Introduction
The Acholla Archaeological Project is a newly-established project with the aim of investigating the ancient site of Acholla (Henchir Botria, Tunisia) (Figure 1). The project is an international collaboration between the Institut National du Patrimoine (INP), Dickinson College and the University of Oklahoma, with additional support from the University of Leicester and the Endangered Archaeology in the Middle East and North Africa (EAMENA) project. It is co-directed by J. Andrew Dufton (Dickinson), Nesrine Nasr (INP), Nichole Sheldrick (Leicester) and Sue Alcock (Oklahoma). The first season of the project took place over two and a half weeks (9–25 June, 2025) and was undertaken by an international team of 17 archaeologists, specialists and students, from Tunisia, USA, Canada, UK and France.
Figure 1. Location of Acholla (Basemap Google Earth via xyz tiles, QGIS).
The main aims of the first season were to establish our methodology and to undertake fieldwalking survey and collection in at least two different areas of the site. In addition, topographic and architectural survey began the process of creating a new and updated map of the site, complementing the fieldwalking by providing further evidence and context to understand the ancient city’s extent, its chronology of occupation and possible zones of urban activity. The 2025 work has already yielded significant new information about this clearly substantial but much understudied city. The methodology established will provide a strong foundation for future fieldwork campaigns and has identified areas to target for further investigation.
Background
From its foundation by Phoenician settlers, the ancient city of Acholla has occupied an important position as a regional port and urban centre for the surrounding countryside. The early life of the city is relatively poorly understood and lacks the type of foundational narrative or historical dates of settlement we have for better-known sites such as Carthage or Utica. Some degree of presence at the site can be reasonably hypothesised from around the fifth century BC, if not earlier, but archaeological and textual evidence for this early period is sparse.
Acholla had clearly established itself as a noteworthy agricultural centre by the third century BC as part of a network of sites along the Mediterranean coast, and the site appears occasionally in Roman accounts of the conflicts with Carthage and subsequent late-Republican infighting. In the historian Livy’s discussion of the Second Punic War, for example, the Carthaginian general Hannibal Barca – defeated by Scipio Africanus at the battle of Zama – fled to his estates on the coast between Acholla and Thapsus (modern Bekalta) before setting sail for the eastern Mediterranean (History of Rome, 33.48). In the Third Punic War, the city allied with the Roman invaders and provided supplies to Roman troops during the final siege of Carthage (Appian, Roman History, 8.94). After the fall of Carthage, Acholla was rewarded the municipal status of oppidum liberum, allowing the city control over its own taxation, agricultural hinterland and even, at one point, the production of coinage (Pliny, Natural History, 5.30; Charles-Picard Reference Charles-Picard1947; Hobson Reference Hobson2015, 35–41). The city reappears – still as a free state – in the history of Caesar’s campaigns in North Africa during his war with Pompey. A convoy from Acholla was sent to Caesar offering supplies and, in return, the city was given a garrison of Caesar’s troops (Bellum Africum, 33). The presence of Caesar’s troops was not without risks, as the city was at one point besieged by enemy forces. The siege ultimately proved unsuccessful, a result of both the walls fortifying the city (not yet attested archaeologically) and the resistance of the Achollans who regularly set fire to the siege engines (Bellum Africum, 43).
The city played a major regional role during the height of Roman power, with evidence suggesting a great deal of wealth was concentrated within the Achollan elite. Well-attested elite families like that of M. Asinius Rufinus, whose house has seen considerable previous study, held substantial agricultural landholdings (Blanc-Bijon Reference Blanc-Bijon, Gozlan and Jeddi2001). Monumental, elite-sponsored developments of the second and third centuries AD include an amphitheatre, theatre and several bath complexes. While we lack firm evidence for the municipal status the city at this time, the overwhelming majority of North-African cities with multiple entertainment structures held the civic rank of municipium or colonia and we might safely assume Acholla achieved the same status at some point during this period (Dufton and Fentress Reference Dufton, Fentress and Hitchner2022).
Less is known from the latter periods of occupation. An established Christian presence is noted through previous excavations, including a large church with two baptisteries (Duval Reference Duval1980) and an active burial area in the northeast (Duval Reference Duval2003). A fortification constructed on the northeast edge of the city is identified as Byzantine in most documentation but lacks firm archaeological dating or textual corroboration. Many sites along this stretch of the coast show signs of contraction and eventual abandonment during the early medieval or medieval periods, a fate which probably also befell Acholla. The precise chronology of Islamic occupation remains uncertain, although we note the surface presence of medieval ceramic evidence concentrating in and around the fortification.
Previous archaeological works
Despite this rich history, the site has seen only a few periods of archaeological excavation focused on elite residences, key monuments and the urban core (Figure 2). The site is best known for the mosaics which were exposed during excavations in the 1940s and 1950s and are now prominently featured in the Acholla Room of the Musée National du Bardo in Tunis. Publications from the early excavations focused primarily on mosaics and epigraphy at the expense of ceramic material, other small finds, and archaeological stratigraphy (e.g. Charles-Picard Reference Charles-Picard1947, Reference Charles-Picard1948, Reference Charles-Picard1953, Reference Charles-Picard1959, Reference Charles-Picard1968).
Figure 2. Plan of Acholla by Gozlan and Jeddi (Reference Gozlan and Jeddi2001, Plate 1). N indicates necropolis.
Suzanne Gozlan and colleagues revisited the remains of the site in 1979–80 during a campaign to re-evaluate the House of the Triumph of Neptune (Gozlan and Bourgeois Reference Gozlan and Bourgeois1981) and undertook additional cleaning and sondages at the House of Asinius Rufinus in 1984 and 1987 (Gozlan Reference Gozlan, Johnson, Ling and Smith1994) (Figure 3). Publications associated with these later works – in particular a monograph dedicated to the mosaics of the urban centre (Gozlan and Jeddi Reference Gozlan and Jeddi2001) and a second focused on the House of the Triumph of Neptune (Gozlan et al. Reference Gozlan, Bourgeois, El Fourgi, Jannin and Prudhomme1992) – provide some of the best and only indications of the city’s long-term occupation and associated ceramic remains. A reassessment by Duval, primarily of the site’s Christian remains, also provides additional context on the discoveries of these earlier works (Duval Reference Duval1980, Reference Duval2003).
Figure 3. The excavated urban core, showing the probable forum in the foreground and the House of Asinius. Rufinus behind.
The site’s sporadic history of fieldwork and heavy focus on mosaic remains has resulted in a diminished stature for Acholla within the scholarly work of the region. However, the city’s size, visible monumental remains, location and history all suggest Acholla was a place of considerable ancient significance. The Acholla Archaeological Project thus offers an opportunity to apply a range of modern survey and fieldwork techniques to answer questions about the city’s long-term evolution.
Pedestrian Survey
One of the main aims of the Acholla Archaeological Project is to establish the extent of the ancient city and, where possible, identify different zones of activity and periods of occupation. Systematic fieldwalking and surface collection over the entire site had never been attempted. Our project decided to undertake this substantial task in order to gather a baseline material assemblage for the site and to determine whether meaningful patterns in the distribution and dates of different types of objects could be identified. Survey limits were agreed with the INP, establishing both a primary urban-oriented zone for survey, which is the main focus of the project, and an extended zone, in which further targeted survey might be undertaken in future (Figure 4). For the 2025 season two different areas of the site, on the east and west sides, were chosen to gain preliminary insights into how finds might compare in these two zones.
Figure 4. Survey limits and tract grid (Basemap: Bing Aerial via xyz tiles, QGIS).
Methodology
Acholla is an ideal site for non-invasive, intensive surface survey. The majority of the site remains open and undeveloped with few modern obstructions. Parts of the ancient site are currently in use for the cultivation of cereals and, in a few places, olive trees. However, planting and harvesting are done without the aid of modern machinery and disturbance is therefore minimal.
Fine-grained spatial data is needed to ask detailed questions of Acholla’s urban development and to identify zones of activity. It was therefore decided to undertake a survey at 10 m intervals, following the precedent set by the urban survey conducted at Leptiminus (modern Lamta), a port city of similar size ca. 70 km north–northwest of Acholla (Stone et al. Reference Stone, Mattingly, Dore, Stone, Mattingly and Lazreg2011, 51–53). At Leptiminus the survey units were determined by modern field boundaries, but the comparatively open landscape of Acholla allowed us to establish a regular grid system of tracts measuring 50 m east–west by 100 m north–south (0.5 ha) for the entire survey area prior to the fieldwork (Figure 4).
Using a similar method to that established by Broodbank et al. (Reference Broodbank, Lucarini, Bokbot, Benattia, Bigoulimen, Brucato, Farr, Garcia-Molsosa, Hachami, Laoutari, Lombardi, Marsilio, Martin, Martínez Sánchez, Mazzini, Morales, Pelegrin, Radi, Rega, Sulas and Wilkinson2024) in their survey at Oued Beht (Morocco), each tract was divided into five 10 m-wide transects running north–south and assigned letters A to E. Each transect was then further divided north–south into ten 10 m x 10 m units assigned numbers from 1 to 10. Each tract therefore comprised a grid of 50 units, identified as A01, A02, …, B01, B02, etc. (Figure 5). Thus, for example, the fourth unit of Transect B in Tract 45, would be recorded as T045-B04. Tracts were numbered in the order that they were surveyed.
Figure 5. Example layout of a tract, showing division of transects (A-E) and units (1-10).
The boundaries of each tract were established using QGIS and set out in the field in advance of collection using handheld Garmin GPS units to locate the corners, which were marked with survey flags. Flags were also placed at the start and end points of each transect and at every 10 m along the east and west edges of the tracts to provide visual guidelines for the survey team.
Each tract was surveyed and documented by a team of six: one survey lead and five walkers. Each walker walked down the centre line of their assigned transect and were instructed to pick up all artefacts within arm’s reach. After every 10 m, walkers changed collection bags so that the location and density of artefacts could be recorded for every 10 x 10 m unit. For the first two days of field-walking (Tracts T001 to T005), team members were instructed to collect any and all visible materials to familiarise themselves with the methodology and to create a strong base dataset for the finds analysis. Given the volume of material present on the site, after the first two days it was decided to exclude ceramic sherds less than 3 cm in diameter from collection, unless diagnostic in some way. The survey lead recorded the survey conditions, took notes and photographs and recorded the location of any notable features or finds within the tract using a handheld GPS unit.
Tracts T001 to T008 were walked from north to south ends, but all subsequent tracts were walked from south to north. This decision was made as it became clear that walking all transects in the same direction was preferable to maintain clarity during the recording process and consistency in the relative position of the 10-metre units. Some areas which were inaccessible due to rubbish piles, dense vegetation, or marshland along the coast were not surveyed (see further details below).
Results
A total of 51 tracts were surveyed across two zones of the ancient city (Figure 6) – covering 25.5 hectares or ca. 25 percent of the intensive survey area designated for the project. The first zone, on the eastern side of the site between the excavated urban core and the coast, extended to the north and south of a linear depression running roughly east–southeast, which has previously been identified as a possible decumanus. A second zone at the western limits of the defined survey area was selected to try and find the limits of the city in that direction, as well as to investigate a previously identified sector of pottery production.
Figure 6. Tracts surveyed during the 2025 season (Basemap: Bing Aerial via xyz tiles, QGIS).
Zone 1
Zone 1 comprised 23 tracts (T001–T023) which were surveyed by the fieldwalking team between 9–13 June. This area was targeted to assess the nature and density of ceramics and other finds in the area east of the excavated urban core. A dirt road running roughly NE–SW through the site served as a general western limit of Zone 1, and the tracts continue to the east to the coastline. Surface visibility was generally high across Zone 1, ranging from 75–95 percent. In the western and southern parts, vegetation coverage consisted of scrub grass, low shrubs and small bushes; cobble-sized stones were also scattered across much of the surface (Figure 7). To the north and east towards the sea, the conditions transitioned to low shrubs and exposed sand (Figure 8). Rows of small bushes sometimes delimited field boundaries, primarily in T001–T008, but these fields were neither tilled nor planted at the time of the survey.
Figure 7. Surface conditions in T013 showing cobbles and small shrubs (facing east).
Figure 8. Surface conditions in T020 showing shrubs and exposed sand near the coast (facing south).
Surface material was most densely concentrated in the southwest tracts, closest to the excavated urban core, and decreased in density closer to the coastline. While this could in theory reflect ancient occupation patterns, it is equally possible that modern conditions are affecting these results. The tracts in the south and west of Zone 1 (T001–T007 and T010–T015) were generally clear of modern obstructions, with the exception of the modern track marking the western edge of the zone and some large bushes. However, the areas further from the excavated urban core, particularly at the north end of Zone 1, have been more affected by modern disturbances. The north ends of tracts T016–T019 and large parts of T022–T023 could not be surveyed properly due to large mounds of trash and the compacted dirt tracks. These tracks also ran through large sections of tracts T008–T009 and T020. In addition, large parts of the easternmost tracts were marshy and/or had standing water and therefore could not be surveyed. It is also likely that the wetter and more exposed conditions closer to the coast have made materials more prone to weathering or being washed into the sea.
Zone 1 contained a number of previously known features and monuments, including the Baptistery (Architectural Feature AF015), Vaulted Building (AF016), Necropolis 2 and the possible decumanus, the latter visible as a linear depression cutting across (from east to west) T006, T005, T004, T013 and T014 (Figure 9). Other visible walls and structures were also noted in the zone during fieldwalking, especially in the tracts closest to the coast, some of which have now been documented as part of the Topographic and Architectural Documentation surveys (see below).
Figure 9. The possible decumanus visible in the foreground in T005 (facing northwest).
The combination of visible modern disturbances and the presence of substantial architectural remains near the coast suggest that the lower density of artefacts collected in the eastern tracts of Zone 1 may be due to the poor surface conditions rather than to a lack of ancient occupation. Further investigation and analysis of both the material collected so far and the structures recorded will be needed to better understand the nature of the remains in this part of the site.
Zone 2
Zone 2 consisted of 28 tracts (T024–T051) which were surveyed between 16–20 June. These tracts were located inland, to the west and south of the excavated area of the urban core. This zone was selected primarily to see if there was a substantial drop-off in finds that might indicate the western limits of the city, as well as to collect material from an area previously identified as a probable site of ceramic production (Peacock et al. Reference Peacock, Bejaoui and Ben Lazreg1990, 61).
Surface visibility over most of Zone 2 was high, ranging from 75–95 percent. Most tracts consisted of previously tilled and harvested fields delimited by thin (ca. 0.5 m wide), untilled borders of low grasses and occasional shrubs (Figure 10). The orientation of fields and direction of tilling rarely matched the north–south alignment of the transects. Some of the fields had recently been in crop and had been harvested by hand shortly before our work, leaving wheat chaff on the surface, but visibility was not drastically affected. Scrub grass, small bushes, scattered thistles and cobble-sized sandstones were found across the whole area; large bushes or trees were occasionally present but uncommon. Untilled fields were also noted within Zone 2 and these generally coincided with an increase in small cobbles (Figure 11). Areas of particular cobble density were noted in Tracts T032, T033 and T049. Larger stones (sandstone, limestone and sedimentary conglomerate) were observed less frequently.
Figure 10. Tilled and harvested fields in T042, with untilled borders (facing south).
Figure 11. An untilled field with dense scatter of cobbles, T049 (facing north).
The western edge of Zone 2 was delimited by a dirt road cutting northeast–southwest through Tracts T024, T039 and T048 and northwest–southeast through T029, T037 and T038. The south border of the zone was marked by the edge of a recently tilled olive grove which occupied the southern parts of T035, T034 and T049; the olive grove was separated from the tilled fields by another dirt road and an earthen bank in some places, and in others by a fence (Figure 12). The areas of the olive grove afforded much higher surface visibility and returned a higher number of artefacts than the untilled portion of each of these tracts.
Figure 12. Boundary between olive grove and fields in T049 (facing west).
There was a general decrease in the density of finds in the south and west part of Zone 2, although whether this genuinely represents the western edge of the ancient city is not yet clear. Earlier published plans of the city identified a cemetery (Necropolis 5, Gozlan and Jeddi Reference Gozlan and Jeddi2001, Pl. 1; see Figure 2, above) in the southwestern part of Zone 2, in the vicinity of T030–T032. This could potentially account for the lower density of artefacts, though it is worth noting that no particular evidence for the necropolis was identified during the survey. The presence of the compacted dirt roads, where few artefacts were collected, combined with a dense, impassable patch of thistles in T037 and T038 may also have contributed to this pattern.
No standing structures were recorded within Zone 2, though several mounds and concentrations of finds were identified which almost certainly represent buried buildings, for example, the large mound in Tracts T050 and T051 (Figure 13). There was a high density of ceramic building material in these two tracts and a large collection of marble was found on the surface of the mound in T050, probably gathered by an unknown person from the immediate vicinity. A fragment of a terracotta figurine was similarly found within the visible bounds of this feature (SF002, see Small Finds section below). Other mounds and fragmentary walls in T027, T043 and T034 almost certainly also represent buried structures; a cistern was also noted at the north end of T043.
Figure 13. Mound in T050, probable buried structure (facing west).
Of particular interest in Zone 2 was an area of probable ceramic production in the northwest corner of the surveyed area (T039–T040, T046, T048). At the north end of T039 and T040 a discrete area of ashy, grey sediment was evident, distinct from the surrounding yellowish-brown soil, and numerous kiln wasters were observed on the surface (Figure 14). Immediately to the northeast, another dense collection of ceramics and kiln wasters was also identified spreading across T048 and T046. This evidence suggests that this was the location of an active ceramic industry in the western outskirts of the city (see also Ceramics section, below).
Figure 14. Probable ceramic kiln in T040, marked by ashy grey discolouration (facing west).
Topographic survey
While detailed architectural plans of various individual buildings at the site have been produced in previous publications, the site plan published by Gozlan and Jeddi (Reference Gozlan and Jeddi2001, Figure 2, above), based on an aerial photograph, is the most recent existing map/representation of the site as a whole. A campaign of topographic survey was therefore begun in order to create an up-to-date, grounded plan of Acholla.
Using a Total Station, the team established seven topographic survey stations in strategic locations from which points were taken on monuments, structures and features across the site. For each structure surveyed, the team created a sketch plan to which survey points were added and exact co-ordinates were measured (Figure 15).
Figure 15. Sketch plan of the House of the Triumph of Neptune.
Surveys were undertaken of several previously known structures and monuments across the site, including the House of the Triumph of Neptune, the amphitheatre, the (presumed) church with two baptisteries and a vaulted structure near the coast. Other visible structures and walls, some of which had not previously been recorded, were also documented. For example, a substantial rectangular building with several rooms of different sizes (some with traces of surviving plaster) and a well with hydraulic mortar were recorded to the southwest of the House of the Triumph of Neptune (Figure 16).
Figure 16. Rectangular structure southwest of the House of the Triumph of Neptune.
Three of the seven previously identified necropoleis at Acholla were also visited by the survey team. The surface remains of Necropolis 1, located along the coast to the north of the Byzantine fort, consist of a few piles of stones. Necropolis 2 extends in a north–south strip next to the coastline and contains the foundations of several tombs, some of which are covered by small- and medium-sized stone rubble. Others are flush with the ground, including some jar tombs. The extent of this necropolis is not clearly defined, but tombs in different locations within it were surveyed.
The extensive Necropolis 3 is located at the southern edge of the site along the coast. Three areas were identified by the topographic survey team with tombs clearly visible at ground level, including jar tombs (Figure 17). The tombs face different directions: while most face east–west, several face north–south, potentially suggesting different phases of use. Necropolis 3 requires urgent documentation and investigation, as much of the surface layer has been lost due to erosion and human intervention. During the field season, the topography team discovered three tombs which had been disturbed, which was reported to local authorities (Figure 17). One of these tombs included an exposed cremation urn which was rapidly documented and collected for study with the other finds from the 2025 season (Figure 18).
Figure 17. Graves visible in Necropolis 3.
Figure 18. Illegally excavated grave in Necropolis 3.
The work of creating technical drawings based on the topographic survey undertaken during the 2025 season of fieldwork is underway. The survey itself will continue during the next season with the goal of producing as detailed and accurate plan of the site as possible, including visible monuments and their current state of conservation.
Architectural documentation
In addition to the topographic survey, during the 2025 season we began a systematic architectural recording of the structural features across the site, cataloguing details such as materials used and construction techniques. The architectural documentation in 2025 focused on Zone 1 where 16 architectural features (AF001–AF016) were identified and recorded (Table 1). Two of these features were previously known from earlier investigations, specifically the probable church with two baptisteries (AF015) and the vaulted structure near the coast (AF016). One of the primary aims of this documentation was to record the many other walls and structures noted during fieldwalking which have never been formally documented.
Table 1. Architectural features recorded in Zone 1
The location of each architectural feature was recorded with a handheld GPS unit and described, measured, sketched and photographed. No cleaning or excavation was undertaken for any of these features; the recording thus depended in large part on the extent of the feature’s surface-level exposure and its visibility.
Most of the architectural features recorded consisted of small sections of walls exposed at surface level, either in isolation or clustered together. The proximity of these features to each other sometimes suggested a larger structure below but it was usually impossible to define the relationship between sections. The architectural features in Tracts T022 and T021 are one such example (AF001–AF006).
Other architectural features identified in Zone 1 had visibly intersecting or related walls, though again it was not possible to fully determine the layout or extent of these structures. The smallest, least well-preserved and least defined of these structures (AF007) comprised two adjoining walls suggesting the corner of a small building. Similarly, two visible walls of AF011 suggest the corner of what may possibly be a cistern.
AF014, AF015 and AF016 were the largest features recorded. AF014 comprised two sides of a rectangular structure, potentially with a smaller rectangular element within, close to the sea. AF015, a probable church, was partially excavated in 1947 but not described in any detail in early reports or publications. Duval provides further description based on his own visits to the site in the 1960s, but the building has never been systematically documented (Duval Reference Duval1980, Reference Duval2003). Duval’s study focused primarily on two baptismal fonts that are still visible today, if largely filled in and covered with rubble. To the east of the two basins, an apse in medium-sized stonework is visible at ground level.
The recording process in 2025 identified characteristics that should help to phase AF015 in future. For example, identification of a wall placed over a mortared floor/preparatory floor layer indicates a renovation. An ashlar block used as a base for the intersection of multiple walls mirrors a construction technique also employed in the House of Asinius Rufinus. Measurements were taken of the apse facing the coast, the baptismal font, the cistern with an opening on the western side, the walls and the maximum length and width of the building. Future excavation or cleaning would clarify the relationship between the apse and the (previously excavated) central area, which would in turn help to establish the limits and layout of the building.
The vaulted structure AF016 was also previously identified by the team of Gozlan and Jeddi (Reference Gozlan and Jeddi2001, Plate 1, Figure 2, above) but, as with the baptistery, has never been formally recorded. Visible structures above ground include the remains of vaulting and walls made of coursed, worked stone and mortar (Figure 19) and multiple large basins with thick hydraulic mortar; satellite imagery clearly demonstrates that these collapsed vaults must originally have been part of a large complex (Figure 20). The recording of the vaulted complex AF016 was not finished due to time constraints and will be completed during the next field season.
Figure 19. Collapsed remains of the vaulted structure (AF016).
Figure 20. Satellite imagery showing large complex around the collapsed remains of the vaulted structure. AF016 and Structure AF014 to the north (Basemap: Bing Aerial via xyz tiles, QGIS).
Finds processing and analysis
A methodology of total collection was adopted during fieldwalking to obtain the widest range of material. As a result, nearly 40,000 finds were collected during the 2025 survey including fragments of ceramic vessels, marble, ceramic building material, fresco, stucco, coins and other small finds (Table 2).
Table 2. Finds collected during the 2025 field season
Finds were sorted, classified and counted based on material type. The counts for each object category were recorded on the reverse of survey sheets and then transferred to a spreadsheet to link the data to a GIS for analysis and visualisation. Some preliminary assessment of the ceramic material was undertaken during the 2025 season by two ceramics specialists, along with basic inventories of the other categories of finds. Preliminary information about the finds collected during this season are included below; detailed analyses will be undertaken by specialists in future seasons.
Ceramics
Over 37,000 individual fragments of pottery were collected from the two zones investigated during the 2025 field season (Figure 21). With only one week possible for intensive study, the analysis of ceramic material was designed to maximise the information available on ceramic type and chronology to quickly characterise the areas studied. A representative sample of the two survey zones was therefore selected to give an initial impression of the characteristics of each area: a single transect from each of T001, T025 and T040 (see Figure 6, above).
Figure 21. Distribution of pottery collected during the 2025 season (Basemap: Bing Aerial via xyz tiles, QGIS).
The ceramics were first divided into classes (e.g. fine wares, cooking wares, amphorae, etc.), then into types; where possible, fabrics and possible production sites were determined with the help of an eyeglass. Following the sorting, the pottery from each context was quantified and noteworthy sherds were set aside, photographed and given separate inventory numbers to be retained for future study and drawing. As much information as possible was recorded from non-diagnostic sherds such that we might identify patterns in the distribution of at least basic classes of vessel, if not detailed type. All pottery from the sampled transects was recorded in an Excel spreadsheet. Once recorded, these sherds were discarded and returned to the site. A study campaign planned for Spring 2026 will enable the inventory of ceramics from the 2025 survey to be finalised.
The material collected during the 2025 season was in a fragmentary state, as expected from a surface context, and the density and composition of the assemblages differed from one tract to another. Collected materials consisted of fine tableware, cooking wares, common ceramics, moulded ceramics and transport amphorae. Lamps are present but only sporadically. The ceramic assemblage examined included imports from Italy and, more rarely, from the eastern Mediterranean basin.
Fine wares
Fine wares were represented by both black-glazed pottery from the Punic period and African Red Slip (ARS). Finds from this group remain less abundant in comparison to other types. Black-glazed pottery (V PUN) is attested only twice (T1.C1.1) in Tract T001. These are most likely to be local productions dating from the Neo-Punic period. ARS is slightly more abundant, recorded in T001 and T025. In terms of forms, we note the presence of Group A2 (Hayes 27 dish) dating from the late second to early third century AD; group A/D represented by a Hayes 31 dish dating from the second to the fourth century AD; and group D2, produced in northern Tunisia, represented by a rim from a Hayes 61 dish (T1.C1.2). Finally, ARS E, produced in southern Byzacena, is represented by a rim from a Hayes 68 dish (T1.C1.3).
Cooking wares
Cooking wares were well represented in all three transects analysed. The repertoire consists of Type C pottery from northern Tunisia, including lids Hayes 195 and Hayes 196, combined with Type B pottery from Byzacena, such as shallow casserole Hayes 181 and slipped lids Hayes 182 and Hayes 185. CB-type cooking wares were frequently represented by the Hayes 183 pot. Sabratha 104 lids are also abundant, reminiscent of Tripolitanian production. Overall, cooking wares cover a wide chronological range from the first to the fifth centuries AD. The chronology of regional production is less clear and merits further detailed study.
Common wares
Common wares were more abundant than cooking wares. The typological repertoire varies between open and closed forms. Open forms consist of traditional Punic dishes such as the Sabratha 243 bowl, dating to the first century BC, and the Sabratha 213–214 bowl, dating from the first century AD. The second century is represented by an abundance of Bonifay 41 lids. Open forms are also represented by the Uzita 1, Uzita 2 and Uzita 3 basins of Punic tradition, datable to between the second and fifth centuries AD, depending on the variant. Some bowls are reminiscent of the typological variants attested at Leptiminus, such as the C SB 26 form. The closed forms are mainly jugs, similar to regional productions such as the Thapsus 1 and Thapsus 2 forms which can be dated between the first and second centuries AD.
Amphorae
Transport vessels are ubiquitous among the artefacts collected. A rim from a Greco-Italic amphora dating from the third century BC provides the earliest evidence of occupation in the areas surveyed. The transition period is represented by two Neo-Punic Sabratha 9 amphorae rims (T025.A09.34), dating from between the first century BC and the first century AD. Italic imports are attested at least once by a Dressel 2/4 wine amphora dating from between the first and third centuries, as well as a possible Lamboglia 2 amphora. African amphorae from the early empire are represented by Africana I, Africana II and, more rarely, Africana III types (and their variants), (T1.C7.21) and (T1.C7.22). The origins of the African amphorae are limited to the workshops of Salakta and, most likely, a regional workshop. Later production of these containers is rarer. We note only the presence of a Keay 26 rim dating from between the fifth and sixth centuries AD (T025.A07.30) and a body fragment of an LR1-type amphora imported from the coast of Turkey dating from between the fifth and seventh centuries. It is also worth noting the presence of an amphora fragment bearing a stamp from T003 (Figure 22), which will provide important evidence about the production of these vessels. However, this tract has not yet been analysed; more detailed information will be provided in future reports.
Figure 22. Amphora fragment with stamp (T003-D03).
Handmade wares
This group of ceramics is represented exclusively by imports from the island of Pantelleria, such as the Baldassari 3 type (T1.C2.13). This production generally dates from between the second and fifth centuries. Alongside this, we note the presence of African handmade ceramics, abundantly represented by the Kanoun rims of the Bonifay 66 type.
Preliminary analysis of the assemblage
One of the main objectives of the ceramics analysis was to provide quick preliminary information on the characteristics of the different areas studied and identify activities such as occupation, production, agriculture, etc., as well as to assess potential changes in density and activity over time. Some basic analyses of the assemblage are presented below, but should be treated as subject to change with additional study.
In Zone 1, one transect (C) of T001 was chosen for the preliminary analysis because of the high density of finds collected in the tract and its proximity to the excavated urban core. A total of 508 sherds were collected from Transect C, around 30 percent of which were diagnostic in some way. The best-represented period was from 150–300 AD, with 59 percent of the datable ceramics in circulation during this period, followed by the period from 0–150 AD, represented by 42 percent of the forms.Footnote 1 The periods immediately preceding and following the broad period from 0–300 AD were also represented, though in far smaller proportions: the period from 150–0 BC with 6 percent and the period from 300–500 AD with 8 percent. In terms of forms, the largest category was amphorae, representing 48 percent of the total of classifiable sherds. Common wares were the second most commonly identified at around 28 percent of the total, followed by cooking wares (16 percent) and fine wares (3 percent). The high proportion of amphorae could suggest that this area was more engaged in trade and/or storage activities than residential behaviour.
In Zone 2, two contiguous transects (A) were selected, from Tracts T025 and T040, both to assess the western edge of the surveyed zone and to undertake a preliminary analysis of the finds from the probable kiln located at the north end of T040. Far fewer sherds (95) were collected in T025 than in T001, suggesting that this area may have been less densely occupied, which makes sense given its location further from the urban core. Only 36 sherds were dateable; as in T001, the most represented period was from 150–300 AD (44 percent), followed closely by the period from 0–150 AD (38 percent). Again, the periods immediately before and after were represented in smaller proportions, 14 percent and 5 percent respectively. Also similar to T001, amphorae were most commonly recorded, making up 68 percent of the assemblage; cooking wares made up 14 percent, common wares 13 percent, and fine wares 2 percent. The high proportion of amphorae in the assemblage suggests that this may have been an area engaged in trade and/or storage activities and is also very likely to be related to the probable ceramic production site identified in the next tract to the north.
In T040, 480 sherds were collected from Transect A, though nearly 75 percent of these were from the two northernmost units, A09 and A10. The area around these units had previously been identified as the location of a probable kiln, identifiable by a spread of ashy grey soil and a very dense collection of wasters (see Figure 14 above). However, despite the high number of sherds collected, only 50 were identified as diagnostic in some way. Of these, the period from 150–300 AD once again was the most represented, with 66 percent of the dateable sherds. Around 19 percent of the forms dated from 0–150 AD, with 12 percent from the period from 150–0 BC and 12 percent from the period from 300–500 AD. The vast majority of the assemblage from T040 was made up of amphorae (53 percent) (Figure 23) and common wares (44 percent), with only small amounts of cooking ware (2 percent) and fine wares (0.4 percent). It seems likely therefore that the identified production site was in the business of manufacturing the former two classes of ceramic and was probably most active between 150–300 AD.
Figure 23. Kiln wasters of Africana I (left) and Africana II (right) amphorae from T040.
Marble, sculpture and inscriptions
Over 1,000 pieces of marble were collected during the fieldwalking of which several different types have already been identified. Marble fragments were recovered in both Zones 1 and 2, but with distinct distributions (Figure 24). In Zone 1, the majority of marble was found in the southern tracts; this may reflect the distribution of public and elite buildings, but it is also worth noting that there was more modern disturbance in the northern tracts of Zone 1, which affected collection. In Zone 2, marble was concentrated in the eastern tracts with the distribution dropping off towards the west. This potentially represents the approximate edge of the urban core.
Figure 24. Distribution of marble collected during the 2025 season (Basemap: Bing Aerial via xyz tiles, QGIS).
Most of the fragments recovered were plain slabs of marble, in most cases probably from floors or walls, but there were also many fragments of architectural sculpture. Seven inscription fragments have also been identified (none of which have more than a few fragmentary letters) (Figure 25), along with at least one possible piece of (non-architectural) sculpture. It is possible that with closer inspection further sculpture and/or inscription fragments will be identified. Cataloguing is in progress and studies will be undertaken by specialists in future seasons.
Figure 25. Inscription fragment IN002 (T011-A04).
Tesserae and mosaic fragments
Small fragments of mosaic and individual tesserae were collected for analysis: 208 individual loose tesserae and 21 fragments of mosaic were collected during the 2025 season. A few small clusters can be identified in the distribution, which probably relate to buried structures. Most of the tesserae recovered were stone, though some glass and terracotta examples were also identified.
Painted plaster (fresco and stucco)
For the most part, only fragments of plaster with evidence of painting were collected, though some representative samples of undecorated plaster were also kept for technical analysis and comparison. In total, 151 fragments of plaster were collected. Fewer fragments were recovered from the eastern part of Zone 1, perhaps owing to the damper conditions closer to the coast which are not conducive to the survival of plaster. There were a large number of fresco fragments that retained their painted decoration, representing a wide variety of colours and patterns, as well as stucco fragments (Figure 26). Cataloguing and analysis will be carried out by a specialist in a future season.
Figure 26. Fragment of fresco (T041-D06), left, and moulded stucco (T027-B10), right.
Other building materials
Survey teams were instructed to collect a representative sample of ceramic building materials such as roof tiles and to note their presence in the field if not collected. A large number of fragments (815) were nonetheless collected during the field survey. This discrepancy between methodology and actual collected remains was due in part to the inexperience of some of the survey team, who were not always able to distinguish between ceramic building material and ceramic vessels (particularly when the sherds were very fragmentary). The distribution of the materials recovered was similar to that of marble, however, with very few fragments recovered in the north part of Zone 1 and the frequency dropping off in the west of Zone 2. Some small fragments of mortars and cements (e.g. opus caementicium) were collected as samples, but the survey team was instructed in most cases only to note their presence in the field.
Coins
A total of 17 coins were recovered during the survey, all of which came from Zone 1. Most of the coins were small (less than 1.5 cm diameter) and not legible in their current state. A preliminary inventory of the coins was undertaken and the coins were assigned inventory numbers (e.g. C001). A specialist will be engaged in future seasons to clean and assess the coins more fully.
Small finds
Ten small finds were recovered during the survey, including three fishing net weights, two loom weights, three basalt grinder fragments, a white marble jeton and a terracotta figure head (Figure 27). The latter (SF002, found in T051-C04) has been identified as Venus with a crown. A basic inventory of the small finds was completed this season and small find numbers assigned (e.g. SF001). Relevant specialists will be engaged in future seasons to assess and catalogue these materials.
Figure 27. Small Find SF002 (T051-C04), terracotta figurine head representing Venus with crown.
Other finds
Other finds collected included glass, metal, slag and bone. Some of these finds are certainly ancient, though others could be more recent in date. These finds will require evaluation in future seasons to identify and assess their potential significance.
Conclusions and next steps
The 2025 season of the Acholla Archaeological Project was a successful first season of work. The project team established a strong survey methodology which will form the basis for future investigations and the data thus far recovered will inform the direction of activity and research at the site. Over 25 ha of the site were surveyed and nearly 40,000 artefact fragments collected. Once these have been fully inventoried and studied, this substantial collection of finds will provide new and vital information about the dynamic, diachronic life of the ancient city, and particularly trade and production not only at Acholla itself but in the surrounding region. Topographic and architectural surveys have already made considerable progress toward an updated plan of the site and its monuments, which will in turn allow us to better understand its layout and long-term development.
In addition to the primary fieldwork, a number of steps were undertaken at the site focusing on building the infrastructure necessary for future fieldwork and artefact storage. Improvements were made to a guardian’s hut constructed by an excavation team in the centre of the site many years previously but which had fallen into disrepair after a long period of disuse. Three new wooden inner doors and iron outer doors were installed to make the hut suitable for storage of the artefacts collected during fieldwork and to ensure their security.
Funds were also provided to repair and waterproof the roof and to add iron bars to the windows to further improve the security of the building. The project also sponsored construction of an awning over the concrete pavement area in front of the hut to create a suitably shaded workspace for finds processing or other on-site work. In addition to these physical improvements, the project provided support for the engagement of a night guardian for eight months. The regular inspection of the site during the night will deter looting, theft and vandalism at Acholla.
Looking forward, a season dedicated to the study of the ceramics is planned for Spring 2026 to complete the inventory and analysis of the materials collected during the 2025 season. A campaign of geophysical and/or drone survey is also envisaged before the next major field season. During the 2026 season, the team will carry on with fieldwalking over the remainder of the site. The topographic and architectural documentation will continue, and additional specialists will be engaged to study the finds from the previous season.
In 2026 we hope also to begin work on the documentation of the ancient necropoleis of Acholla. We intend to begin with the documentation of Necropolis 3, where, as mentioned above, several tombs have unfortunately been recently disturbed by illegal excavations, making documentation of this, and the other known necropoleis of the site, an urgent priority.
Acknowledgements
We wish to thank the Institut National du Patrimoine of Tunisia for the opportunity to study the site of Acholla and their support of this project. We are grateful to all our participants this season for their hard work and dedication, including Nabil Bel Mabrouk, Wided Ben Abdallah, Fadhila Ben Messaoud, Abbir Daada, Claire Davis, Fatma Haddad, Rached Hamdi, Olfa Hsini, Raja Krida, Victoria Leitch, Katelin McCullough, Siobhan Morgan, Abdallah Neili and Anthony Rogers.The 2025 season was funded primarily by Dickinson College, including generous support from the Neil B. Weissman Fund for Faculty Research. Dickinson student participation was supported by the Christopher Roberts Travel Fund. Additional support was also provided by the Endangered Archaeology in the Middle East and North Africa (EAMENA) project, which is funded by Arcadia (2312-5142).
Open access
