Gallia

This study examines the emergence and decline of sanctuaries from the Alps to the Mediterranean between the 5th century BC and the 4th century AD. It draws on a corpus of 114 sanctuaries used from the reign of Augustus to the 4th century AD, and 45 places of worship or collective ritual practices from the Late Iron Age. A second approach focuses on approximately 30 Roman sanctuaries established on Late Iron Age oppida (fortified hilltop settlements) in southern Gaul.Given the extensive literature on the presumed continuity of sacred sites, this article examines changes in the religious landscape, mainly between the Iron Age and the Early Roman Empire, to understand the reasons behind continuity, abandonment and reappropriation.What memory of sacred sites survived between the Iron Age and the Early Roman Empire?A few examples suggest long-term continuity into the Early Roman Empire, sometimes from an early Iron Age date (Glanum, Cavaillon and Correns). However, they are scarce and sometimes questionable (Vernègues, Tourdan). The two best-documented cases of continuity between the Iron Age and the Roman period are in the Pennine Alps (Martigny, En Zibre and the Great-Saint-Bernard Pass), a region conquered much later and more closely connected to Gallia Comata.The scarcity of continuity among sacred sites may reflect the specific characteristics of southern Gaul. Can it be explained by the absence in this region of the “grands sanctuaires installés au cœur de ‘territoires ethniques’ dans le courant du second âge du Fer, devenant ensuite – pour la plupart – de grands centres religieux et civiques à l’époque romaine” (Roure et al. 2019, p. 31)? The early development of urban centres in southern Gaul may indeed account for this absence (Roure et al. 2019). The continuity of many major urban centres as civitas capitals (Glanum, Cavaillon, Avignon, etc.) may also explain the preservation of some earlier places of worship, driven by local elites.One phenomenon might seemingly challenge this observation: the establishment of Roman sanctuaries on former fortified hilltop settlements (oppida). Around 30 cases have been identified in Narbonensis and neighbouring regions, a phenomenon also attested elsewhere in Gaul (Woolf 1996; Barral, Nouvel 2019; Bossard 2024). These sites initially appeared particularly promising for demonstrating continuity. Yet the available evidence does not establish it (Golosetti 2009a; 2014b; 2014c; 2016; 2017).Although excavations are often old and limited, and Roman construction has disturbed archaeological layers, no excavation of a Roman sanctuary on an oppidum has provided conclusive evidence of an Iron Age origin (La Fageole in Recoules-Prévinquières, Truc in Saint-Bonnet-de-Chirac, L’Impernal in Luzech, La Granède in Millau, Cluzel in Toulouse). Even where sedimentation from the Gallic period is preserved, no structures and no artefacts, including residual material (e.g. Châtelard du Lardiers, Grand Rochefort in Varces-Allières-et-Risset, and Combalou in Roquefort-sur-Soulzon), are characteristic of votive deposition, although the archaeological features of Gallic ritual deposits are now relatively well known (Gruat and Izac-Imbert 2007; Golosetti 2009b).Some Roman sanctuaries are known only from out-of-context inscriptions (Castellar de Cadenet, Hières-sur-Amby, Soyons, Serviers-et-Labaume), surface finds or insufficient test pits (Saint-Michel-de-Valbonne, Vidauban, Moustiers-Sainte-Marie, Roussas). It is therefore impossible to determine whether or not the Roman sanctuary replaced a Gallic place of worship.Lack of evidence does not necessarily mean lack of continuity. Nevertheless, the available data suggest the reappropriation of oppida rather than continuous cultic activity. The monumental or dominant position of former fortifications may have given these places particular significance in the landscape. Memories of the ruins of former settlements, their fortifications or their topographical position may have influenced sanctuary location without requiring an undocumented earlier cultic phase. This leads us to consider the concept of an “archaeology of memory”.This is all the more significant as Iron Age sanctuaries securely attested on certain oppida—La Cloche, Constantine, Entremont and Teste-Nègre—disappeared at the end of the Iron Age or the beginning of the Roman period, along with the settlements that housed them. Gallic fortified settlements and their places of worship were thus abandoned simultaneously, while some oppida were later reappropriated for Roman sanctuaries with no proven earlier origins.Towards a new religious landscapeThe low frequency of continuity points to the cessation of activity at many Gallic places of worship around the turn of the era and the emergence of a new religious landscape in the Roman period. When did this occur? At the time of the conquest? During the establishment of the Provincia of Transalpine Gaul? Or with the Augustan provincial reorganisation? Before answering, we should consider communal buildings that some scholars regard as places of worship but that were initially excluded from this study due to the absence of associated ritual artefacts (Golosetti 2016, p. 41).Of the twenty-two Late Iron Age communal buildings examined here, all located within fortified settlements (oppida) in south-eastern Gaul, none underwent Roman-period reoccupation that can clearly be interpreted as cultic. At Glanum (XXXIIa and XXXVIII), Arles (cryptoportiques) and Mouriès (Jean-Jean), these buildings may have remained occupied in the Roman period, but no evidence indicates cultic activity. Nor is any other communal building known to have been reused during the Roman period. Whether these structures were sanctuaries belonging to a different category from those identified through metal votive deposits, or, as proposed here, communal spaces that were not places of worship, none shows continuity with Roman sanctuaries, despite their relatively large number. In either case, they fit the broader pattern of sanctuary abandonment at the end of the Iron Age.Indeed, several Gallic ritual deposition sites (Peiro de l’Autar in Cuers, La Cloche in Les Pennes-Mirabeau and Entremont in Aix-en-Provence) show that activity ceased at the end of the Iron Age. However, this abandonment of places of worship identified by depositional practices occurred at the same time as the abandonment of the oppida that either housed them or lay in their immediate vicinity.The study of monumental sanctuaries confirms this process of abandonment. These include the in situ Italic temple at Glanum and the “temenos” at Constantine (Lançon-de-Provence), as well as displaced blocks from possible Italic temples at Saint-Blaise and Constantine. All were destroyed between the late 2nd and mid-1st century BC. Here again, the abandonment of monumental sanctuaries often coincided with that of their settlements, except at Glanum.Forgetting is thus an integral part of the process of constructing memory (Lavabre 2013), as also shown by Susan Alcock (2001; 2002), Simon Price (2012) and Greg Woolf (1996). This paper also seeks to demonstrate that some Gallic communities did not simply forget their past but constructed, reconstructed and even reinvented it. Landscape analysis can shed light on the loss of memory (the abandonment of La Tène places of worship), its preservation (continuity of sacred sites) and its reinterpretation (Roman sanctuaries on oppida?).The tension between memory and oblivion is particularly evident at Glanum. This civitas capital developed a new urban religious landscape during the late Octavian and Augustan periods. It combined the preservation of a “historical” cultic nucleus around the spring with the site’s reappropriation through the temple of Valetudo, donated by Agrippa. Further away, a new Roman-style monumental centre was established after the destruction of the earlier complex and its Italic temple. A new place of worship was then created there, probably linked to the imperial dynastic cults, towards the end of the 1st century BC.The emergence of a new religious landscape at Augustan Glanum reflects a broader process. Our study indicates that a new religious landscape took shape towards the end of the 1st century BC. This differs from the pattern apparently developing at the same time in central-eastern and western Gaul, where major Late Iron Age sanctuaries were more widely maintained, though neither systematically nor uniformly (e.g. Corent, Gournay-sur-Aronde, Mandeure and Tintignac) (Barral et al., eds. 2003; Bouvet et al. 2003; Nouvel 2007; Van Andringa 2017a; Barral, Nouvel 2019; Bossard 2024).This transformation in southern Gaul was not an abrupt or uniform break. Some Gallic sanctuaries remained in use before disappearing during the 1st century BC. Meanwhile, a few early phases of new places of worship dating to the same century remain poorly identified. These sanctuaries differed in status (public or private) and location, from rural settings (Présilly, Blauvac) to towns (Faverges) and civitas capitals (Glanum, with the temple of Valetudo, and possibly at Mas de Guigue in Tarascon).Most new places of worship were therefore established during the Augustan period. This applies both to public sanctuaries, linked to Augustus’ reorganisation of the civitates of Narbonensis in 27 BC, and to private sanctuaries, which demonstrate the phenomenon’s predominantly local impetus.Considering only the best-dated sites, the evidence suggests, unsurprisingly, that the new Augustan religious landscape first developed in civitas capitals, notably through the early construction of monuments dedicated to the imperial dynastic cult at the centre of urban fora. Other public places of worship, particularly within towns in civitas territories, also seem to have appeared early (Vernègues, Fox-Amphoux). The creation of public and private places of worship, both across civitas territories and in their capitals, could nevertheless extend throughout the Augustan period (Riez, Arles, Châteauneuf, Lioux, Lamanon, La Bastide-des-Jourdans and Saint-Vincent in Saint-Paul-Trois-Châteaux).ConclusionThe conclusion drawn from this analysis seems as clear-cut as its interpretation is delicate. Alongside the very few cases of long-term continuity from the 6th or 5th century BC into the Roman period, there is an even more striking scarcity of continuity between Late Iron Age sanctuaries and the Roman period, contrary to observations in Gallia Comata. Moreover, although Roman sanctuaries established on Late Iron Age oppida might have concealed such continuity, the available evidence does not support this hypothesis. Instead, emphasis should be placed on the cessation of activity at a series of Gallic places of worship around the turn of the 1st century BC, including those located within oppida. During the Republican period in Transalpine Gaul, Gallic and, likewise, Massaliote places of worship gradually declined.New places of worship were not established on a large scale until the Augustan period, when they gave rise to a new religious landscape. This landscape remained dynamic: new sanctuaries appeared during the 1st and 2nd centuries AD, while others declined early (Gap, Vaugrenier, Châteauneuf) (Golosetti 2016). Yet these changes between the 1st and 4th centuries AD were minor compared with the major transformations between the Iron Age and the Roman period.These transformations appear to be linked to spatial dynamics of settlement patterns. The abandonment of several Iron Age settlements (oppida) coincided with that of their places of worship, whether monumental or not. Conversely, the creation of small towns at the beginning of the Roman period was accompanied by new sanctuaries, apparently responding to the needs of newly established communities (Fox-Amphoux, Cabasse, Châteauneuf, Faverges, Annecy, etc.). Where the continuity of sacred sites between the Iron Age and the Roman period may be explained by the persistence of urban settlements (Glanum, Correns, Cavaillon), the limited apparent survival of places of worship between the Gallic and Roman periods in southern Gaul might also be explained by the absence of “grand sanctuaire civique” within civitas territories, probably owing to their smaller size compared with those of Gallia Comata. Yet several major civic sanctuaries in Gallia Comata had precisely a La Tène origin.Similarities should also be noted. Inland Gaul underwent major changes in settlement patterns and its religious landscape. In eastern Gaul, one study shows that half of all sanctuaries had no protohistoric antecedents. These included sanctuaries within or on the outskirts of civitas capitals and other settlements established ex nihilo, as well as spring sanctuaries and those associated with rural establishments (Barral, Nouvel 2019). Southern Gaul shows a similar emergence of spring sanctuaries in the Roman period (Golosetti 2015), as well as places of worship linked to rural establishments.Thus, rather than adopting an overly simplistic notion of continuity, our understanding of the evolution of Gaul’s religious landscape must incorporate greater nuance. The persistence, forgetting, reappropriation and creation of places of worship form a “bricolage de faits religieux […] mêlant continuité, ruptures et innovations” (Isnart 2006, pp. 211–212), as anthropological research demonstrates (Rivière 2003).

This study presents the identification of two anthropomorphic lantern uprights recovered from the shipwrecks Plage de la Corniche 2 and Rieu 1, located off the Hérault coast, as part of a doctoral research project on the collections of the Musée de l’Éphèbe et d’archéologie sous-marine (Cap d’Agde). These objects were not recognized as lantern components at the time of their discovery, highlighting the difficulties of identifying metal artefacts in archaeological collections and the need to reassess material from earlier excavations. Both shipwrecks date to the Roman Imperial period and represent distinct maritime contexts: Plage de la Corniche 2 carried a cargo of marble blocks and sculptural elements, while Rieu 1 transported olive oil amphorae, copper ingots, coins, and items likely belonging to the ship’s onboard equipment. The lantern uprights add to a very limited corpus of lantern remains preserved in underwater contexts, with only a few comparable discoveries documented in the Mediterranean. While terrestrial finds emphasize the symbolic and religious significance of lanterns, particularly in funerary contexts, their function aboard ships remains debated. Iconographic and textual sources suggest that lanterns were used primarily as position or signalling lights rather than for general illumination, especially during convoyed or controlled nocturnal navigation. These findings contribute to a better understanding of Roman ship equipment and maritime signalling practices, and underline that night navigation was governed more by custom and risk assessment than by formal regulation.

The Gallic settlement of Saint-Blaise (Bouches-du-Rhône, France) is a major archaeological site in Provence, located between the River Rhône to the west and the Greek foundation of Marseille to the east. The first excavations undertaken by H. Rolland in 1935 uncovered a “Greek-type” ashlar fortification more than 400 m long, indicative of the economic and political importance of the town in the 2nd century BC. The archaeological site has been managed by Aix-Marseille Métropole since 2007, which has commissioned several archaeological excavations on the Gallic habitat as a whole, including one in 2011 on the southern sector of the plateau. This operation permitted a detailed survey of the already visible remains as well as two cross-sectional stratigraphic cuts of the tower, but in the absence of any real excavation and new ceramic finds, it casts no further light on the dating of the building already established in 1981 (Bouloumié, Fincker, Tréziny 1984) and 1984 (Arcelin, Cayot 1984 and Cayot 1984). This study begins by presenting the stratigraphy of the cladding applied in the 2nd century BC against the archaic flanking structure. Following on from the technical analysis of the structure already carried out by J.-C. Bessac (1980) and the metrological study undertaken by H. Tréziny (1989, pp. 26-27), this article goes on to detail the various architectural components used in the large-scale construction, from the foundation to the crown, specifying certain technical aspects of the construction site and the metrology employed in this section of the enclosure. The technical and architectural findings contribute to the debate on the supposed Mediterranean influences on this remarkable structure, which is as impressive for its defensive function as for its ostentatious appearance.The Archaic enclosureThe Archaic fortification served as a support and internal structure for the large-scale cladding added in the Hellenistic period. H. Rolland’s 1936 survey of side E revealed two distinct structures. A primitive core (MR 20.025), completely devoid of stelae, might correspond to a defensive system dating from the Late Bronze Age or late 7th-early 6th century BC. It seems likely that the lining of this first enclosure with a wall made almost exclusively of stelae constituted a refurbishment, like the two successive states also noted in the archaic fortification of the lower town, attributed to phases VIIa, dated to around 600 BC, and VIb, dated in the second quarter of the 6th century BC (Arcelin, Pradelle, Rigoir 1983, p. 142). In the current state of excavation, a monumental subcircular tower stands out on the surface, measuring some 14.50 m wide and protruding at least 9 m from the curtain wall. In addition to the military defence of a strategic sector of the promontory and a possible access route to the site, the building’s size serves a direct deterrent function, ostentatiously emphasising the community’s economic and human power. Its elevated position, 60 m above sea level, on the tip of the promontory of the upper town, accentuates this monumental effect, making it clearly visible from the coast. Last but not least, the systematic use of stelae to form the tower’s facing contributes to its monumental appearance. In addition to the aesthetics of a building carefully crafted out of modular blocks, the structure probably possessed a ritual value, intended to reinforce the defence of the site and undoubtedly associated with the perpetuation of a place of remembrance. It was not until the 2nd century BC that the Gauls decided to dress this “wall of stelae” with a large-scale construction.Renovation of the enclosure in the Hellenistic periodBuilding materials were extracted from nearby quarries, in the soft limestone of the Burdigalien period characterised by the presence of seashell fossils. Extraction was carried out in the “Gallo-Greek tradition”, using rounded wedges made of hardwood (olive or juniper) (Bessac 2002, p. 38, fig. 25; Chausserie-Laprée, Duval 2009, pp. 232-233). This practice would indicate that a Gallic workforce—free and/or bonded—was recruited to contribute directly to the extraction, transport and construction of the fortification. The rutted tracks indicate the supply routes taken from the quarries to the places where the blocks were deposited near the curtain walls of the Hellenistic fortification.The southern promontory’s defensive plan consists of a pentagonal tower with a solid base at least 3 m high, connected to curtain wall B, which protects the western postern. At the base of each corner of the pentagonal tower a heel protrudes from the bare facing, helping to verify the orientation of the facings: each obtuse angle described by the ground plan of the building would be more easily corroborated by the acute angle of the protruding heel. The angle determined by faces B and C measures 120°, while the two angles between faces C/D and E/F are 140° for two sections of masonry of very similar lengths. These two equal angles stand opposite the apex of the triangle formed at the beak between faces D/E, whose angle is 110/115°.Our study details the various architectural components, from the foundations to the crowning of rounded merlons: the rare blocks placed off-centre on face B; the bending of the chords and the offsets that facilitate the corner ties; the oblique sawing, some of which is deliberately very inclined; the clamp holes used to tighten the blocks, indicative of the direction of progress of the site; the herringbone cut which, in its variations, reveals the probable intervention of two to three workers specialised in this task; the drainage of rainwater... This focus also enables us to locate and describe the lapidary marks found on the southern promontory: two Bs, six examples of recumbent digamma, as well as the letters X (Late Antique mark?), T and the sign Γ. Finally, the interpretation of the crenellation with rounded merlons is still a matter of debate. It has been postulated that the “highly decorative” nature of the rounded merlons at Saint-Blaise (Tréziny 1993, p. 39) was chosen to recall the shape of Gallic votive stelae (Tréziny 1989, p. 30; 1992, p. 345). This borrowing would be a fine illustration of the rendition and reappropriation of a form transposed to a different host cultural system. However, this possible memorial projection only adds to the real defensive character of this form, which would be indicative of a borrowing from a Mediterranean model. But known references from the Punic world (Sicily, Tharros) or Eastern Greece (Lycia), already cited by H. Rolland and H. Tréziny, are all too old to claim direct influence. Marseilles may have borrowed this type of parapet from Phocaea, but the absence of such merlons from those two cities does not lend credence to this hypothesis. Although other examples of Eastern borrowings, such as the pentagonal towers characteristic of the Caria and Lycia regions, are widespread in the Mediterranean, they are not associated with the adoption of rounded crenellations in the Greek context. The other models of this type of crenellation that could be cited are late and correspond more to contexts in the direct orbit of the Carthaginian sphere. There has been little growth since the 1980s in the documentation concerning the southern part of Gaul, and what there has been mainly concerns the contemporary enclosure of Glanum. However, new data are eagerly awaited from current excavations on the fortifications of Les Baux-de-Provence (Corniot, Chapelin, 2025).MetrologyA few ideas are put forward regarding the metrological system used. Despite the variations inherent in large-scale cutting on the scale of a monumental defensive building, the distribution and frequency of block dimensions are consistent with three possible metrological units. The dominant foot would correspond to a standard of 31/33 cm (Doric foot). The cubit would be calibrated between 51 and 53 cm (a variant of the “Ionian” cubit, “Massaliète” and the basic unit of the Punic world). Another unit of measurement seems possible and would appear to correspond to the fraction of a 29/29.8 cm foot (attic-cycladic, Phocaean, Punic and Roman foot), including a number of measurements between 37 and 45 cm (1 + 1/2 to 1/4 of a foot) and a 40 cm modulus emphasising the heels and merlons. Based on these metric data, a master plan is proposed for the symmetrical construction of the pentagonal tower. A dozen erratic boulders cut to an atypical modulus (26 to 33 cm thick and 62 to 90 cm wide) may have served as paving slabs for the parapet walk or as stair treads.Large-scale construction for an indigenous fortificationModels of similar pentagonal towers are known to have existed in the 3rd century BC in Asia Minor and in the Iberian world in an indigenous context (Castellet de Banyoles). The construction of the Saint-Blaise wall may have taken anywhere from a few months to two years, depending on the resources available to the settlement in a troubled historical context marked by Ligurian revolts. However, the work was left unfinished, interrupted by the Roman military siege. While the large-scale fortification of Saint-Blaise undoubtedly had an ostentatious prestige value, it also reflected the military need to adapt to improvements in siege techniques. The cosmopolitan port of Marseille, like the Gallic settlement at the mouth of the Rhône, may have directly received travellers from various backgrounds, including Greeks and/or Punics, architects, scholars, and military personnel, who were able to disseminate this technological knowledge throughout the Mediterranean.

The Col de la Forcle (2,543 m) is in the western Swiss Alps, 26 kilometres east of the border between the French Department of Haute-Savoie and the Swiss Canton of Valais. On its northern side, the Forcle Glacier has been melting faster as a result of global warming. Between 2013 and 2023, surveys on the pass and glacier yielded around 350 fragments of wood and some 10 pieces of metal.Several stakes have been identified among the artifacts collected. The study of more than 195 fragments revealed 49 items with points carved in the shape of a spear, cross-sections of 2 to 3 cm and measuring around 150 cm in length. They are mainly made of softwood (Picea abies, Larix decidua, Abies alba) and, more rarely, hardwood (Corylus avellana).Their tips are carefully carved and sometimes decorated with notches. The spearheads are rounded or oval in cross-section, rarely quadrangular. The butt-spikes are pointed. The fragments are concentrated in two places, one near the pass, the other on a natural flat area about a hundred metres below. Generally, the “spears” appear to have been naturally broken and then scattered down the slope by gravity and the movements of glacier meltwater. 14-C analysis and dendrochronological measurements determined that the posts form a coherent set made from wood harvested in 158 AD or a few years later.These spear-shaped stakes form a one-off that is difficult to interpret. Their morphology would suggest that they could be weapons, but this purpose is belied by the use of softwood. Moreover, ancient texts mention the existence of wooden practice swords, but never spears. These stakes may be evidence of a cult site similar to Conjux (Savoie), dated by dendrochronology between 131 and 169 AD, or Oberordla in Thuringia (Germany), which was occupied throughout the Roman period. Both cult sites yielded numerous stakes around which offerings (ceramics, bones, etc.) had been placed. At the Col de la Forcle, a coin and a Roman shoe nail were the only objects discovered near the stakes.In 2021, a wooden spear dated by 14-C analysis to the Roman period was found on the Colerin pass (3,475 m) in the Savoy Alps. It could be linked with a wooden statue discovered eighteen years earlier at the same spot and dated to the 1st-2nd century AD (Flavian or early Antonine period). The situation could be similar in the Col de la Forcle, where the artefacts found could have had a symbolic function and might have been deposited deliberately as an act of veneration to a divinity. Another possible interpretation is that the Col de la Forcle, where the wood was found, was not its intended destination. Finally, although no convincing parallel can be proposed, they could simply have been utilitarian objects.Analysis of archaeological data from the Roman period shows that the areas around the pass were densely populated at low altitudes (settlements, villae rusticae, sanctuaries and necropolises). The permanent settlements extended at least to an altitude of 1200 m, where several necropolises have been located. A few isolated finds attest to the presence of humans in the high mountains (cattle grazing, movement of people and goods between valleys, etc.).Other wood collected from the pass has been dated using 14-C. A large squared fragment, whose function remains undefined, has been dated to the Iron Age (7th-5th century BC). Two artefacts, interpreted as walking sticks or marker posts (beacons), have been dated to the 13th-14th centuries AD. Finally, a set of timbers measuring around 70 cm in length and 2 to 3 cm in diameter have been dated to the 15th-17th centuries. The latter are spruce branches (Picea abies), which have been debarked with an axe, while some have also been split in half. They could have been bundles of firewood or, in the case of the split pieces, bindings for wooden containers.Although the archaeological finds from the Col de la Forcle have a long history, from the Early Iron Age to the modern period, the wooden artifacts discussed here seem to relate to one-off events (intentional deposition, accidental loss, etc.) separated by significant chronological gaps. Further 14-C analyses on some fragments not yet analysed here could reveal evidence of more regular passage use.The discoveries made at the Col de la Forcle were made possible by the accelerated retreat of the glacier as a result of global warming. Following the finding of the Hauslabjoch mummy in the eastern Alps in 1991, glacial archaeology research has grown at varying pace in the Alpine regions concerned. However, it should be stressed that most of the discoveries in glacial environments are made by non-archaeologists. There is therefore an urgent need to raise awareness of the importance of this fragile archaeological heritage.

North of Nîmes (Gard), the scrubland called the garrigue is home to a rich archaeological legacy that has received little archaeological intention. A quarry operation for the extraction of rock, which after 15 years would become the city of Nîmes’ largest run-off water retention basin, covered an 8-hectare area in a sector covered with agro-pastoral structures. An archaeological assessment conducted in 2015 revealed ancient quarries beneath the dry-stone huts (capitelles) and pasture walls. These quarries covered a significant portion of the marly limestone plateau included in the project area.Following this assessment, the Regional Archaeology Service prescribed excavations. The Antiquailles excavation began in September 2016 and continued until the end of the year, covering 26,000 m2. After studying the modern vernacular structures—probably built primarily from ancient extraction waste—the excavation focused on clearing and conducting a comprehensive study of one of the quarry zones. Additional trenches and windows were dug to obtain a better picture of the scale of the operation. The quarries appear to have been concentrated within the prescribed area, with some extraction fronts observed to the south and southwest during the assessment, which were not revisited. The trenches also revealed the presence of agriculture between the four identified quarries, dating to the early Imperial period.All the trenches showed that the exclusive target of extraction was a specific type of module: slabs of very large dimensions but often minimal thickness. The extraction techniques were identical across all observed layers and zones. A detailed study of one quarry gave precise insights into the extraction methods employed, and how rock layers, waste, and pathways were handled.This quarry, which had a total surface area of approximately 1,400 m2 and a maximum depth of about 3 metres, was divided into three main sectors separated by residues which demarcate them. Extraction would begin in the centre of one sector, and once it was completed the space would be occupied by waste, contained by small walls. Development then proceeded on multiple fronts, with debris accumulating while passageways were maintained between storage areas. Roadways were constructed between the walls, remnants or fronts, on quarry floors, and sometimes on embankments. These roadways were made by the juxtaposition of polygonal and fragmented slabs that had been rejected. They facilitated the movement of the extracted materials to the exterior. The absence of traces of wear on their surface suggests that they were used briefly, just long enough to open a few new fronts that were quickly abandoned.The extraction traces were systematically recorded in order to define the procedures employed by the quarrymen, the type of production, and the volume of extracted materials. Two main types of traces were identified. In the first type—rarely observed and only on the quarry floor—a groove was cut in order to delineate the slab to be extracted. These grooves were made with a launched percussion tool, such as a flat-edged chisel of Hellenistic tradition, or a pointed tool like a pick. The second type of trace consisted of roughly egg-shaped hollows at the opening, slightly flared and regularly aligned, forming “dotted lines” found on the quarry faces and floors, where they always delineated large rectangular surfaces. These hollows were made with a mortise chisel and sometimes a drill bit, accommodating small wedges, two examples of which were found, probably combined with shims that would have been used to detach the slabs from the rock along their perimeter. This technique offered several advantages: it was faster than the traditional cutting method, reduced waste, and allowed for specialisation in work methods.The thickness of the layers varied little, averaging about ten centimetres. The imprints of extracted slabs or remnants in the quarry were studied to obtain information about the types of products extracted from these thin layers. While widths varied, lengths were the most interesting criterion for understanding their use. Two groups stood out, with average lengths of 3.01 metres and 4.30 metres. The particularly large dimensions of these slabs relative to their thickness indicate that they would have required delicate handling both in the quarry and during transport to the city, which probably explains the large quantities of waste. The total extracted volume was approximately 2,330 m3, with an estimated usable volume of 25-30%, or about 580 m3, representing roughly 1,120 slabs.The rare archaeological artifacts found in the excavation waste date to the very beginning of Augustus’s reign. These excavations indicate that the quarrying must have been a response to a single, standardised order, intended for a specific use. Indeed, these thin slabs, in which length was the only criterion, were used exclusively to pave the bottoms of some large, vaulted collectors identified beneath the old city. All were lined with sections of paving identical in thickness to the Antiquailles slabs. They mark the beginning of the development of the city of Nîmes within the context of the Augustan city-building programme.