Introduction

It is a sad reality that the later prehistoric funerary and ritual landscapes of North Africa west of Egypt remain, despite extensive investigation over the past 200 years, the least widely known and understood in the Mediterranean region. This is particularly problematic considering the diverse conditions in its modern countries, characterized by rapid urban development, political and economic instability, and widespread looting. Urban expansion and agricultural practices have contributed to the destruction of many known and unknown archaeological sites, hampering efforts to understand the broader social landscapes of this region, while the high visibility of cemeteries and monuments, and their association with well-preserved artifacts, has further fuelled an illegal market for collectors. For example, in Morocco, treasure hunters, known locally as “Souassa,” have been active since at least the late medieval period, and their often systematic looting remains a significant issue today (Camps, 1961, pp. 11–12).

Alongside these current challenges, recent historical factors also contribute to the gaps in our understanding of later prehistoric landscapes in Mediterranean North Africa. In particular, during the colonial era, post-Pleistocene archaeological research primarily focused on the Roman period, largely neglecting the pre-Roman and later prehistoric heritage (Díaz-Andreu, 2015; Sanmartí et al., 2012, pp. 23–24). The one salient exception, of course, is the investigation of the numerous dolmen-like burials (for a synthesis and further references, see Camps, 1961, pp. 115–152). Consequently, the periodization of post-Pleistocene phases is poorly delimited and is based on old cultural labels and borrowed terminologies applied during the colonial period, whose local relevance has yet to be demonstrated (e.g., “Copper Age,” “Bronze Age,” “Protohistory,” and “Iron Age”). Regrettably, until new data, particularly radiocarbon dates, becomes available, avoiding the use of this terminology may prove challenging.

Subsequent research on Holocene, pre-Roman Mediterranean Africa has focused more on the so-called neolithization process and the later “Phoenician Iron Age,” with far less attention paid to the intervening later prehistoric millennia (for insights, see Broodbank & Lucarini, 2019). This selective focus, as opposed to a more comprehensive, diachronic approach, has resulted in an uneven distribution of radiocarbon dates, creating an uncomplete mosaic of knowledge that hinders researchers from analyzing late prehistoric social, economic, and cultural dynamics over the longue durée (Lucarini et al., 2021). Finally, the paucity of data regarding settlements, particularly those dated between the fourth and early first millennia BC, is equally regrettable given the relevance of settlement sites for understanding broader social landscapes.

This paper focuses on the ritual landscape of the Tangier Peninsula from c. 3000 to 500 BC, an expanse of land located at the northwestern tip of the African continent, in modern Morocco (Fig. 1). Given the growing need to bridge gaps in non-Nilotic North African archaeology, the Tangier Peninsula emerges as an ideal case study for several reasons. First, its unique geographic position as a connecting gateway between Europe and Africa, as well as the inner Atlantic and the Mediterranean, suggests that this area may be crucial for understanding northwest Africa’s ritual and funerary landscapes. Second, the Tangier Peninsula is comparatively extensively researched, with major projects conducted over the past two decades, yet its funerary and ritual landscapes remain under-analyzed. Third, of all non-Nilotic Mediterranean African regions, the Tangier Peninsula displays a uniquely wide variety of burial traditions, rites, symbolic sites, rock art, and megalithic monuments with close parallels in both late prehistoric Atlantic Europe and the Saharan and pre-Saharan regions. While previous studies have often treated cemeteries and rock art as separate elements, this paper adopts an integrative approach, considering rock art, cemeteries, standing stones, and ritualized artefacts as interconnected components of a broader symbolic and ritual landscape. Accordingly, our fieldwork in the area (conducted within the framework of the Tahadart Archaeological Project and the Kach Kouch Archaeological ProjectFootnote 1) has sought to understand, through survey and excavation, the region’s late prehistoric social, cultural, and economic dynamics and their links to other Mediterranean and Atlantic regions.

Fig. 1
Fig. 1
Full size image

(a) Map showing the regions mentioned in the text; (b) map showing all the known settlements in the Tangier Peninsula and potential terrestrial connections. Basemap: GMTED2010 and Landsat 8. Figure by Hamza Benattia

Tangier Peninsula’s Geographic Setting: Terrestrial and Maritime Connections over Time

The Tangier Peninsula is roughly 8000 km2 in area and constitutes the African side of the Strait of Gibraltar, separated by c. 14 km from the European continent at its shortest point. Bordered by the Atlantic and the Mediterranean on three sides, the River Loukkous to the south and the River Laou to the southeast, it lies at a strategic maritime and terrestrial crossroads (Fig. 1). Ecologically and geologically, the Tangier Peninsula is rich and diverse, split between a sandstone-dominated Atlantic side, with vast plains and large rivers, and a limestone Mediterranean side, characterized by steep valleys and mountains. Consequently, the Mediterranean coast, in the northern and eastern strip, is rugged but today offers better anchorage due to the presence of bays and coves as well as less tidal rise and fall. The Atlantic western coast, though now sandy and battered by extreme tides and currents, in late prehistoric times featured open bays, large estuaries, and lagoons (Carmona-González et al., 2010, pp. 53–60; El Gharbaoui, 1981; Khalfaoui et al., 2023) that can be tentatively reconstructed with GIS software (Fig. 1b).

Notably, the maritime connectivity of the Tangier Peninsula may have been fostered by the inner Atlantic setting of the Ibero-Moroccan Gulf (Vanney & Ménanteau, 2005), an extensive maritime region c. 700 km wide, if we assume Cabo da Roca in Portugal and Cape Beddouza in Morocco to be its ends (Fig. 1a). Additionally, the geographic shape of the Tangier Peninsula likely made land connections to other parts of northwest Africa easier by following the natural contours of its orography. However, coastal routes along the Atlantic strip during prehistoric times may have been avoided due to lagoons and estuaries, or alternatively, in part navigated by boat. The use of GIS software provides a better understanding of terrestrial mobility within the Peninsula, allowing the identification of communication corridors and crossroads (Fig. 1b). This can be modeled by linking known third to earliest first millennia BC settlements, employing least cost paths tool (Fig. 1b; see Supplementary Material 1 for more information on the methodology employed). As part of our field project, two of the thereby identified crossroads have been targeted by extensive survey. Crossroad A (Fig. 1b) yielded significant new evidence of late prehistoric ritual activity including cemeteries, standing stones, and rock art, as reported in this paper.

Additionally, crossroad D (Fig. 1b) has yielded new information regarding settled life (Benattia, 2025). Particularly notable is the unparalleled farming settlement at Kach Kouch, occupied from the late third to earliest first millennia BC and featuring round mud-built houses, rock-cut silos, evidence of domesticated plants and animals as well as connections with the Mediterranean and the Atlantic (Benattia et al., 2025).

Additionally, the Peninsula’s location opposite southern Iberia and separated by a few kilometers may have fostered maritime connections since the Late Stone Age. This is supported by evidence of shared material culture found on both sides of the Strait (Raissouni et al., 2016). The arrival by c. 5400 BC of the Neolithic package from other areas of the Western Mediterranean, first to the Moroccan northern coast, and subsequently extending to inland areas, hints at increasing connectivity by sea (Guilaine, 2018; Martínez-Sánchez et al., 2021). The subsequent Middle Neolithic (c. 4500 BC) is, by contrast, characterized by a certain Saharan influence, probably as a result of human mobility caused by rapid environmental changes in the Sahara (Mercuri et al., 2011). This period is characterized along northwest Morocco’s Atlantic strip by the so-called Achakar Ware, with distinctive flat-bottomed, finely decorated vessels (Martínez-Sánchez et al., 2018). While the Final Neolithic in the Tangier Peninsula remains poorly understood, new discoveries further south at Oued Beht reveal extensive later fourth to early third millennia BC connections with Iberia, evidenced by a shared distinctive painted pottery (Broodbank et al., 2024) and mirrored across the strait by ivory objects, ostrich eggshells, and potentially, donkey remains (Cardoso et al., 2013; Schuhmacher, 2017). While the Copper Age transition occurred in southern and eastern Iberia around 3200–3000 BC, this phase remains less defined in northwest Africa, but nonetheless, the presence of Beaker pottery and metallic objects in northern Morocco and northwestern Algeria provides substantial hints of a regional equivalent and continued connections across the Strait during the third millennium BC (Bokbot, 2005). The subsequent Bronze Age (c. 2200–800 BC in Iberia) remains one of the least known and radiocarbon-dated periods of northwest Africa’s later prehistory (Broodbank & Lucarini, 2019). Yet this situation is gradually improving due to new radiocarbon dates, notably from Kach Kouch (Benattia et al., 2025) as well as cave burials (Ramos et al., 2011; Vijande-Vila et al., 2019) and cist cemeteries (as reported in this paper, Fig. 3). Understanding the social, economic, and cultural dynamics of the second millennium BC, in particular, is naturally vital for evaluating ensuing regional dynamics during the subsequent period, known locally as Mauretanian 1, which spans from c. 800 to 500 BC (Akerraz et al., 2020b, p. 77, Table 4.1). This final phase is better known thanks to renewed excavations and other research at major Phoenician settlements along Atlantic and Mediterranean Africa (Aranegui-Gascó, 2005; Callegarin et al., 2016; Docter et al., 2004; López-Castro et al., 2016; López-Pardo & Mederos-Martín, 2008; Vuillemot, 1965).

Finally, it should be noted that the extensive connections evidenced above may have been more intense, variable, and complex than initially assumed. This connectivity positions the Tangier Peninsula and the Strait of Gibraltar at the center of an extensive network. Although later Classical sources place the Pillars of Hercules, as a sort of finis terrae, in this area, the reality is that outside the Mediterranean sphere, the inner Atlantic also bustled with its own dynamics of mobility and connections. As this paper aims to demonstrate, such all-round connectivity is reified, from c. 3000 to 500 BC, in the occurrence of a rich record of prehistoric monuments, rock art, and rituals that span imposed continental divides.

Cemeteries and Burial Traditions in the Tangier Peninsula and Beyond

The strategic positioning of the Tangier Peninsula, which fostered its maritime and terrestrial connections, played a crucial role in shaping the region’s deathscapes. This section explores the available funerary evidence in the Tangier Peninsula and broader northwest African region, from the Early Holocene to the end of the Mauretanian 1 period (c. 8000–500 BC).

Background: Early and Middle Holocene Cemeteries and Burial Traditions (c. 8000–3000 BC)

The earliest burial recorded in the Tangier Peninsula is at Hattab II Cave, in the Western Rif mountains near the Mediterranean coast (Fig, 2b: 27). This inhumation of a male (c. 25–30 years old), lacking upper incisors, included grave goods such as a burnt lithic artefact dated by TL at 8900 ± 1100 BP, confirming a late Iberomaurusian chronology and suggesting the continuity of burial practices into the following period (Barton et al., 2008). Additionally, a poorly understood burial context at Ghar Cahal (Fig. 2b: 24), in the northern Tangier Peninsula, may be tentatively linked to the Iberomaurusian Period (Fusté-Ara, 1961), although the lack of systematic excavation and radiocarbon dates hinders a better understanding of its chronology. During the subsequent Early Neolithic period (c. 5400–4500 BC), burials in wider northwest Morocco occurred in both caves (El Hajraoui et al. 2012; Fregel et al., 2018; Simões et al., 2023; Turek & Vintr, 2016) and open-air graves, the latter notably attested within the Tangier region and linked to a large settlement at the extinct bay of Tahadart (Otte & El Abdellaoui, 2004, pp. 163–168; Fig. 2b: 47; Fig. 3). Funerary evidence for the ensuing Middle and Final Neolithic (c. 4500–3000 BC) is absent from the Tangier Peninsula, but comparanda from broader northwest Morocco suggest a predilection for caves and open-air cemeteries. The former is exemplified by Kef El Baroud (Fig. 2a: 30), where an individual was radiocarbon dated to 3702–3536 BC cal. at 2 sigmas (Fregel et al., 2018; Fig. 3), while the long-known site at Rouazi-Skhirat located in the fossilized dunes at the mouth of the Oued (river) Cherrat is the earliest example of a large open-air cemetery, encompassing c. 100 burials (Texier et al., 2008; Fig. 2a: 43). Recent radiocarbon dates on human bone suggest a Middle Neolithic chronology (Fig. 3), centered around 4500 BC (Simões et al., 2023; Turek & Vintr, 2016), further supported by the presence of Achakar Ware vessels among grave goods, also attested in the mid- 5 th millennium BC layers at the Achakar Caves and Zafrín Island (Martínez-Sánchez et al., 2018; Rojo et al., 2010). Finally, the funerary record of the Final Neolithic later fourth to earliest third millennia BC is poorly understood throughout northwest Morocco except for the intriguing cemetery at El Kiffen rock shelter (Fig. 2a: 19), where a radiocarbon date on human bone yielded a date cal. at 2 sigmas 3320–2630 BC (Daugas et al., 1989, p. 682; Fig. 3). However, a recent publication suggests a significantly older chronology, spanning from the early sixth to the late fifth millennium BC (Lacombe, 2023: 484). These new radiocarbon dates indicate that El Kiffen cemetery may have been in use from the Late Iberomaurusian to the Middle Neolithic. Renewed excavations and radiocarbon dating are necessary to shed light on this poorly understood site.

Fig. 2
Fig. 2
Full size image

(a) Map showing the main cemeteries mentioned in the text; (b) map showing all the known cemeteries in the Tangier Peninsula. In alphabetical order: 1. Achakar plateau (pit and hypogaea burials) 2. Ain Berriche 3. Ain Dalia Kebira 4. Ain Misteheyia 5. Ali Thalat 6. Atlas tumuli 7. Benzú 8. Beni Maadane 9. Beni Snassen 10. Bray’s Cave 11. Carthage 12. Carmona 13. Cerro Peluca 14. Daroua Zaydan 15. Djebila 16. Dougga 17. Eastern and central Maghreb dolmens 18. El Argar cist cemeteries 19. El Kiffen 20. El Khil C (Achakar Caves) 21. El Mers 22. El Mriès 23. Ferradeira culture cist cemeteries 24. Gar Cahal 25. Gharb tumuli 26. Hafa II 27. Hattab II cave 28. Ifri N’Amr Ou Moussa 29. Kannoua 30. Kef el Baroud 31. Laqueleya 32. La Traviesa 33. Loma del Puerco 34. Menkal I 35. Middle Wadi Draa tumuli 36. Moulouya tumuli 37. Mzoura 38. Orden seminario 39. Oued Ksiar 40. Oued Liane VII 41. Oulad Zin 42. Rabat Caves 43. Rouazi-Skhirat 44. Roummane 45. SE-B, SE-K and Jardín de Alá (Pits and cist burials) 46. Sidi Slimane 47. Tahadart 48. Toscanos. Basemap: GMTED2010 and Landsat 8. Figure by Hamza Benattia

Fig. 3
Fig. 3
Full size image

Radiocarbon sequence of cemeteries and burials in northwestern Africa, performed exclusively on human bone samples. Caibrated with OxCal v4.4.4, Bronk-Ramsey, 2021. Figure by Hamza Benattia

Late Holocene Monuments, Cemeteries, and Burial Traditions (c. 3000–500 BC)

Building on this overview of antecedent patterning in the funerary record, the following section discusses available evidence from the Tangier Peninsula and wider northwest Africa from c. 3000 to 500 BC. Although precise dates will be provided when available, the general lack of radiocarbon dating and data requires the occasional use of terms such as “Copper Age” (mid-to-later third millennium BC), “Bronze Age” (terminal third to early first millennia BC), and “Iron Age” (early-to-later first millennium BC).

Cave Burials

The funerary evidence for the core centuries of the third millennium BC, contemporary with the Iberian Copper Age and marked in parts of northwest Africa by the presence of Beaker sites, currently remains largely blank in the Tangier Peninsula, except for an intriguing, yet unconclusive radiocarbon date on a human tooth cal. at 2 sigmas of 2470–2240 BC (Martínez-Sánchez et al. 2018, p. 432, Table 5) from the El Khil C (Achakar Caves; Fig. 2b: 20), near modern Tangier (Fig. 3). Though it is unclear whether this derived from a lost burial, previously attested burials in a flexed fetal position from the Achakar Caves may suggest that this find was part of a larger cemetery (Bleicher, 1875, p. 210). Further east, within the Peninsula, cave burials are attested at Benzú, calibrated at 2 sigmas to 2127–1902 BC, and, a few centuries later, at Hafa II, likewise calibrated to 1610–1450 BC (Ramos et al., 2011; Vijande-Vila et al., 2019) (Fig. 3). Benzú and Hafa II are located around the Djebel Moussa (Fig. 2b: 7,26), a symbolically and geographically prominent locus associated with the Pillars of Hercules legend and suggestively mirroring second millennium BC burials across the strait at Bray’s Cave, in Gibraltar (Giles-Guzmán et al., 2017; Fig. 2a: 10)—though whether as part of a now symbolic maritime space or simply a continuation of burial practices in the Tangier Peninsula since the Iberomaurusian is unclear.

Pit Burials

A little further south of our area, the only potential Beaker burial (c. 2500–2200 BC) is that of Sidi Slimane in the Gharb Region (Fig. 2a: 46), as suggested by the stray find of several complete vessels presumably from a funerary context, the detailed information concerning which is largely lost (Souville, 1984). Pit burials in the Tangier Peninsula during this period consist of bodies placed in extended (and from the size of some pits presumably also flexed) position within simple pits often covered by one or several slabs (Buchet, 1907, p. 399). Apparently, some of these burials may have been part of larger cist cemeteries (Jodin, 1964a, p. 40), mirroring similar practices across the Strait of Gibraltar, where mixed pits and cist burials cemeteries have been attested (Hunt-Ortiz et al., 2007). Our study of the archive of Cesar Montalban, who was appointed in the 1940 s as Head of Archaeology at the Tangier International Administration, found evidence of previously undocumented pit burials on the Achakar plateau, above the famous caves (Supplementary Material 2: 1, Fig. 13). These burials contained the remains of an undetermined number of individuals buried in pits with their heads pointing north and feet south along with several pottery vessels. These pottery vessels were later rediscovered at the Archaeological Museum of Tangier by Jodin (1964b). While Jodin recognized these finds as excavated by Montalban, he could not determine their precise location and archaeological context, mistakenly associating them with cist burials (Jodin, 1964b). However, the presence of a singularly boat-shaped vessel, noted by Montalban (Supplementary Material 2: 1, Fig. 13b), has allowed us to make the correct association (Jodin, 1964b, p. 327, plate 1b). Although the exact chronology of these late prehistoric burials remains unclear, the shape and fabrics of the vases strongly suggest a third to second millennia BC date. Pit burials are a well-known type in wider North Africa. As mentioned earlier, such practices have been recorded at Middle Neolithic Rouazi-Skhirat (Fig. 2a: 43), although the Achakar plateau examples represent the first evidence for the little-known third to second millennia BC Morocco. Further East in Tunisia, pit burials inside Capsian shell-middens at Ain Misteheyia and Ain Berriche (Fig. 2a: 2,4), previously thought to date from the Medieval period, are now shown through radiocarbon dates to fall within the third and second millennia BC (Jackes et al., 2015; Lubell et al., 2009). Similarly, several pit burials beneath the Numidian settlement at Dougga (Fig. 2a: 16) yielded two radiocarbon dates centered in the first half of the second millennium BC (Khanoussi et al., 2004). Finally, in neighboring Iberia, pit burials are frequent during the third and second millennia BC (Fig. 2a), with large cemeteries documented both in coastal and inland areas (Deamos et al., 2015; González-Toraya et al., 2010; Hunt-Ortiz et al., 2007; Linares-Catela, 2020).

Cist Burials

From at least the late third millennium BC (Daroua Zaydan: see Fig. 3 and below), cist burials constitute a new funerary type that emerged in the Tangier Peninsula. For the first time, the monumentalized nature of these constructions, compared to previous burial traditions, required a significant investment of time and resources, perhaps suggesting a more stable and consolidated territorial control. In the wider North African region, especially along the Mediterranean strip of modern Algeria and Tunisia, megalithic burials were labeled as dolmens by early French researchers, associating them with similar European monuments (Camps, 1961, p. 115–152). Camps argued that the Tangier Peninsula’s cist burials differed from the rest of northern Africa’s dolmen tradition given the former’s distinctive cist shape, older chronology, and Ibero-Atlantic filiation, from where they supposedly originated, together with Beaker pottery and metallic objects (Camps, 1961, p. 152). In contrast, the central and eastern Maghreb’s dolmens (Fig. 2a: 17) emerged as a result of contacts and exchanges with Italy and Sardinia in the Late Bronze Age, immediately before the Phoenician Iron Age, disappearing prior to the Roman conquest (Camps, 1961, p. 152–153). Intriguingly, an apparently isolated dolmen, containing polished axes but lacking metallic objects, was excavated in 1873 at Beni Snassen (Fig. 2a: 9) in northeastern Morocco (Velain, 1885) and may argue for an older, generically Neolithic, origin for dolmen monuments in North Africa. Unfortunately, though, this is now impossible to verify, and the general lack of documentation, radiocarbon dates, and more modern research currently prevents a deeper understanding of the origins and development of these burials across North Africa. The most that can be said is that their restriction to north of the Atlas ranges does indeed suggest a Mediterranean origin and offer some continued currency for Camps argument (1961, p. 125) that they represent an initially non-local burial tradition in contrast to the deep north African and Saharan ancestry of tumuli.

Returning to the Tangier Peninsula, cist burials were firstly documented in the region by French explorers (Tissot, 1876), with subsequent research during the colonial and postcolonial periods expanding knowledge through survey and excavation (Buchet, 1907; Jodin, 1964a; Salmon, 1904). The main research on cist cemeteries is attributable to Ponsich (1967, 1970), who synthetized the available information and excavated new sites around the Tahadart (Fig. 2b), proposing differences between Bronze Age and Iron Age cists, based on distinctive burial goods and construction techniques (Ponsich, 1970, p. 37–168). Bronze Age cists were typically constructed placing four large vertical slabs arranged in a trapezoidal disposition, covered by a larger horizontal slab and often displaying a crescent-shaped stone accumulation outside the eastern slab (Fig. 4b). In some cases, vertical slabs were replaced on one or more sides by a rough wall built with medium-sized unworked stones without bonding (Jodin, 1964a). These Bronze Age cists closely resemble southwestern Iberian examples, though the crescent-shaped stone accumulation is unique to the northwest African structures. In both the Tangier Peninsula and Iberia, cist burials are frequently found either isolated or grouped into cemeteries concentrated on a single hill or dispersed across several adjacent hills (Del Amo-y de la Hera 1993; Ponsich, 1970). Typically, they contain one individual but can host up to two or even three, as seen at the Ali Thalat cist (Fig. 2b: 5), in the middle Oued Laou river (Quintero-Atauri, 1941). Bodies were interred in a flexed fetal position, often sprinkled with ochre and accompanied by handmade pottery, tortoiseshell, and metallic items (Ponsich, 1970). The latter includes awls, Palmela points, and a halberd/dagger (Fig. 4a), with closest parallels in Early Bronze Age Iberia and potentially elsewhere in Atlantic Europe (Delibes-de Castro, 1977; Lull et al., 2018; Ponsich, 1970). Occasionally, tombs contain exclusively grave goods, and no detectable human remains, while completely empty tombs are also frequent (Ponsich, 1970). The absence of human remains, also observed in Iberian cists, has fueled a long debate as to whether this is due to post-depositional factors or if tombs were intentionally built as cenotaphs (for insights, see Manuel-Valdés et al., 1998). In the Tangier Peninsula, the presence of empty tombs alongside others containing human remains suggests cultural practices rather than post-depositional factors, assuming these tombs have not been previously looted (a similar argument has been applied to Iberian cists by Del Amo-y de la Hera 1993, p. 172).

Fig. 4
Fig. 4
Full size image

(a) Halberd/dagger and palmela point in copper/bronze recovered at El Mers and Ain Dalia cist cemeteries, in the Tangier Peninsula; (b) Mers Tomb 10; (c) earliest type of Iron Age cist burial, resembling a sort of tumulus. Re-elaborated from Ponsich, 1970

Notably, in the Tangier Peninsula, Iron Age burial traditions (dated from the eighth to fifth centuries BC) exhibit continuity with those of the preceding period. Cists remained the primary burial type, often lying in the vicinity of previous Bronze Age cist cemeteries. Intriguingly, the oldest type of Iron Age cists closely resembled Bronze Age ones, namely large slabs arranged in a trapezoidal disposition, but Iron Age cists were covered by a stone accumulation, resembling a small tumulus (Ponsich, 1970, p. 90; Fig. 4c). Over time, slabs were gradually replaced by stone walls made of squared ashlars, while the trapezoidal arrangement characteristic of Bronze Age cists became progressively rectangular (Ponsich, 1970). Burial practices continued to follow a similar pattern to the previous period, with bodies placed in a flexed fetal position and scattered with ochre. In contrast, Iron Age grave goods included wheeled and handmade pottery, iron objects, decorated ostrich eggshells, and eastern-type gold and silver jewelry. Notably, some of the later cist cemeteries seem to have hosted a larger number of burials, as in the case of Djebila and Ain Dalia Kebira (Fig. 2b: 3,15), comprising 107 and 84 cists, respectively (Blaschta et al., 2024; Ponsich, 1970, pp. 72–81).

Hypogaea

In the context of changes resulting from the increasing connectivity that characterizes the early first millennium BC, and contemporaneous with the Iron Age cists, a radically new burial tradition of hypogaea emerges at the Tangier Peninsula (Koehler, 1930; Ponsich, 1970, pp. 68–71). These burials find their closest parallels among similar seventh and sixth century BC tombs at Trayamar in southern Iberia and other Phoenician settlements in the central and eastern Mediterranean (López-Pardo, 1990, pp. 33–35; Niemeyer, 1982). The Tangier Peninsula’s early first millennium BC hypogaea are located on the Achakar plateau (Fig. 2b: 1), above the famous caves, a place that may have had a certain symbolic meaning among local communities, since it was used as a ritual and burial ground from at least the third millennium BC. Intriguingly, the accompanying goods show no difference from those found in contemporary cist graves (Koehler, 1930; López-Pardo, 1990; Papi, 2019). The absence of aDNA material from these interments prevents an understanding of the genetic background of the buried individuals, though their local or mixed origins might plausibly be inferred based on the burial practices recorded.

Tumuli and Bazinas

Of all the North African burial traditions, tumuli and bazinas (the latter, a sort of truncated pyramids) are the most characteristic and widely spread types of funerary buildings, being present across the Mediterranean, Atlantic, and Saharan domains (Fig. 2a), including dozens in the Tangier peninsula (Fig. 2b), and either encountered singly or clustered into large cemeteries. As mentioned earlier, Camps supported a local North African origin for tumuli and bazinas, as opposed to other built burial traditions (Camps, 1961, p. 125). Tumuli in modern Morocco are most frequent along the Atlantic strip, but they also extend south and east, well beyond the Tangier Peninsula and broader northwest (Akerraz & El Khayari, 2016; Akerraz et al., 2020a; Akerraz et al., 1995; Baranello, 2017; Boudouhou, 1997; Ponsich, 1966; Rebuffat & Limane 2011; Souville, 1973; Stone, 2016) into the Atlas and the pre-Sahara and Sahara regions where their number is astonishingly high (Bokbot, 2019; Bokbot et al., 2008), with large concentrations of thousands of tumuli in relatively small areas, notably along the Middle Wadi Draa (Mattingly et al., 2017; Fig. 2a: 35). Although their dating is still notoriously open to debate, those northwest of the Atlas have traditionally been associated with the first millennium BC onwards, largely due to the presence of wheeled pottery among their grave goods (hence their late treatment here, despite their frequency). While Camps (1961) proposed a detailed typology based on shape and construction techniques, ranging from simple accumulations of stones to more complex bazinas (the latter exemplified in the Tangier Peninsula by Beni Maadane, Fig. 2b: 8, near the Mediterranean coast; Ghirelli, 1931), the lack of systematic excavation and radiocarbon dates hinders a better understanding of chronology and evolution. In fact, the existence of late prehistoric tumuli in the wider Sahara (Mori & Ricci, 2013; Sáenz-de Buruaga, 2013 p. 57) and Iberia (Benítez de Lugo-Enrich et al., 2014; Linares-Catela, 2020) may argue for an older origin of northwest Moroccan tumuli, perhaps dating back to the third or second millennia BC, a possibility further suggested by the fact that in northwest Africa (as in southern Iberia, Lillios, 2015), some tumuli cover one or more cist (Fig. 4c), prompting the hypothesis that tumuli evolved from an earlier cist tradition or developed independently.

Mzoura

The remarkable site of Mzoura, some 40 km southwest of modern Tangier (Fig. 2b: 37), represents a large megalithic (and possibly in part funerary) monument, encompassing at least four or five clusters of standing stones arranged atop a gentle plateau (Fig. 5a). The largest, known locally as the “Cromlech of Mzoura,” consists of 176 standing stones disposed in a circle (Fig. 5b), the tallest reaching an impressive 5.3 m, encircling a 6-m high and 55-m wide tumulus (Daugas et al., 2006). Despite the lack of a comprehensive publication of the extensive excavations in the early 1930's by Montalban, his short report suggests that the tumulus had three levels of retaining or defensive walls, with the southern side described as a “straight ashlar-built wall” (Supplementary Material 2: 2, Fig. 14a, b). The inner structure, as described by Montalban, is reminiscent of bazina-like monuments, which are generally dated to the Mauretanian period and later (Camps, 1961), a chronology further supported by the discovery of a Libyan inscription within the tumulus (Supplementary Material 2: 2, Fig. 14b). Conflicting information exists regarding a possible cist burial at the tumulus’ center (Gozalbes-Cravioto, 2018); albeit this is never mentioned in Montalban’s original report, he does describe six or seven small trilithons (Fig. 5d), a pit containing charcoal, a cattle tooth and a handmade bowl (Supplementary Material 2: 2, Fig. 14a), and a burial outside the northern side of the tumulus, which was never excavated and was subsequently looted.

Fig. 5
Fig. 5
Full size image

(a) Plan showing the position of the different clusters of standing stones at Mzoura; (b) plan of the so-called ‘Cromlech of Mzoura’, showing the extensive excavations performed by Montalban in the 1930’s; (c) aerial image of Mzoura taken probably in the 1960’s; (d) image of the Trilithons, from the inner part of the tumulus; (e) photograph of a section showing the presence of “straight ashlar-built walls” within the inner part of the tumulus. Re-elaborated from Daugas et al. 2006 (a-b); Ponsich 1966 (c); and Montalban archives (d-e)

A more recent study of Mzoura (Daugas et al., 2006) suggests that while the standing stones were shaped with stone tools, the ashlar blocks used in the construction of the tumulus walls were crafted with metallic tools. This has led researchers to propose a two-phase construction for the monument (Daugas et al., 2006, p. 764): an earlier phase for the standing stones and a later, Mauretanian, phase for the tumulus. Although undoubtedly the inner part of the structure was built as a burial in the mid-first millennium BC, further observations suggest that Mzoura was indeed part of a more complex and earlier structure that needs to be properly investigated. These include the lack of parallels for the standing stones elsewhere in northwest Africa during the Mauretanian period, the existence of similar monuments in third millennium BC Atlantic Europe, and the presence of at least three or four smaller clusters of standing stones without tumuli (Fig. 5a). Unfortunately, as ever, the absence of radiocarbon dating and fuller archaeological documentation prevents a deeper understanding of this extraordinary and under-considered monument. It is tempting to view Mzoura as a large aggregation site in use since later prehistory that was subsequently reused in the mid-first millennium BC as a burial ground for one or more prominent individuals. Such individuals may have exploited the significance of an ancestral place of memory as a powerful means of legitimation, in a context of increasing social and economic complexity during the emergence of the Mauretanian monarchy. Considering the wider landscape, while prehistoric burials at Mzoura itself cannot as yet be confirmed, the monument’s location along a major communication route in the region (Fig. 2b: 37), as well as the nearby presence of cist cemeteries and tumuli suggest that Mzoura was part of a broader later prehistoric ritual landscape.

Newly Discovered Standing Stones, Cemeteries and First Radiocarbon Date on a Northwest African Cist Grave

As part of our research, extensive field survey and excavation were conducted in the northwestern Tangier Peninsula. Survey efforts concentrated in the designated crossroad A (Fig. 1b), a major crossroads identified through GIS analysis, resulting in the discovery of three previously unknown cemeteries (Fig. 2b: 14, 39, 41). Among these, Daroua Zaydan consists of a single cist burial (Fig. 6a). In contrast, Oulad Zin comprises a solitary tumulus along with a group of at least 3 cists, situated atop two neighboring hills (Fig. 6b). Finally, Oued Ksiar, the most extensive site documented, features two tumuli and three clusters of cists (potentially four, with one of them destroyed) spread across five adjacent hills overlooking the Oued (river) Ksiar (Fig. 6c). Evidence of extensive looting was apparent across all three cemeteries, with slabs displaced from their original positions and signs of destructive excavation within some cists and tumuli. The specific period during which this looting occurred remains unknown; however, as noted earlier, looting in Morocco has been documented since the late medieval period. Some of the cists remained nonetheless apparently intact, with their covering slab in its original position. While the scarce surface materials lacked diagnostic features, the architectural characteristics of certain burial structures, such as the trapezoidal arrangement of the slabs and the presence of stone accumulations outside some of the cists (Figs. 6c and 7), hinted at a late prehistoric chronology for some of the tombs. To minimize the risk of attracting further attention to, and thus the looting of, the larger cemeteries, excavation efforts focused on Daroua Zaydan.

Fig. 6
Fig. 6
Full size image

Images of the newly discovered cist cemeteries and tumuli. (a) Daroua Zaydan; (b) Oulad Zin; (c) Oued Ksiar. Photographs by Hamza Benattia

Fig. 7
Fig. 7
Full size image

Plan and section of the single, apparently isolated cist burial excavated at Daroua Zaydan; (a-e) images showing the process of excavation of the cist; (f-g) position of the human fibula on which radiocarbon dating was performed. Photograph and drawing by Hamza benattia

Daroua Zaydan consists of a single trapezoidal cist burial, situated on a moderate hill at the bottom of the ancient Tahadart bay (Fig. 2b: 14). This bay was partially open to the Atlantic during prehistoric times and silted gradually, forming the current marshy lagoon (Khalfaoui et al., 2023). Archaeological excavation at Daroua Zaydan revealed a straightforward stratigraphy. The natural sediment layer had been dug into to introduce four vertical sandstone slabs, covered by a single large horizontal slab, with the upper parts of the vertical slabs and the horizontal slab remaining visible (Fig. 7). A crescent-shaped pit excavated outside the eastern slab, filled with medium-sized blocks (Fig. 7b–e), possibly served as an access point to the burial space, mirroring features documented at El Mers cemetery (Ponsich, 1970, p. 52; Fig. 4b). Inside the burial structure, a limited number of flint flakes were unearthed (Fig. 8a), with no human remains identified. Nonetheless, the presence of scattered human bones buried outside the structure at around 10–15 cm from the modern surface (Fig. 7f-g), suggestive of possible ancient looting, offered an opportunity for radiocarbon, collagen isotopes, and DNA analyses. Radiocarbon dating on this material indicates that the cist was in use at least since cal. at 2 sigmas 2119–1890 BC (Fig. 3), which in neighboring Iberia correlates with the Early Bronze Age, where cist burials are common (Baldomero-Navarro & Ferrer-Palma, 1984; Hunt-Ortiz et al., 2007; Lillios, 2015), and in the Tangier Peninsula with contemporaneous burials at Benzú cave (Fig. 2b: 7; Fig. 3). This marks the first radiocarbon dated cist burial in northwestern Africa, and its date helps to explain the apparent absence, so far, of earlier Beaker sherds within cists, as well as the later typological affinities of the associated metallic grave goods (Ponsich, 1970), which share similarities with items from Iberian early Argaric and Ferradeira funerary contexts, dated to terminal third and early second millennia BC (Lillios, 2015). While DNA results are still awaited, isotopic analysis on bone collagen reveals the individual’s diet was primarily terrestrial, comprising animal proteins and C3 plants, with minor contributions from marine resources (Fig. 8b).

Fig. 8
Fig. 8
Full size image

(a) Chipped stones recovered from the inner part of the cist; (b) results of the isotopic analysis on bone collagen performed on the radiocarbon dated human fibula. Photographs by Hamza Benattia

Additionally, within the framework of our survey, previously known Bronze and Iron Age cist cemeteries were revisited in order to assess their preservation and extract further information. The extensive Bronze Age cemetery at El Mers (Fig. 2b: 21) seems to have been completely destroyed as a result of urban expansion, while Bronze Age El Mriès and Iron Age Jebila (Fig. 2b: 22, 15) have both sustained severe damage. At El Mriès, an excavator has completely dug the upper hill, to a deep ranging between 3 and 4 m, destroying around 40% of the burials and exposing large quantities of archaeological materials, mostly pottery (Fig. 9). The surface collection comprises an awl of pure copper and c. 200 sherds encompassing rims, handles, and decorated pottery (Fig. 9). Some of the decorations have parallels in similar sherds from the Achakar Caves (Jodin, 1958, p. 307, plate XII 1–10), a remark already made by Buchet (1907) and later picked up by Jodin (1964a, p. 42). Additionally, these decorated sherds find their closest known parallels in pottery from Bronze Age contexts at Ghar Cahal and Kaf That el Ghar caves (Bokbot, 2020, p. 481, Fig. 12.2). El Mriès sherds (Fig. 9), although without secure stratigraphic context, are relevant for they represent some of the only known examples of pottery from cist cemeteries. Last but far from least, the recovery of a pure copper awl suggests that other metallic items at the cemetery may have gone unnoticed. Indeed, the study of Montalban’s archives indicates that El Mriès likely contained more metallic objects than previously assumed. Notably, a report from September 1952 (Supplementary Material 2: 3, Fig. 14c) mentions the looting of a dagger/halberd and several Palmela points by a person residing in a nearby village. This information significantly increases the number of known metallic items at El Mriès and doubles the recorded daggers/halberds in the region, raising the total from one to two. Furthermore, it underscores the impact of widespread looting in the region—evidenced by looting recorded at Daroua Zaydan and other cemeteries—which may have resulted in an irretrievable loss of valuable information.

Fig. 9
Fig. 9
Full size image

Decorated pottery recovered from El Mriès cist cemetery. Photographs and drawing by Pau Menéndez, Arnau Pou & Hamza Benattia

Alongside these newly discovered and revisited cemeteries, other forms of monumental structures such as standing stones offer further evidence of the ritual landscape of the Tangier Peninsula. Two standing stones were located along potential routes connecting different areas of the peninsula. One is a large pointed standing stone over 2.5 m long (Fig. 10a), placed within an important corridor that connects the Tahadart plain with the Djebel Mousa (Fig. 2b: 31). The second is smaller in size and different in shape, but a break at its base suggests it could have been part of a larger standing stone (Fig. 10b). Its position at the base of a gentle hill overlooking the interior road leading from the Tahadart to Tétouan (Fig. 12b: 44), avoiding coastally inundated areas, again suggests its emplacement is on a strategic area. The relevance of their position lies also in their location not far from the Tahadart, one of the greatest concentrations of cist and tumuli cemeteries. It is unclear whether these standing stones are part of larger structures, as in the case of Mzoura, although their disposition along routes and crossroads suggests that their distribution during late prehistory may have been more extensive, with the disappearance of most from the landscape potentially due to centuries of agricultural activities.

Fig. 10
Fig. 10
Full size image

Images of the two standing stones identified in the region of Tangier. (a) Roummane; (b) Laqueleya. Photographs by Hamza Benattia

Interpreting the Tangier Region’s Cemeteries and Burial Traditions from c. 3000 to 500 BC

While the preceding sections have focused on examining the evidence from previously known and newly discovered cemeteries, monuments, and standing stones, this section aims to interpret their significance within the broader landscape. As the reader may note, the currently available data is often limited, making it challenging to draw definitive conclusions about the social, cultural, and economic dynamics of late prehistoric northwestern Africa. Consequently, the interpretations offered here should be approached with caution, as they are drawn from the limited information available. Future fieldwork and additional data will be essential for developing a more comprehensive and accurate understanding of these topics.

A significant particularity regarding the third to second millennia BC Tangier Peninsula is its diversity of burial practices, notably cave, pit, and cist burials. Suggestively, these practices mirror those documented across the Strait, especially in areas around modern Gibraltar (Giles-Guzmán et al., 2017) and the lower Guadalquivir plain (Martín-de la Cruz & Garrido-Anguita, 2015, p. 189). Early researchers anticipated uncovering Bronze Age settlements linked to the cist cemeteries (Jodin, 1964a, p. 44), as exemplified in southern Iberia by pairings at sites such as La Traviesa in Seville and Cerro de la Peluca in Málaga (Ferrer-Palma et al., 1990; García-Sanjuán & Hurtado-Ortiz, 1998) (Fig. 2a: 32,13). In the Tangier Peninsula, however, such associations remain ambiguous. Recent investigations have suggested potential links between Bronze Age cist cemeteries and adjacent settlements (Fig. 1b), particularly at Menkal I and Oued Liane VII (Fig. 2b: 34, 44), although the lack of excavations and further research leaves such claims unverified (Bernal et al. 2016, pp. 181–182, 334). More troubling, mapping the known Bronze Age and Iron Age cist cemeteries reveals a significant concentration of cemeteries around the Tahadart (Fig. 2b), which should correspond to an equivalent number of settlements, yet despite extensive archaeological research in the area over the past 120 years, little evidence has emerged regarding third and second millennia BC settlements (Palumbo & Akerraz, 2020, pp. 67–69). This apparent absence may be attributed to the ephemeral nature of the archaeological record and subsequent erosion, or perhaps to successive multiperiod building activity, as identified at Kach Kouch (Benattia et al., 2025). Alternatively, this scenario may argue for the potential role of Tangier as a nuclear Bronze Age settlement in the region, with its remains lying undiscovered beneath the modern city (Broodbank, 2024, p. 267).

Until recently, this issue persisted in the subsequent Iron Age period, marked by limited data regarding Mauretanian settlements in the Tahadart. However, a recent research project has successfully identified several previously unknown eight to fifth centuries BC settlements (Akerraz et al., 2020b, pp. 78–79; Fig. 1b). Although mostly have been only surface explored, these newly identified sites provide a better context for understanding the relationship between Iron Age cist cemeteries and settlements. Notably, the site of Kannoua features a sixth century BC settlement (Figs. 1b & 2b: 29), which may prove to have older origins, alongside an extensive cemetery featuring both cists and tumuli (Akerraz et al., 2020a, pp. 216–218, 255). This site constitutes a strong candidate for examining the relation between Tahadart’s Iron Age and potentially Bronze Age settlements and cemeteries.

As highlighted in a previous section, the Tangier Peninsula’s terrestrial connections may have played a key role in shaping its funerary and ritual landscape. The significance of routes and connections in the distribution of Tahadart’s cist cemeteries was already noticed by Ponsich (1970, p. 40). Subsequent discoveries have reinforced this notion at a regional scale, indicating that cist cemeteries, as well as standing stones, tumuli, and rock art (see below), were often placed atop outstanding hills linked to routes and crossroads (Fig. 2b), likely acting as territorial markers (Brown, 1995). Intriguingly, no Bronze Age or Iron Age cist cemeteries have been discovered to date beyond the Tangier Peninsula, especially significative in areas of central, Atlantic, and Mediterranean Morocco known to have hosted eighth to fifth centuries BC Mauretanian settlements. This invites further consideration into whether this is simply the result of lack of research or has deeper implications, notably the existence of a sort of “funerary koine,” restricted to the Tangier Peninsula, that may be tentatively linked to social (tribal) structures, evidence of which exist in later, classical, sources (for insights see Pliny, Nat. Hist. 5, 17, and Ptolemy, Geogr. 4, 5).

Given the paucity of data, unpacking the complexity of the diachronic social, cultural, and economic dynamics of the Tangier Peninsula during later prehistory remains a challenging task. Based on existing data, namely the presence of small sites (with the exception of Kach Kouch, so far only explored through surface survey; Fig. 1b) and particularly the absence of large settlements, defensive structures and evidence of extensive social control suggest that local communities in the Tangier Peninsula may have been organized as kin-based communities constituted by small family groups. In this social context, differences in burial practices, such as variation in grave goods or tomb size (Ponsich, 1970, pp. 64–65), may reflect personal actions undertaken during the individual’s lifetime, rather than accumulation of power and wealth. This kind of social organization contrasts strikingly with that of Southern Iberia, in the Early Bronze Age, which was marked by increasing social and economic complexity, inequalities as well as walled settlements with evidence of extensive territorial control (Broodbank, 2024, pp. 345–444). Alternatively, the Tangier Peninsula’s overall settlement pattern could be understood as peripheral to a broader social and economic landscape that encompasses southwestern Iberia, or other areas to the South, with nuclear settlements located across the Strait, or under Tangier, or in the Gharb plains just to the south (Fig. 1). Significantly, similarities between Iberian cists, particularly in the Southwest, and northwestern Africa cists raise important questions about whether these parallels result from contacts and exchanges across the strait, human mobility across the same, or independent convergent paths resulting in the same phenomena. In suggestive support of the second option, DNA analysis of a second millennium BC individual interred at Loma del Puerco’s cemetery (Fig. 2a: 33), in neighboring Cadiz, revealed a mixed local and African ancestry (González-Toraya et al., 2010; Olalde et al., 2019). This finding emphasizes the importance of extracting complementary DNA data from human remains south of the strait.

Notably, first millennium BC rituals and burial grounds in the Tangier Peninsula exhibit continuity with the preceding period. While the Phoenician-inspired hypogaea may suggest foreign practices, the continuity of burial grounds on the Achakar plateau (Supplementary Material 2: 1; Fig. 13) and the similarity of grave goods with those uncovered in cist burials suggest instead that the individuals interred may have been locals or of mixed ancestries and identities. Moreover, while the presence of distinctive eastern Mediterranean grave goods within cists may indicate economic and social inequalities, the persistence of indigenous burial practices suggests that the economic, social, and cultural meanings of these objects may have been reinterpreted from a local perspective. Consequently, their perceived value and cultural significance within a Phoenician funerary context may differ from that of a local context (Papa, 2010, p. 68). This scenario sharply contrasts with the divergent path of southern Iberia, where the late second millennium BC is characterized by the disappearance of cist cemeteries and the early first millennium BC by the generalization of the cremation ritual (Brandherm & Krueger, 2017; Lorrio, 2008). It is also notable that some African Iron Age cist cemeteries appear to be concentrated in a single location and contain a higher number of tombs, possibly indicating larger settlements and demographic expansion. This contrasts with Bronze Age cist burials, which are often found either isolated or grouped into cemeteries concentrated on a single hill or dispersed across several adjacent hills. These variations may reflect differences in the social organization of these communities, with earlier cists suggesting kin-based relations.

The chronic paucity of data on contemporary settlements complicates efforts to determine whether the above-mentioned phenomena are the result of changes in the social configuration and increasing social complexity. But since at least the sixth and fifth centuries BC, the emergence of new settlements in Atlantic, Mediterranean, and central interior Morocco seems to signal an incipient process of urbanization (Akerraz et al., 2020b; Akerraz et al., 2010; Benattia, 2025; Bridoux et al., 2009; Callegarin et al., 2016; El Khayari & Akerraz, 2012; Kbiri-Alaoui, 2008; Raisouni et al. 2016). In this context, the gradual disappearance of cist cemeteries and the increasing incidence of tumuli, potentially as monumental or monumentalised burials, may be a consequence of deep socioeconomic changes in the region right at the end of our period. These changes, characterized moreover by social inequalities and the emergence of elites, gradually led to the consolidation, in North Africa, of Hellenistic-type monarchies—a path that decisively diverges, in the centuries preceding the Roman conquest, from that attested in the northwestern Mediterranean.

Further Ritual Places and Rock Art in the Tangier Peninsula and Beyond

While cemeteries and burial traditions are essential for reconstructing past beliefs and cultural dynamics, the following section explores sites that, given the nature of the ritual practices that North African communities may have performed in them, as well as the wider meaning they hold, can be considered part of the broader ritualized landscape (Fig. 11). One intriguing site on the southern edge of the Tangier Peninsula lies on the bed of the river Loukkous (Fig. 11b: 8), where the stray find of two metal objects hints at connections with Atlantic Europe in the terminal second millennium BC. One is a flat copper/bronze axe, and the second object is a Ballintober-type rapier (Fig. 12), with closest parallels in the British Isles (Brandherm, 2007). Apart from the singularity of the rapier, the only known to date in Africa, the nature of this deposition has parallels in Atlantic Europe (Bradley, 1990), with the closest at Huelva (Almagro-Basch, 1940; Ruiz-Gálvez, 1995).

Fig. 11
Fig. 11
Full size image

(a) Map showing the main ritual and rock art sites mentioned in the text; (b) map showing all the known rock art and ritual sites previously known in the Tangier Peninsula. In alphabetical order: 1. Beni Hkim 2. Boubana cup marks 3. Idol’s Cave (Achakar) 4. Ifri n’Ammar 5. Jebabra cup marks 6. Kef el Baroud 7. Kef el Kerma 8. Loukkous metallic deposit 9. Maaziz 10. Magara Sanar 11. Marsa shelter 12. Mzoura 13. Nkheila 14. Oued Zireg 15. Oukaimeden and Yagour 16. South Western Iberia’s schematic rock art 17. Tamanart 18. Toulkine shelter 19. Ziaten cup marks. Basemap: GMTED2010 and Landsat 8. Figure by Hamza Benattia

Fig. 12
Fig. 12
Full size image

 Ballintober type rapier recovered in the 1920’s from the bed of the river Loukkous and currently held at the Staatliche Museen zu Berlin, Museum für Vor- und Frühgeschichte. Photographs by Claudia Plamp

Notably, caves and shelters have long been considered sacred in North Africa (Basset, 1920), as exemplified by the so-called Idol’s Cave in the Achakar complex (Fig. 11b: 3), in which several hundred low-fired clay figurines of a late prehistoric chronology were discovered in the past century (Biarnay & Péretié, 1912; Messili, 2001). These figurines, portraying unknown schematic representations, have been associated to solar cults potentially connected to agricultural practices (Basset, 1920, p. 44) and with phallic figures linked with fertility and apotropaic rituals (Biarnay & Péretié, 1912, p. 395). Additionally, ritual practices in caves are also manifested in rock art, including paintings, cup marks, and other carvings. Before our fieldwork, Magara Sanar 1 was the only known painted shelter in northwestern Morocco (Fig. 11b: 10), featuring numerous anthropomorphs, dots, and animals (García-Hernández, 1941). Its paintings were created employing mainly a red substance (most likely iron oxides), but other colors such as yellow and white are also attested (Solís-Delgado et al., 2023). While the chronology of the Magara Sanar paintings remains unclear, evidence suggests they were created in multiple phases, ranging from the Late Stone Age to the Mauretanian Period (Solís-Delgado et al., 2023), with their closest parallels in Late Stone Age rock art sites in Cadiz (Collado-Giraldo et al., 2019; Fernández-Sánchez et al., 2021). In the wider area of modern Morocco, paintings are especially frequent in the southern desert and pre-desert areas, as well as in the Atlas Mountains (Mas et al., 2023; Moumane et al., 2019; Searight, 2004), while the only known examples in Central Morocco’s Atlantic area and the eastern Rif are those at Kef El Baroud and Ifri N’Ammar cave (De Wailly, 1975, pp. 61–72; Eiwanger, 2003) (Fig. 11a: 4, 6).

On the other hand, in the Tangier Peninsula, engraved cup marks have been previously recorded in at least three open-air sites (Boubana, Jebabra and Ziaten), and on some of the standing stones at Mzoura (Fig. 11b: 2,5,12,19; Gozalbes-Cravioto & Gozalbes-García, 2015, 2018; Koehler, 1929). In the wider area of modern Morocco, cup marks are attested at Oued Zireg, northeast of Fes (Fig. 11a: 14; Grébénart & Pierret, 1966), at Kef el Kerma, in the Atlantic Strip, near modern Casablanca (Fig. 11a: 7; Antoine, 1935), in the High Atlas Oukaimeden and Yagour (Fig. 11a: 15; Collado, 2014; Ewague, 2016, p. 63), and in the pre-desertic Tamanart Valley, south of the Atlas (Fig. 11a: 17; Skounti et al., 2019). In contrast, other type of rock carvings in the Tangier Peninsula includes lines and symbols of difficult interpretation at Marsa IV shelter (Fig. 11b: 11), on the northern slope of Jebel Mousa (Bernal et al. 2016, pp. 280–281), and on the Atlantic side (Fig. 11b: 1), near Beni Hkim (Cheddad & Bouazza, 2020). Further south, at Nkheila and Maaziz (Fig. 11a: 9, 13), near modern Rabat, three engraved standing stones (stelae) have been identified, one depicting a human figure and all three featuring concentric lines (Souville, 1973). These stelae are without known parallels in northwest Africa and differ from the southwestern Iberian Late Bronze Age tradition, which typically portrays warriors and weapons (Díaz-Guardamino, 2012). A later prehistoric chronology has been proposed for the Moroccan stelae, though their chronology remains uncertain.

Our extensive field survey in the Tangier Peninsula resulted in the identification of twenty-four previously undocumented rock art sites (Supplementary Material 3): nineteen clustered around an important crossroad (Fig. 15: 1–19) and five in the immediate vicinity of Magara Sanar 1 (Fig. 15: 21–25). Of the newly documented sites, three are open-air sites with cup marks (Supplementary Material 3: 17–19), five rock shelters with cup marks (Supplementary Material 3: 12–16), and sixteen painted rock shelters, two of which contain both paintings and cup marks (Supplementary Material 3: 1–11, 21–25). These discoveries increase the number of known painted shelters in the Tangier Peninsula from one (Magara Sanar 1 Supplementary Material 3) to seventeen, positioning the western Rif’s sandstone mountains as a substantial theater for the study of rock art. This region is extremely interesting due to its position as a linking point between Africa and Europe, which is further highlighted by the variety of depicted scenes and styles, reflecting influences from Iberia’s schematic art as well as known sites to the south.

The Tangier Peninsula’s Rock Art Sites in Wider Trans-Regional Perspective

By examining the newly discovered rock art sites (Harman, 2006; see Supplementary Material 3 for more information on the methodology employed), we can establish comparisons with similar artistic traditions in Atlantic Europe and, further south, with the Atlas and the pre-Saharan and Saharan areas. Among the newly discovered shelters with painted rock art, several of the themes are closely related with Southwestern Iberia’s Schematic rock art (Fig. 11a: 16); meandering, horizontal, and vertical lines are present across the Strait, while dotted compositions, bi-triangulars, and anthropomorphs have been recorded at Magara Sanar 1 cave as well as in southern Iberia and southern Morocco. Other motifs, such as squares with dots and lines inside, draw their closest parallels from the Toulkine shelter in the Moroccan Atlas (De Bayle-Des Hermens et al., 1984) and Saharan rock art (Moumane et al., 2019; Searight, 2004). While the two red-painted Arabic inscriptions (Supplementary Material 3, Fig. 17 h & Fig. 25e), one clearly linked to a Muslim religious term, further indicates that ritual activity within shelters persisted into historic times, it also seems likely that some of the new sites are relatively early. In this context, some of the newly discovered caves show paired dotting compositions (Supplementary Material 3, Fig. 17e & Fig. 22c-e), which in both Magara Sanar 1 and southwestern Iberia have been dated in the Late Stone Age (Solís-Delgado et al., 2023). However, it is relevant to highlight some differences with the Magara Sanar 1 paintings, for example, at the latter, depictions of horses and riders represent a relatively late feature likely dating to the Mauretanian Period (Solís-Delgado et al., 2023). Additionally, the newly documented shelters are painted exclusively in red, whereas multiple colors have been recorded at Magara Sanar 1. While dating rock art remains challenging, except in cases where post-prehistoric features have been documented, the parallels between the Tangier Peninsula’s rock art and other Atlantic and northwestern African sites suggest a multiperiod chronology, as early as the Late Stone Age.

The presence of engraved cup marks in some of Mzoura’s standing stones, as well as at open-air outcrops and rock shelters, implies that the distribution of such features over the landscape is widespread and diverse (Fig. 11b). Cup marks have been recorded on both sides of the Strait, although their arrangement in a circle, so common in Atlantic Europe, has not been documented in northwest Africa previous to the new finds (Supplementary Material 3, Fig. 28 & Fig. 31). Notably, the two newly discovered shelters containing both schematic paintings and cup marks are the first of their kind recorded in the region (Supplementary Material 3, Fig. 20 & Fig. 22), highlighting cavities as spaces used for the celebration of complex rituals. In the Tangier Peninsula, locales with cup marks have been previously viewed as sanctuaries, closely associated with ritual practices involving the use of water or liquids (Gozalbes-Cravioto & Gozalbes-García, 2015), but in Atlantic Europe, other explanations have been proposed. They have, for example, been considered in their landscape context, as marking liminal places related to symbolic interactions with the earth or other spiritual entities (Waddington, 1998), or as serving a dual role as territorial markers and ceremonial spaces (Bradley & Watson, 2019). Whatever the precise answers, similarities in rock art along Atlantic Europe, and now also including northwest Africa, may suggest a shared cultural network conveying shared knowledge, connections, or even messages among different communities (Valdez-Tullett, 2021).

Interestingly, some rock art and ritual sites in the Tangier Peninsula are located near cemeteries, suggesting an intentional spatial relationship. This is exemplified by the so-called Idols Cave in the Achakar complex, where burials have been attested. Additionally, the Achakar plateau, above Idols Cave, has served as a burial ground since the later Prehistory. Similarly, the Ziaten cup marks (Fig. 11: 19), situated atop a hill overlooking the Boukhalef plain, are in close proximity to multiple cist and tumuli cemeteries (Biarnay & Péretié, 1912; Koehler, 1948). Additionally, several of the rock art sites (Supplementary Material 3, Fig. 15: 6–7, 15–16 & 18) are situated near the newly documented cists and tumuli cemeteries (Daroua Zaydan, Oulad Zin and Oued Ksiar; Fig. 2: 14, 39, 41), often within direct visual range. For instance, Ghar Aharel (Supplementary Material 3, Fig. 15: 8, Fig. 23a) has a clear line of sight to the cists and tumuli cemetery at Kannoua (Fig. 2: 29).

Notably, the newly documented cemeteries and rock art sites are suggestively clustered around one of the most significant crossroads in the Tangier Peninsula, aligning with key prehistorical communication routes. Unfortunately, the social landscape surrounding Magara Sanar remains largely unexplored, though its location along a major valley suggests it may have been situated within a significant communication corridor. While the specific function of some of these ritual sites and their relationship with cemeteries remains difficult to determine, the available evidence suggests parallels with similar practices in neighboring regions. For instance, in Iberia, rock art and burial sites often occur in close proximity (Lazarich et al., 2010, and particularly the numerous publications by Primitiva Bueno-Ramírez: e.g., Bueno-Ramírez et al., 2009, 2022). Further south, in the pre-desert areas of the Middle Draa and the Tamanart Valley, rock art is often found near burial structures (Mattingly et al., 2017; Skounti et al., 2019), reinforcing the potential symbolic connection between these two phenomena. Although Tangier Peninsula’s rock art studies and its relationship with the wider social landscape are still in an embryonic stage, the evident connections of this area with Atlantic Europe, and the Sahara, may yet prove pivotal for fully understanding the shared cultural dynamics and trans-regional connections.

Conclusion

From around 3000 BC, northwest Africa’s Tangier Peninsula experienced increasing interactions with other regions, facilitated by its strategic position as a gateway between Europe and Africa, as well as the Atlantic and Mediterranean. This growing interactivity—initially through contacts and exchanges and later potentially through marriages and migration—is reified in a rich record of prehistoric monuments, burials, rock art, and ritual places that span anachronistically imposed continental divides. The closest parallels to these practices can be found in both late prehistoric Iberia and the Sahara (and pre-Sahara), as evidenced by a modest but robust dataset of sites, archaeological materials, and, so far, slight clues from aDNA.

This paper has summarized the available evidence and presented new data from recent fieldwork, suggesting that a complex and evolving ritual landscape existed in the Tangier Peninsula between approximately 3000 and 500 BC. This landscape included a wide array of ritual sites, rock art, and burial practices, including pit burials, cave burials, hypogaea, cist, and tumuli cemeteries. These heterogenous burial practices find their closest known parallels across the Strait in southern Iberia, where similar practices have been attested, raising questions about whether this reflects shared social and cultural dynamics between the two regions. While since the later second millennium BC cist burials disappear in southern Iberia, replaced in the subsequent millennium by the cremation ritual, in contrast, the Tangier Peninsula shows continuity since at least 2100 BC to as late as approximately 400 BC. Moreover, the discovery of new cist cemeteries, often along major crossroads, suggests that this burial phenomenon is more complex and widespread within the Tangier Peninsula than previously thought and that cemeteries may have served as territorial markers for communities competing for the land in a context of increasing territoriality. Although evidence of late prehistoric settlements exists in the Tangier Peninsula, identification of paired open-air settlements and cemeteries remains a milestone still to be achieved for northern Africa’s late prehistoric research, in contrast to the situation across the Strait, where the association of settlements and cemeteries is a commonplace. In northwestern Africa, cist cemeteries were largely confined to the Tangier Peninsula and seem to have been replaced by tumuli from the mid-first millennium BC. While most tumuli in northern Morocco have been dated to the first millennium BC, the existence of third and second millennia BC tumuli in Iberia and the Sahara suggest that tumuli in the Tangier Peninsula may have an older origin than previously assumed. Additionally, the presence of cist burials within some tumuli in both North Africa and southern Iberia raises questions as to whether tumuli evolved from earlier cist burial traditions or developed independently. Most likely, tumuli were built as monumental or monumentalized burials, potentially reflecting social and economic changes linked to the emergence of new forms of political organization in the region. Admittedly, the paucity of data regarding northwest Africa’s late prehistoric communities hinders a deeper understanding of the region’s long-term social, economic, and cultural dynamics. However, understanding the dynamics of the second millennium BC remains crucial for assessing ensuing regional developments during the subsequent Mauretanian Period, the latter characterized by the consolidation, in the southwestern Mediterranean, of Hellenistic-type monarchies.

The importance of investigating local communities in the Tangier Peninsula is further highlighted by the continuity of ritual, economic, and cultural practices. Such observations invite re-evaluation of longstanding core concepts such as colonization and acculturation, as well as the role of external influences in shaping North African dynamics (Benattia, 2025). When Phoenician communities arrived and settled in the area during the early first millennium BC, they assuredly encountered not a passive local populace, but rather a network of settlements and individuals with extensive Atlantic and Mediterranean connections and agency, already practicing a full farming economy based on cultivation of domesticated species and animal husbandry (Benattia et al., 2025). The Phoenician presence initiated a long and complex process of negotiation with local communities that probably involved resistance, assimilation, and hybridization of cultural practices as seen both in Kach Kouch and the ritual practices and burial traditions of the first millennium BC.

In addition to cemeteries, which offer insight into the funerary practices of the Tangier Peninsula, standing stones, rock art, and ritual practices in rivers and caves suggest a broader ritual and social landscape that remains largely underexplored. These elements have traditionally been studied separately, yet their spatial associations suggest potential interconnections, particularly in their placement along major communication routes and crossroads, as suggested by GIS analysis. While these spatial patterns further highlight their possible ritual and territorial significance, the limited available evidence does not allow for firm conclusions, and additional research is needed. Indeed, while the rock art research focus until now has been on the Atlas, pre-Saharan, and Saharan areas, the newly discovered rock art sites (Supplementary Material 3) underscore the importance of this region for the study of prehistoric imagery, indicating that northern Morocco’s Rif Mountains may host hundreds of undocumented painted shelters. Tangier Peninsula’s rock art is particularly relevant due to its clear parallels both with southern Moroccan rock art and sites in Atlantic Europe, particularly in southern Iberia, potentially indicating longstanding contacts between these regions.

Finally, the extensive damage recorded at both newly identified and previously known sites in the Tangier Peninsula, caused by looting, urban development, and agricultural practices, underscores the urgent need for conservation efforts to protect the rich heritage of this region. The destruction of these sites would not only erase invaluable examples of northwest Africa’s cultural heritage but may also hinder our ability to fully grasp the complex social, economic, and cultural dynamics of this region. By publishing this new data and assessing what we know and still do not know about northwest Africa’s later prehistoric ritual and funerary landscapes, we hope to propel further research on one of its least investigated periods and prompt a reassessment of the region’s place within the broader dynamics of the late prehistoric western Mediterranean and inner Atlantic.