹ Certainly, further excavations of 6th-c. dwellings in these towns would paradoxically shed light on the apparently increased centrality of the Church in civic life. Excavation might also reveal whether these churches were reached by colonnaded, processional thoroughfares, as occur in the Eastern Empire.Footnote ¹² This said, the towns of Epirus were undoubtedly experiencing the same transition of power – from a rich aristocracy to a clerical leadership invested in the notion of salvation in the afterlife – that has been identified in other regions of the Central and Western Empire in the 6th c.Footnote ¹³

As Peter Brown and Ian Wood have argued, this was a major step toward an early medieval “temple” society that hastened the end of the Roman world.Footnote ¹⁴ Both historians have drawn attention to the challenge posed to the Later Roman economy by the massive concentration of wealth in the Church.Footnote ¹⁵ Where were the taxes to come from to sustain the armies and bureaucracies of the era, let alone the building of defenses? In a world of “democratized” clergy, as Brown puts it, who was to build, or make pots and tools, let alone manage livestock and cereal strategies? Certainly, too little is known about the economy of the Later Roman Church.Footnote ¹⁶ At Byllis, for example, Basilicas A–E possessed buildings that may have been used for storage and production. At Butrint, likewise, the buildings associated with the bridgehead basilica in the suburb had buildings that might have been used for storage. And at the erstwhile maritime villa at Diaporit, a silo was excavated and several large dolia were found.

This report on Byllis's Basilicas A, C, and D also shows that each was remodeled after its construction. Particular emphasis in each case was on ever-grander entrances. Associated with this may have been the increasing embellishment of the church interiors with extensive polychrome mosaic pavements and ornamented furniture. Increasing display mattered.

Short-lived overinvestment in conspicuous consumption, rather than the arrival of the Slavs, may explain the sudden abandonment of all these towns. The archaeology of this abandonment phase has not yet been presented for Byllis. The apparently unworn pavements, for example, and the limited evidence of burning and destruction appear to rule out desertion following a cataclysmic event. On the other hand, evidence of ruralization within a ruined urban fabric in the mid- to later 6th c., such as was found at Butrint, is not mentioned in this report.Footnote ¹⁷ (Is this a result of the excavators overlooking these levels, as remains of later 6th-c. dwellings of an insubstantial character were cited by early excavators of the town?Footnote ¹⁸) Strikingly absent are the intramural burials that occur in significant numbers at Butrint, as well as in its suburb and at the ecclesiastical complex at Diaporit. Only one such late burial is published in this volume on Byllis: that of a child from the exonarthex of Basilica C (311–13). Abandonment due to the Justinianic plague, for example, seems unlikely given the absence of graves. This suggests that the reign of Justinian may have marked not only the apogee of the Late Antique town but also its sudden abandonment. Tacit support for this may be the absence of later 6th-c. ceramics and coins published in this volume.Footnote ¹⁹

On balance, Bowden's compelling argument that investment in church-building without return in a place that had once prized a rich pastoral economy perhaps best explains the end of Byllis.Footnote ²⁰ Rather than waiting for the arrival of the barbarians, as this first volume devoted to the Byllis excavations suggests, an explanation for its abandonment may lie in the increasing investment in display by the Church as opposed to by its secular citizens, and concurrently an unsustainable economy created by clerics principally promising salvation in the afterlife. As is now clear, Epirus was just one of many regions in the central Mediterranean to experience this fate over the course of the 6th and earlier 7th c. Judging from this report on part of the excavations, Byllis is a compelling inland illustration of this ultimately ubiquitous historical episode. Future reports in combination with this admirable first one will come to hold an important place in demonstrating the decline of the Roman Empire, as the social and economic nuances of the town's basilicas and their associated buildings are revealed and analyzed.