Research Article

Urban life in early Islamic Morocco: new light from the excavations at Walīla (Roman Volubilis)

Authors: , ,

Abstract

This article presents a research update from the INSAP-UCL excavations at the UNESCO site of medieval Walīla (Roman Volubilis), Morocco. Though the site is best known as the Roman city of Volubilis, in the Middle Ages, and by then called Walīla, it took on a new importance as a Berber centre, the probable locale of an Umayyad or Abbasid garrison and the capital of Idrīs I, the founder of one of the earliest Islamic states in Morocco. It is the only site in Morocco – and in North Africa more generally – where excavations have uncovered substantial evidence of eighth-century urbanism. As such, Walīla provides an exceptional opportunity to investigate the nature of a Berber town and its transformation into, for a short time, the centre of one of the earliest Islamic states in North Africa. This article presents a summary of our results from the four seasons of fieldwork that have taken place so far, between 2018 and 2022.

Keywords: Islamic archaeology, medieval archaeology, North Africa, Morocco, urbanism

How to Cite: Fentress, E. , Fenwick, C. & Limane, H. (2022) “Urban life in early Islamic Morocco: new light from the excavations at Walīla (Roman Volubilis)”, Archaeology International. 25(1). doi: https://doi.org/10.14324/111.444.ai.2022.08

Introduction

The rule of the caliphate lasted barely a hundred years in North Africa. In 740–1, the so-called ‘Kharijite revolt’ broke out in Morocco, rapidly spreading to Iberia as well as the rest of North Africa. In its wake, new Muslim states emerged as rivals to the Umayyad and then the Abbasid caliphate. In Morocco and western Algeria, an important conduit for trans-Saharan gold, the local elites embraced variants of Islam and created their own dynastic states, at the same time urbanising earlier oasis, sedentary and semi-transhumant populations. Sometimes, as with the Idrisids at Walīla, the Salihids at Nakuūr or the Rustamids based at Tāhart, incomers from the East were involved in these nascent processes of state formation. The emergence of these rival states coincides with a phase of intense urbanisation and an economic boom in Morocco during the ninth to eleventh centuries: new cities were built, old settlements re-occupied and new trade links developed with the Atlantic, the Sahara and the Mediterranean (Cressier 1998; Fenwick 2020a).

What little is known about Umayyad and Abbasid activity in the far West and the first Islamic successor states comes from much later historical sources that are imbued with a strong sense of religious and political propaganda. Small-scale archaeological interventions at several of the new dynastic foundations established in the ninth and tenth centuries, such as Nakūr, Sijilmāsa and al-Baṣra, have begun to fill in our understanding of Islamic urbanism (Cressier 2017; Messier and Miller 2015; Benco 2004). At the same time the archaeology of the medieval phases of Romano-Moroccan sites (Volubilis, Lixus, Rirha) suggests that many earlier sites were re-occupied from the ninth century onwards (Akerraz 1992; Callegarin et al. 2016). Taken collectively, these excavations hint tantalisingly at a radical transformation of the Moroccan landscape from the ninth century (and perhaps earlier) comprising the appearance of new forms of architecture (mosques, hammams, courtyard houses), the spread of Islam and Muslim dietary and burial practices and the introduction of new crops, husbandry practices, manufacturing techniques and ceramic forms. However, we still lack a holistic understanding of urbanism and daily life, particularly for the pivotal eighth century when many new groups were moving in and around this region.

The site of Walīla, Roman Volubilis, is an ideal place to explore these issues. The town of Volubilis is one of the most spectacular Roman sites in the Mediterranean and the most visited archaeological site in Morocco. However, its medieval history as the town of Walīla has often been neglected, despite its importance in the history of Islamic North Africa. Walīla was a Berber centre, the probable locale of an Umayyad or Abbasid garrison, and the capital of Idrīs I (r. 788–91), the founder of one of the earliest Islamic states in Morocco. The Idrisid state of northern Morocco ruled by Idrīs and his sons was arguably the most significant of the rival states – and the greatest perceived threat – to the Abbasid caliphate (Fenwick 2022). Excavations between 1887 and 2015 aimed at reaching Roman levels have produced the largest corpus of medieval burials and medieval finds (coins, glass, metal and ceramics) anywhere in North Africa. Moreover, it is the only site in Morocco – and in North Africa more generally – where excavations have uncovered substantial evidence of eighth-century urbanism (Fentress and Limane 2018; Fenwick 2013). As such, it provides an exceptional opportunity to investigate the nature of a Berber town and its transformation into, for a short time, the centre of one of the earliest Islamic states in North Africa.

We present here a research update from the INSAP-UCL excavations at the UNESCO site of Volubilis, Morocco, where we are exploring how this urban settlement changed and adapted through this pivotal period. Our knowledge of the medieval town is still extremely partial, based primarily on the discovery of a series of medieval coin hoards in the early twentieth century (Eustache 1956; 1966); excavations by Aomar Akerraz in the 1980s, which uncovered a sixth-century fortification wall and an eighth-century workshop with kilns (Akerraz 1983; 1998); and a first phase of INSAP-UCL excavations (2000–4), which revealed a section of the Berber town, within the Roman walls, and what were probably the headquarters of Idrīs I, just outside the walls (Fentress and Limane 2018). The new project aims further to reconstruct the urban development of this North African town in the Islamic period through excavation of multiple zones in the town and to understand the diet, health, lifestyle, origins and mobility of its medieval inhabitants through environmental, isotopic and aDNA analysis. Equally, it seeks to train Moroccan and British students collectively in the latest archaeological techniques and to foster a new generation of Islamic archaeologists in the two countries. This article covers the four seasons of fieldwork that have taken place so far, between 2018 and 2022.

Volubilis-Walīla: open questions and prior research

Volubilis was founded at some point in the third century bce, probably by members of the Mauretanian elite, and became the provincial capital of the new Roman province of Mauretania Tingitana after its annexation in the first century ce. Positioned on the lower slopes of the foothills of the Zerhoun mountains, then (as today) covered with olives, overlooking the wheat-growing plain of the Gharb, the town sits on a low hill with a river (the Oued Khoumane) flowing past at its base. Over 200 years of Roman rule, Volubilis acquired the standard public architecture of a Roman town: a forum, baths, a basilica and a triumphal arch, as well as houses decorated with splendid mosaics and bronze statuary (Rebuffat 1965). The Roman administration departed around 285 ce, but this was not the end of the town. The houses continue to be decorated by fine mosaics through the fourth century and African Red Slip ceramics, Roman coins and imported metalwork are found until the first quarter of the fifth century. The town seems to have been hit by one of the many earthquakes of the early fifth century and was subsequently deserted for a time, at least in part (Fentress and Limane 2018).

By the late sixth century, the town was re-occupied. A small group of tombstones written in Latin and dating from the late sixth and seventh centuries reveals the existence of a Christian community that still used the Roman provincial calendar (Thouvenot 1969). One inscription of the year 655 even names a vicepraepositus, probably referring to the leader of a Berber kingdom who was drawing on Roman symbols and patterns of rule. This community lived in the western third of the city, protected by a new north–south fortification wall that joined two segments of the earlier Roman walls to enclose around 18 hectares (Akerraz 1998). To the east of the new rampart, much of the Roman monumental centre was given over to burials and a cemetery with Latin tombstones dating from the sixth to seventh centuries was located around the Arch of Caracalla. The new town, although reduced, was a large and important agricultural settlement, and its name reverted to what was probably the original Berber form, ‘Walīla’ – a toponym likely derived from the Berber word for Oleanders.

After the Muslim conquest of Morocco in the early eighth century, we know that it was the residence of the Berber Awraba tribe who had reportedly converted to Islam (presumably, but not certainly, the same group who had settled there in the sixth and seventh centuries ce). The town of the Awraba was a substantial agro-settlement that minted its own coins, produced its own ceramics and was linked into larger networks of trade and exchange. The residents of the town seem to have lived in rather simple scattered housing within the late walls. The first INSAP-UCL excavations uncovered a series of small two-room houses on the edge of the town (Area D) dating to the late seventh to ninth centuries. The houses were divided into domestic space and stabling/workrooms, with storage silos outside (Fentress and Limane 2018, 74–81).

Probably in the early eighth century, the so-called ‘Arab quarter’ was established outside the early Roman walls near the river. This area was excavated after a fashion by French archaeologists in the 1950s and 1960s, but their findings were never published. The excavations revealed dense medieval housing, an irrigation canal and a building with 12 columns and five doors that may have been a mosque, as well as a gold dīnār, two large silver dirham hoards and hundreds of copper fulūs (Eustache 1956). Copper coins with Arabic legends were minted at the town, most probably in the late eighth century. Some were inscribed simply with the town’s name ‘Walīla’ while others bore the names of otherwise unknown individuals (Rashīd b. Kadīm, Muḥammad b. Khalifa, Ibrāhīm al-Ḥasanī). These have variously been identified as Abbasid generals (Eustache 1956) or leaders of an independent Berber city-state (El Harrif 2018). An early conversion of Walīla’s inhabitants to Islam is supported by an absence of faunal evidence for pork consumption in seventh- and eighth-century contexts across the site (King 2018), as well as burials laid out in accordance with Muslim funerary rites in the latest layers of the late antique cemetery near the Arch of Caracalla (Akerraz 1998, 297, 303).

It is at this point that the site became pivotal in the history of Morocco, for it was here that Idrīs I (745–91), a descendant of the Prophet, established himself during his brief reign. The later Arabic literary sources relate that he arrived in 787–8 fleeing persecution, married an Awraba woman and was proclaimed imam. With his Berber allies he conquered much of northern Morocco and established a state in this region (see Fenwick 2022). Walīla briefly became the headquarters of the Idrisid state and Idrīs and his descendants minted silver and copper coins there. Outside the walls and to the south of the ‘Arab quarter’ (Area B), the first INSAP-UCL excavations uncovered an extra-mural complex of three courtyard buildings and a hammam (bath building) constructed in the late eighth to early ninth century. This may have served as the residence of Idrīs I and his son Idrīs II (Fentress and Limane 2018, 82–102).

Walīla was to be only briefly the centre of the Idrisid state. Idrīs I founded the city of Fès and his son Idrīs II, on reaching his majority, moved the court there. Walīla continued to be an important Idrisid centre before apparently being abandoned by the tenth century (Ibn ‘Idhārī 1948–51, 82–3). The town disappears from the historical sources thereafter, but there is archaeological evidence of a significant fourteenth-century occupation in the Merinid period – perhaps related to the new importance of the Idrisid heritage for legitimising the Merinid right to rule (Beck 1989).

The current project is excavating in the different medieval quarters in eighth-century Walīla to understand the nature of early medieval urbanism at the town. Our first area of focus is the centre of the ‘Berber’ medieval town (Figure 1, Area A), where survey showed several substantial medieval buildings (30 × 30 m) in the centre. These look very different from the small one- or two-roomed houses on the periphery of the settlement that were uncovered in the excavations of 2000 to 2005, and were abandoned by the tenth century (Figure 1, Area D). Our second area is the so-called ‘Arab quarter’ (Figure 1, Area E), outside the sixth-century ramparts near the wadi which has never been excavated stratigraphically. Our hypothesis is that this is an Umayyad or Abbasid extra-mural settlement, similar to those identified elsewhere in the Middle East, North Africa and Spain (Whitcomb 1994; Fenwick 2020b). The aim of excavating in these different zones is to explore whether there are differences in architecture, lifeways and consumption practices between the two quarters, and whether these may be related to ethnicity, wealth or function.

Figure 1
Figure 1

Composite plan of medieval Walīla showing the location of 2000–5 excavations (Areas B and D) and the 2018–22 excavations (Areas A and E) (Source: Corisande Fenwick)

Area A: the ‘Berber town’

The excavation of Area A was aimed at establishing whether the simple two-room housing discovered in Area D during the 2000–5 campaign was an anomaly or characteristic of domestic housing in the town. The site was also chosen as an example of a series of large enclosures in the area (Figure 2). These are large, roughly square enclosures, visible on the surface as double rows of squared blocks, all re-used Roman blocks (spolia), with an infill of smaller stones. Each has only one course and may have been constructed in rammed earth in their upper levels. They clearly represent separate lots, but of what period? Clearance and preliminary excavation immediately revealed that they were the latest features on the site, dating in fact to the Merinid period in the fourteenth century. They served, however, as a convenient limit of excavation.

Figure 2
Figure 2

Area A, UAV photograph with phasing (Source: photograph by Hallvard Indgjerd and Elizabeth Fentress)

Although excavation is not yet complete, we have now managed to reveal a general picture of the development of the area from the seventh century (Figure 2). The damage caused by the fifth-century earthquake postulated in the earlier excavations is amply documented here. Almost 3 metres of destruction deposit, generally the thick yellow clay deriving from earthen walls, separate late antique layers from the Roman buildings beneath it. The earliest features so far found cut into this destruction deposit are a group of three simple kilns, without raised floors, built into the slope of the hill (Figure 3). Shaped like figures of eight, they seem to have produced water jars, bowls and dishes in forms similar to earlier Roman productions. Above them a nearby building with rooms H-S-T in the northeast corner of the site (see Figure 2) seems to be a courtyard structure with rooms around two sides. Relatively well built, it has a small vestibule to the west, creating a bent entrance to the courtyard. This structure awaits excavation next season, along with confirmation that it is of a similar date to the kilns, perhaps the seventh century.

Figure 3
Figure 3

The kilns, from the north (Source: photograph by Asmae El-Kacimi)

Some time after the abandonment of the kilns, a large rectangular house was cut into the slope. The building was terraced along a north–south line and possessed a courtyard, delimited, perhaps, by a boundary wall to the south. The house was divided into two rooms (Figure 4). The larger, B, a multipurpose room, was probably used by the family. It had a hearth and a pair of postholes in front of the door that may have supported an upright loom. The smaller room, on the east side, was used for storage and covered by a loft on which jars and other pottery were stored. From the smaller room, a door led out to the courtyard to the north, into which was dug a substantial silo. In the courtyard to the south was found a cesspit containing mineralised seeds, identified by Dr Ruth Pelling as fig, grape and watermelon. This pit cut the remains of the kilns and the construction of the building covered one of them.

Figure 4
Figure 4

House A–B, from the north-west, in 2018. The two postholes for the upright loom are visible in the foreground (Source: photograph by Elizabeth Fentress)

The compound seems to be associated with an unusual underground workshop aligned with the building. The structure was created by cutting 2 metres down into the Roman destruction levels. A retaining wall, built against its west side, seems to have served as a staircase, although access using a ladder was also possible. The workspace was arranged with two earthen counters on either side of a lower area used for movement in the space. To the south, the counter was cut by a large pit, filled with ashy earth. To the north, a substantial circular hearth or furnace was cut into the counter, filled with small, burned stones and ash. To the east of it a succession of two pits lined with white clay cut the counter; they appear to have been intended for holding water. Some 300 fragments of pellet moulds were found in the destruction layers which fill this space: these are heavily burned and vitrified on their undersides.

Although we do not have enough rims or joins to reconstruct a whole mould, a tentative reconstruction of one is proposed in Figure 5. Preliminary analysis by Raluca Lazarescu (UCL) has identified copper and silver prills in the individual pockets of the moulds. This demonstrates that they were used to produce pellets using a measured amount of powdered metal, melted over the hearth. Coins must have been struck from these pellets, but it is still unclear where this would have taken place. Such coin moulds are known from Iron Age Britain (Landon 2016) and are also found in early medieval contexts elsewhere in Morocco (El Ajlaoui 1994), as well as in Mali (Nixon et al. 2011) and Pakistan (Khan 1990). The ceramics suggest that this activity was taking place around the middle of the eighth century, the time at which Walīla had begun to mint its own coins. Continued use of the space is shown by a build-up of clay and ash, as the first hearth was replaced by later ones.

Figure 5
Figure 5

Reconstruction of a pellet mould (Source: photograph by Elizabeth Fentress)

At some point the house burned down – its loft collapsing on the floor below and its walls falling outwards. The debris from the burning of the house was used to backfill the workshop and the silo, both of which contain large, squared blocks in their upper layers. In the debris from the loft was found an incense burner and a clay bottle inscribed with a character resembling a Tifinagh letter from the Berber alphabet.

However, this was not the end of occupation in this area, which appears to have begun again shortly afterwards, probably in the late eighth century. The door that led from the old house, now in ruins, was blocked, and a new wall built perpendicular to it, creating a new room. The new building does not seem to have been domestic: its floor was composed of large pebbles around a drain through the wall that may imply some sort of use with water, perhaps for laundry, as ethnographic parallels suggest. The space once occupied by the workshop was now used as a courtyard, onto which was built an oven. The surface contained a coin of Idrīs I (777–89 ce), as did the destruction of the vault of the oven.

The destruction of this building seems to date to the ninth century. There is no further occupation until the construction of the large enclosure in the fourteenth century when Morocco was under Merinid rule. One house dates to this latter period, on the southern edge of the trench. This had a careful plaster floor on which sat a large dish, certainly fourteenth century in date. However, there is little in the way of domestic refuse and nothing to suggest a long occupation of the space. The aim of this curious settlement, with its laboriously constructed precinct walls, remains to be determined. It may be related to the discovery of the tomb of Idrīs I, which the fourteenth-century author al-Jaznā’ī (1923, 15) tells us took place in the year 1340, causing vast crowds to flock to Walīla. In any case, it does not appear that the settlement survived very long, and the body of Idrīs is currently said to lie in the great mausoleum at Moulay Idrīs, 3 km away.

Area E: the ‘Arab quarter’

Our second focus is the zone known as the ‘quartier arabe’ immediately outside the West Gate of the Roman wall and on the flat ground in front of the river (Figure 6). Our aim here was to plan the dense medieval housing and possible mosque exposed by the unpublished French excavations in the 1950s, to clarify through targeted excavation the occupational history of the sector and to determine whether there is evidence for an Arab settlement or garrison. Our excavation strategy focuses on areas of intact stratigraphy left by earlier French excavations within a large sector in the East. As a result, it is not easy to reconstruct the full extent of buildings or exterior spaces in the same way as is possible in Area A. Nonetheless, we have so far identified five phases from the early Roman period through to the Middle Ages.

Figure 6
Figure 6

Area E, UAV photograph with phasing (Source: photograph by Corisande Fenwick)

During the life of the Roman town, this area lay outside the massive Antonine Walls and the two-bay West Gate with its dedicatory inscription of 168/9. It was occupied by a substantial Roman cemetery with mausolea and cupolae tombs of the second and third centuries: the bases and substructures of some of these are still visible. A series of cobbled surfaces with ceramics and coins of the first century ce marked the road leading out of the city and across some kind of a bridge or crossing over the Oued Khoumane. The presence of hearths/furnaces with iron and copper slag and vitreous waste suggest that some industrial activity may also have taken place here in the early Roman period.

In the sixth or seventh century, presumably contemporary with the re-occupation of the site after the earthquake, the area near the gate seems to have been used as a midden for the Late Roman town. Large amounts of animal bones (including those of pigs), imported Tunisian lamps and Spanish and Eastern amphorae of seventh-century forms suggest that Volubilis was far better integrated into Mediterranean trading networks than previously believed.

In the next phase, probably in the late seventh or early eighth century, a settlement was established in this area. Several large houses with courtyard spaces have been identified so far, some separated by narrow alleyways. At least one house was inserted between the two bays of the Roman gate, creating a far narrower main street in place of the wider early Roman one. These houses are built in a roughly opus africanum technique re-employing architraves, columns, tombstones and other stones from the Roman cemetery, while internal rooms are characterised by fine plaster floors. Cooking seems to have taken place in the courtyards: in Trench 3 a small circular bread oven was found in the centre of the open space, as were postholes to support a roof.

One building is on a much larger scale than the others: it consists of a large room at least 15 m long, facing the main street. Behind it is a large courtyard and well. Based on its size and orientation, we are evaluating the possibility that it was originally a mosque (Figure 7). Inside it very little intact stratigraphy survives, but a series of partially preserved plaster floors and substantial stone pavements are visible in the sections. The contexts of this phase are characterised by large numbers of pre-Idrisid copper coins (fulūs), new forms of ceramics, including filter jars, and a much wider variety of crop types than were found in Area A. A glass weight inscribed in Kufic Arabic of possible Umayyad date was also found. A substantial burn layer dating to the mid-eighth century closed this phase and is found across the excavated area. It contained a glass-paste ring setting inscribed bismillah (‘in the name of God’) (Figure 8). It is very similar to a silver seal-ring discovered in 1933 (perhaps in a grave) to the west of the ‘maison au chien’ that is inscribed ‘Allah is enough for me. He is the best guarantee’ (Alaoui et al. 2014). These seal-rings are characteristic of early eighth-century cemeteries in al-Andalus (De Miguel Ibáñez 2016, 64–9), but these are the earliest discovered in Morocco to date. Taken together, this supports our interpretation of this new urban zone with its different housing types as a settlement of Arabs or Muslims living outside the town walls in the first half of the eighth century.

Figure 7
Figure 7

The large building with associated courtyard, UAV photograph (Source: photograph by Corisande Fenwick)

Figure 8
Figure 8

Glass-paste ring setting reading ‘bismillah’ found in the destruction layer of the first Islamic phase (Source: photograph by Ruth Pelling)

This first settlement seems to have met a violent end in the middle of the eighth century, as extensive traces of burning are found throughout our trenches. In the next phase, dated securely to the late eighth to ninth century, and apparently following a period of abandonment, the quarter expanded significantly. New multi-room dwellings were built in a drystone masonry, while existing houses seem to be subdivided or extended. One of these, next to the partially standing fortification wall, had a bent-axis entrance characteristic of Islamic houses. Several of the houses have wells cut into their courtyards. Evidence of artisanal activities has also been found, including a furnace for iron and copper smelting, several pits used for dumping ash and a tank for fulling or processing agricultural goods. A significant number of worked bone tools were found in association with the furnace (Figure 9). Similar tools have been identified in medieval and early modern metal workshops in Morocco and Iberia and are interpreted as bone anvils for saw-tooth sickles (Benco et al. 2002; Grau-Sologestoa 2012; Belatik et al. 2020).

Figure 9
Figure 9

Example of worked camelid bone identified by Marie Middleton: proximal camel radius from SU 6228, Area A, ID no. 3188 (Source: photographs by Hallvard Indgjerd)

This phase also ended abruptly; thick burning and collapse layers rich in coins and locally produced ceramics (cookwares, jars, tablewares) have been found across the site. The coins are predominantly of Idrisid date and include several silver dirhams and many bronze fulūs minted under Muhammad b. Idrīs (828–36). These not only attest to the wealth and prosperity of the inhabitants in the extra-mural quarter, but also demonstrate that this area continued to be occupied in the Idrisid period, perhaps as part of a significant expansion of the town.

A final phase of occupation comes after a substantial period of abandonment and repeated flooding events. This consists of ephemeral housing scattered across the entire area, mostly destroyed by the French excavations in the 1950s. It may relate to the fourteenth-century Merinid activity identified elsewhere on the site or be yet later still.

New light on an early medieval settlement

Volubilis-Walīla offers new interpretive insights into urbanism in the late antique and Islamic West. Here the city continued to thrive after the Roman administration left in 285 ce, but it was eventually abandoned as early as the fifth century – a cautionary tale for archaeologists who seek to tie abandonment of classical cities to political events or for those who see medieval settlement as proof of continuous occupation. This first abandonment was short-lived and by the late sixth century the site was partially re-occupied by Berber groups. The small simple housing of the Berber town was entirely different to the large insulae of the classical town. However, this was still a substantial settlement with a (perhaps Christian) elite who produced their own pottery, had access to imported luxury goods, used Roman titles and still dated their year by the Roman provincial calendar. Houses were widely spaced, with ample courtyards in which activities such as pottery production could be found. Urban settlements of this size were rare in late antique Morocco. Walīla therefore provides a tantalising glimpse into a nascent but as yet poorly understood Berber urbanism of the late sixth and seventh centuries.

This Berber town continued to thrive in the early eighth century during the Arab conquest and partial rule of Morocco (whatever this may have meant in practice). What emerges from our findings in the ‘Arab quarter’ is a phenomenon paralleled at other sites in North Africa: a reluctance on the part of the newly arrived Arabs to settle within the confines of the existing settlements (Fenwick 2020a, 31–52). The styles of housing and building are radically different inside and outside the walls, as are the finds, palaeobotanical and numismatic records. Coins, in particular, are extremely abundant in the extra-mural settlement and very rare inside the walls. If we are right in identifying the extra-mural quarter at Walīla as a new Arab settlement, perhaps even an Umayyad garrison, it is tempting to associate its violent destruction with the Berber revolts of the 740s. However, the minting of coins at Walīla under its own name and by otherwise unknown individuals with Arab names in probably the late eighth century suggests that the picture is not so clear-cut (El Harrif 2001, 2018). Precise dating of this phase must await the detailed analysis of the ceramics and coins, as well as a targeted radiocarbon dating programme.

Under Idrisid rule, in the late eighth and ninth centuries, the town grew rapidly. Traces of Islamic occupation discovered outside the rampart to the north-west as well as in the extra-mural zone shows that its extent far exceeded that of the Late Roman city. The metal-working and other industrial activities in the extra-mural settlement, as well as the potters’ quarter in the Maison au Compas, would have been related to the new needs of this larger and bustling city, which functioned as the base for Idrīs’s conquest of northern Morocco. However, Walīla’s prosperity proved short-lived. The mid-ninth century date of the second burn layer outside the walls in Area E (the ‘Arab quarter’), the abandonment in Areas A and D inside the walls, as well as in the potters’ quarter in the Maison au Compas, suggests that we may be seeing a site-wide abandonment. Ibn Khurradādhbih (d. 912 ce) provides an explanation for this abandonment that fits very well with the archaeological and numismatic evidence: ‘Muhammad [b. Idrīs] resided at Walīla, the last town in the province of Tanger, and he died there. His descendants moved to Fez and are still there’ (1889: 265–6).

After an interval of several decades, it was briefly re-occupied by a settlement over the ruins of the old Idrisid headquarters (Area B). The site was then abandoned for half a millennium, when a new and substantial, if short-lived, Merinid settlement appears to have occupied much of the site within the walls.

Acknowledgements

The project is a collaboration between the Institut National des Sciences de l’Archéologie et du Patrimoine (INSAP) and the UCL Institute of Archaeology and we remain indebted to Professor Aomar Akerraz and Abdelouahed Ben-Nçer, Directors of INSAP, and Professor Sue Hamilton, Director of the Institute of Archaeology, for their support of the project. We are extremely grateful for the contributions of the specialists and students who participated in the fieldwork campaigns of 2018, 2019, 2021 and 2022: Mohamed Alilou, Leila Araar, Zineb Askaoui, Mustapha Atki, Victoria Amoros-Ruiz, Dirk Booms, Rowena Banerjea, Anna Bekési, Omar Bouka, Fatima Zohra Cadi, Gabriella Carpentiero, Scott Chaussée, Samantha Cox, Fatima Zorha El Harrif, Asmae El Kacimi, Asmae Elqably, Louis Falkingham, Maria Gajewska, Meriam Hansali, Joyce Ho, Guy Hunt, Uzma Haque, Hajar El Khouda, Hallvard Indgjerd Rübner, Hanae Jerjou, Raluca Lazarescu, Basma Mejrihi, Marie Middleton, Imane Qouti, Efthymia Nikita, Tarik Moujoud, Ruth Pelling, Tim Penn, Alice Poletto, Caitlin Sutherland, Abdelilah Sguiri and Paige Steen. We also thank Sandra Bond, David Bone, Jo Dullaghan and Timea Grego for their logistical support.

Funding

The first two seasons of work in 2018 and 2019 were primarily funded by grants from the Barakat Trust and the Gerald Avery Wainwright Foundation. The third season in 2021, delayed due to the Covid-19 pandemic, was funded by grants from the Barakat Trust, the Society of Antiquaries and an Institute of Archaeology Award. A European Research Council (ERC) grant to Corisande Fenwick for the project EVERYDAYISLAM-Becoming Muslim: Cultural Change, Everyday Life and State Formation in early Islamic North Africa (600–1000) (Grant no: 949367) provided additional support for the 2021 excavation season and funded the 2022 study season and subsequent post-field research.

Declarations and conflicts of interest

Research ethics statement

Not applicable to this article.

Consent for publication statement

Not applicable to this article.

Conflicts of interest statement

The authors declare no conflict of interest with this work. All efforts to sufficiently anonymise the authors during peer review of this article have been made. The authors declare no further conflicts with this article.

References

Akerraz, Aomar. (1983).  Note sur l’enceinte tardive de Volubilis.  Bulletin du Comité des travaux historiques et scientifiques 19B : 429–36.

Akerraz, Aomar. (1992).  Lixus, du Bas-Empire à l’Islam.  Collection de l’École française de Rome 116 : 379–85.

Akerraz, Aomar. (1998). Recherches sur les niveaux islamiques de Volubilis In:  Genèse de la ville islamique en al-Andalus et au Maghreb occidental. Cressier, Patrice, García-Arenal, Mercedes Mercedes (eds.),   : 295–304. Madrid: Casa de Velázquez.

Alaoui, Abdellah; Ettahiri, Ahmed S.; Fili, Abdallah. (2014). L’archéologie islamique au Maroc, les acquis et les perspectives In:  Maroc médievale. : 44–7. Paris: Louvre éditions.

al-Jaznā’ī, ‘Alī. (1923).  Zahrat el-âs (La fleur du myrte): Traitant de la fondation de la ville de Fès. Bel, A (trans.),   Algiers: Jules Carbonel.

Beck, Herman L. (1989).  L’image d’Idrīs II, ses descendants de Fās et la politique sẖarīfienne des sultans marīnides: 656–869/1258–1465. Vol. 3 Leiden: Brill.

Belatik, Mohamed; Limane, Hassan; Ettahiri, Ahmed S; Ammar, Hakim; Kacimi, Asmae El; Imrani, Chaimae; Askaoui, Zineb; Benjaafar, Fadwa. (2020).  Découverte d’un quartier de potiers d’époque médiévale à Dār al Barād à Salé.  Bulletin d’Archéologie Marocaine 25 : 153–77.

Benco, Nancy L. (2004).  Anatomy of a Medieval Islamic Town: Al-Basra, Morocco. Oxford: Archaeopress.

Benco, Nancy L; Ettahiri, Ahmed; Loyet, Michelle. (2002).  Worked bone tools: Linking metal artisans and animal processors in medieval Islamic Morocco.  Antiquity 76 (292) : 447–57, DOI: http://dx.doi.org/10.1017/S0003598X00090554

Callegarin, Laurent; Alaoui, Mohamed Kbiri; Ichkhakh, Abdelfattah; Roux, Jean-Claude. (2016).  Rirha: Site antique médiéval du Maroc, Volume IV: Période médiévale islamique (IXe–XVe siècle). Volume 153: Collection de la Casa de Velázquez. Madrid: Casa de Velázquez.

Cressier, Patrice. (1998). Urbanisation, arabisation, islamisation au Maroc du Nord: Quelques remarques depuis l’archéologie In:  Peuplement et arabisation au Magreb occidental: Dialectologie et histoire. Cressier, Patrice, Aguadé, Jordi; Jordi and Vicente, Ángeles Ángeles (eds.),   : 27–39. Madrid: Casa de Velázquez.

Cressier, Patrice. (2017). Nakur: Un émirat rifian pro-omeyyade contemporain des Aghlabides In:  The Aghlabids and their Neighbors: Art and Material Culture in Ninth-Century North Africa. Anderson, Glaire, Fenwick, C; C and Rosser-Owen, M M (eds.),   : 491–513. Leiden: Brill.

De Miguel Ibáñez, María Paz. (2016).  La maqbara de Pamplona (s. VIII): Aportes de la osteoarqueología al conocimiento de la islamización en la Marca Superior. PhD thesis. Universidad de Alicante.

El Ajlaoui, El Moussaoui. (1994).  Mine d’argent d’Imidar et la question de Todgha (VIIIe–Xe siècle): vers une hypothèse.  Hespéris Tamuda 32 (1) : 11–34.

El Harrif, Fatima-Zohra. (2001). Monnaies islamiques trouvées à Volubilis: Liberté locale et pouvoir abbaside In:  Actes des Ières journées nationales d’Archéologie et du patrimoine, Rabat, 1–4 juillet 1998. : 142–59. Rabat: Société Marocaine d’Archéologie et du Patrimoine.

El Harrif, Fatima-Zohra. (2018). Les monnaies In:  Volubilis après Rome: Les fouilles UCL/INSAP, 2000–2005. Fentress, Elizabeth, Limane, Hassan Hassan (eds.),   : 308–38. Leiden: Brill.

Eustache, Daniel. (1956).  Monnaies musulmanes trouvées à Volubilis.  Hespéris 43 : 133–95.

Eustache, Daniel. (1966).  Monnaies musulmanes trouvées dans la Maison du Compas, Volubilis.  Bulletin d’Archéologie Marocaine 6 : 349–64.

Fentress, Elizabeth; Limane, H. (2018).  Fouilles de Volubilis 2000–2005. Leiden: Brill.

Fenwick, Corisande. (2013).  From Africa to Ifrīqiya: Settlement and society in early medieval North Africa (650–800).  Al-Masāq 25 (1) : 9–33, DOI: http://dx.doi.org/10.1080/09503110.2013.767008

Fenwick, Corisande. (2020a).  Early Islamic North Africa: A new perspective. London: Bloomsbury.

Fenwick, Corisande. (2020b). The Umayyads and North Africa: Imperial rule and frontier society In:  The Umayyad World. Marsham, Andrew (ed.),   : 293–313. London: Routledge.

Fenwick, Corisande. (2022). How to found an Islamic state: The Idrīsid and the rivals to the Abbasid caliphate in the Far Islamic West In:  The Islamic Historian at Work: Essays in Honour of Hugh N. Kennedy. van Berkel, Maaike L, Osti, Letizia Letizia (eds.),   : 91–116. Leiden: Brill.

Grau-Sologestoa, Idoia. (2012).  Agriculture and ironwork in the Middle Ages: New evidence of bone anvils in Spain.  Munibe Antropologia-Arkeologia 63 : 305–19.

‘Idhārī, Ibn. (1948–51).  Al-Bayān al-muġrib fī aḫbār mulāk al-Andalus wa-l-Maġrib. Colin, G.S, Lévi-Provençal, É É (eds.),   Leiden: Brill.

Khurradādhbih, Ibn. (1889).  Kitāb al-masālik [wa-]al-mamālik. de Goeje, MJ (ed.),   Leiden: Brill.

Khan, Ahmad Nabi. (1990).  Al-Mansurah: A forgotten Arab metropolis in Pakistan. Vol. 2 Karachi: Department of Archaeology and Museums, Government of Pakistan.

King, Anthony. (2018). The faunal remains In:  Volubilis après Rome. Fentress, Elizabeth, Limane, H H (eds.),   : 369–86. Leiden: Brill.

Landon, Mark. (2016).  Making a Mint: Comparative studies in late Iron Age coin mould. Oxford: Archaeopress.

Messier, Ronald A; Miller, James A. (2015).  The Last Civilized Place: Sijilmasa and its Saharan destiny. Austin: University of Texas Press.

Nixon, Sam; Rehren, Thilo; Guerra, Maria Filomena. (2011).  New light on the early Islamic West African gold trade: Coin moulds from Tadmekka, Mali.  Antiquity 85 (330) : 1353–68, DOI: http://dx.doi.org/10.1017/.S0003598X00062104

Rebuffat, René. (1965).  Le développement urbain de Volubilis au second siècle de notre ère.  BCTH 1 : 235.

Thouvenot, Raymond. (1969).  Les origines chrétiennes en Maurétanie Tingitane.  Revue des études anciennes 71 (3) : 354–78.

Whitcomb, Donald. (1994). The Misr of Ayla: Settlement at al-‘Aqaba in the early Islamic period In:  The Byzantine and Early Islamic Near East. Volume II: Land Use and Settlement Patterns. King, GRD, Cameron, Averil Averil (eds.),   : 155–70. Princeton, NJ: Darwin Press.