Previous studies

Surveys and excavations in Tripolitania have rarely focused on cemeteries but when they were the subject attention was oriented to specific elements. For example, Richard Goodchild studied the inscriptions from a large cemetery in the Wadi Dreder (Goodchild Reference Goodchild1954; Reference Goodchild and Reynolds1976), stating that ‘the tombs of Bir ed-Dreder show very little variation of type’ (Goodchild Reference Goodchild and Reynolds1976, 62). This was later disproved by the ULVS, which demonstrated that this cemetery was far more complex with at least six different grave types being evident (Buck et al. Reference Buck, Burns and Mattingly1983, 49–50; Mattingly Reference Mattingly1996, 78–79, Dd001; Mattingly and Dore Reference Mattingly, Dore and Barker1996, 146–48). Also, Olwen Brogan's investigation of funerary remains in Tripolitania largely focused on mausolea and funerary furniture. While noting ‘lesser’ graves at some sites (such as at a cemetery in the Wadi el-Amud: Brogan Reference Brogan1964, 47–56), or the presence of large rectangular graves at others (such as in the Wadi Maymun: Brogan Reference Brogan1976–1977, 105–06), many funerary remains went unrecorded. For instance, no mention was made of the 33 circular graves next to the three obelisk mausolea at Mselletin in the Wadi Merdum (Site 1–3: Brogan Reference Brogan1976–1977, 106; Mattingly Reference Mattingly1996, 173, Md001).

One of the most significant studies on cemeteries in the pre-desert was conducted by Brogan and Smith at the Roman-period settlement of Ghirza (Brogan and Smith Reference Brogan and Smith1957; Reference Brogan and Smith1984; Smith Reference Smith, Buck and Mattingly1985). Next to the two already-known mausolea-cemeteries they recorded additional undocumented funerary areas, giving in total five major cemeteries and numerous outlaying graves. Overall, they estimated that at least 412 graves were evident, while many others had been destroyed by sand or human activity. Nevertheless, the focus of their work at Ghirza clearly was on the monumental tombs and the settlement (Brogan and Smith Reference Brogan and Smith1984).Footnote ⁴

The ULVS was the first team to systematically survey and record large areas of the Tripolitanian pre-desert, albeit with a focus on the structures of the living such as farms and wadi walls. They investigated and published numerous cemeteries in varying levels of detail, including several cemeteries in the Bir Scedua area (Buck et al. Reference Buck, Burns and Mattingly1983, 45–49; Mattingly Reference Mattingly1996, 48–55, Bs010, Bs011/012, Bs072), a cemetery in the Wadi Dreder, (Buck et al. Reference Buck, Burns and Mattingly1983, 49–51; see also Goodchild Reference Goodchild1954; Mattingly Reference Mattingly1996, 78–79, Dd001), as well as a cemetery in the Wadi Umm el-Kharab (Mattingly Reference Mattingly1996, 149, Kh9511–9515; Welsby Reference Welsby1992, 42–45). Other cemeteries were recorded in more detail but not fully published. Nevertheless, the ULVS team made some important general observations in relation to the funerary landscape and its features: they noted that circular graves covered by a stone mound were the most common burial type across the pre-desert (see Davies Reference Davies2013 and Gatto et al. Reference Gatto, Mattingly, Ray and Sterry2019 for further discussion on the subject), although variations existed such as rectangular graves and platforms; structural differentiations were loosely linked as indicators of sociological differentiations amongst the burials (Jones Reference Jones, Buck and Mattingly1985, 286–87); and they observed that many cemeteries were placed in highly visible locations, sometimes a considerable distance from the settlement, and, most likely, marked property boundaries (Buck et al. Reference Buck, Burns and Mattingly1983, 53; Mattingly Reference Mattingly1995, 149; Mattingly and Dore Reference Mattingly, Dore and Barker1996, 144).

Mausolea sites and cemeteries: an overview

The significance of considering mausolea within their cemetery context is highlighted when we examine the archival data for surveys in the Tripolitanian pre-desert, which has provided evidence for 52 mausolea sites within the region. The majority (34 sites, 65%) can be classed with certainty as either being part of a wider cemetery or having a cemetery associated with them (shown in the supplementary material Figure 1). Seven mausolea sites (13.5%) are classed as ‘probable’ since satellite imagery suggests that there are additional features present, although we cannot verify that they are funerary features. For another seven (13.5%), such associations are ‘unknown’ because the exact locations of the sites are unclear. This means that only four mausolea sites (8%) have no associated burials. Some of the cemeteries that surround the mausolea are substantial, such as An001 in the Wadi Antar which holds at least 203 individual tombs (shown in the supplementary material Figure 2) or Kn005 in the Wadi Khanafes which covers an area of approx. 100m² (An001: Mattingly Reference Mattingly1996, 33; Kn005: Mattingly Reference Mattingly1996, 152–53). Others are much smaller and consist of just a few graves such as in the Wadi Lella and at Mm004 in the Wadi Mimoun (Wadi Lella: SLS Archive, Brogan Collection, unpublished notes Wadi Lella; Mm004: Mattingly Reference Mattingly1996, 213).

Cemetery layout

The arrangement of cemeteries is not a standardised component of the funerary landscape, with topography and chronology playing a major part. The influence of the latter is compounded when a long history of activity at a site repeatedly obscures previous patterns, as seems to be the case at An001 in the Wadi Antar, which demonstrates a continuity that runs from prehistory to the Islamic period (Mattingly Reference Mattingly1996, 33–39). An001 exemplifies the variety of funerary structures present in pre-desert cemeteries, including small circular graves, large circular tombs, cist graves, rectangular graves, rectangular platforms, mausolea, and numerous Islamic tombs in the north-western part of the cemetery (Figure 2, supplementary material).

Some cemeteries indicate structures in a linear alignment that represent either a familial lineage or clear visibility in the landscape, or both (Buck et al. Reference Buck, Burns and Mattingly1983). For example, the site of Ag002 in the Wadi Umm el-Agerem contains six mausolea, labelled A–F, dating to the second to fourth centuries AD and arranged in a north-south line;Footnote ⁵ mausolea A and E both exhibited inscriptions that indicate a familial relationship between individuals (shown in the supplementary material Figure 3; Mattingly Reference Mattingly1995, 162–66; Reynolds Reference Reynolds1955). To the east of these monuments is an often-overlooked ‘group of minor tombs’ (Mattingly Reference Mattingly1996, 20), comprising 10–12 rectangular tombs, a single circular grave, and eight or nine groups of upright stones (Mattingly Reference Mattingly1996, 23, Fig.1.6; SLS Archive, Brogan Collection, unpublished notes, Wadi el-Agerem). Approximately 130 m to the northeast of Ag002 is a second cemetery (Ag003), interpreted as lower status than Ag002 (Mattingly Reference Mattingly1996, 20), which displays a similar predominant linear arrangement, with additional features also extending on the eastern sides.Footnote ⁶ It is not possible to make definitive statements relating to the extent to which these two sites were associated but the similarity in arrangement does invite some speculation. Could it be that the mausolea were adhering to existing cultural traditions, represented by Ag003? This is clearly something that needs further consideration.

It is evident that although mausolea usually occupied prominent positions they did not necessarily represent the earliest funerary features. Instead, it was not uncommon for mausolea to be added to already existing sites, including cemeteries, (or close to them) and integrated into the funerary landscape. The large cemetery at An001 in the Wadi Antar incorporates prehistoric features such as hunting walls, and flints are scattered across the site. It is likely that this cemetery is much older than the late first or early second century mausoleum (for the dating of the mausoleum see Brogan and Reynolds Reference Brogan, Reynolds, Buck and Mattingly1985, 20). Cemetery 3 at Ghirza provides one of the clearest examples for the addition of mausolea to existing areas of funerary activity. Pottery from the first century AD was found between the burials but the earliest mausoleum dates to the mid-third century and was placed at the edge of this cemetery, 2 km south of the main settlement (Brogan and Smith Reference Brogan and Smith1984, 100–18, 185). Here, the cemetery probably functioned as a territorial marker, and the owners of the obelisk mausoleum strived to form a connection with the earlier inhabitants of the settlements (Jones and Barker Reference Jones and Barker1983, 53; Mattingly with Flower Reference Mattingly, Flower and Barker1996, 188–89).

There are also examples where the mausolea are clearly not the last features to be added to a cemetery, further highlighting the longevity of these sites. There are numerous instances of architectural fragments from mausolea, or other monumental tombs, being re-purposed in later burial structures. For example, mausoleum masonry was probably re-used at Khanafes (Kn005) and Antar (An001h), and a section of broken column – very probably originating from nearby Mausoleum B (An001f) – was re-purposed as a stele for a nearby grave.

While mausolea and smaller tombs were located within one cemetery, the sheer size and embellishment of the mausolea created a division between the more powerful elite families and the rest of the population (see Pollock Reference Pollock1999, 175 for similar ideas). Furthermore, the spatial distribution of tombs in mausolea-cemeteries indicates that the monumental burials had some form of exclusion zone around them. Satellite images and maps suggest that minor burials were not placed against, or next to, the mausolea. For instance, at Umm el-Agerem (Ag002) at least ten rectangular graves are located approximately 100 m to the east of the mausolea, but no burials have been recorded immediately adjacent to the monument. At Ghirza, Brogan and Smith make no mention of burials located between the mausolea. However, there is no evidence of a physical division such as is the case in neighbouring Africa Proconsularis, where walls surrounded mausolea within cemeteries, such as at Segermes (Dietz et al. Reference Dietz, Sebai and Ben Hassen1995, P11-2 and O10-2) or at Leptiminus (Ben Lazreg and Mattingly Reference Ben Lazreg and Mattingly1992, Mausoleum 20; see also Moore Reference Moore, Stone and Stirling2007, 91; Sterling Reference Sterling, Stone and Stirling2007, 132). The exclusion zone around mausolea may not only have created a physical division between the elite and their subordinates, but may also have functioned as a ritual space to venerate the dead.

Veneration of the dead

The veneration of the dead already had a long-standing tradition before the Roman period, which continued throughout this period and perhaps beyond,Footnote ⁷ evident by the presence of tomb furniture such as offering tables or libation bowls. Many examples of offering tables have been found in cemeteries in the Tripolitanian pre-desert, occurring with both mausolea and other grave types (e.g. at Antar, An001; Khanafes, Kn005).Footnote ⁸ The presence of such funerary furniture, and libation channels in mausolea, show that small graves as well as mausolea were continuously revisited, reflecting the importance of ancestral worship within North African funerary practices while also highlighting that it was not exclusive to the prosperous, leading families (Mattingly Reference Mattingly1995; Reference Mattingly2003; Reference Mattingly, Stone and Stirling2007b). Reliefs on the mausolea include imagery linked to ritual and memory to venerate the dead, be it in the form of figures depicted in gestures of prayer or holding offering vessels. Inscriptions from the mausolea at Ghirza also suggest the veneration of the dead was very much a communal affair,Footnote ⁹ which was probably not exclusive to the monuments of the elite. People clearly engaged with both the graves and the monumental tombs, creating a social relationship through the acts of mourning, ritual and remembrance that took place in the cemeteries (Assmann Reference Assman2011, 149; Osborne Reference Osborne and Osborne2014, 10, 13; Wendrich Reference Wendrich and Osborne2014, 412). This is also reflected further south in Fazzan where offering tables have been widely recorded in association with non-monumental funerary structures in the Wadi al-Ajal (Mattingly Reference Mattingly2003; Reference Mattingly2007a; Reference Mattingly, Stone and Stirling2007b; Mattingly et al. Reference Mattingly, Sterry, Ray, Gatto, Mattingly, Ray and Sterry2019b).