The study of the Indian Ocean trade in Antiquity has increased noticeably during the last few years in most regions involved in the system, from Egypt to India.¹ In comparison with other areas, the coasts of the Horn of Africa have received little attention, with the exception of the ancient port of Adulis, in Eritrea.² The Horn, however, figures prominently in the Periplus of the Erythraean Sea, written during the 1st century AD; several ports are mentioned along the coasts of Eritrea, Somaliland and Somalia. Previous archaeological research has revealed the great potential of the Somali coast for exploring commercial relations between African communities and the Roman and Parthian empires. Excavations in Ras Hafun, south of Cape Guardafui in the Puntland region of Somalia, revealed two important sites with a rich assemblage of imported pottery from the first half of the 1st millennium AD: Hafun West is dated to the 1st-century BC-3rd century AD and Hafun Main from the 2nd to the 5th centuries AD.³ The materials came from the Roman Empire, the Gulf and South Asia. Also in Puntland, another site with Roman materials was discovered in Daamo.⁴ In Somaliland, the presence of Roman merchants is known since the late 19th century, when French explorer Georges Révoil discovered an assemblage of Roman glass and pottery in the cairns of Heis, a locality currently known in Somali as Xiis.⁵ The site was visited by Neville Chittick in 1976, who found Roman glass “of about the 4th century AD” in a small robbed cairn, and potsherds on the surface of the site that he dated between the 2nd and 5th centuries AD.⁶ G. Révoil’s glass finds were studied in 1986 and the entire collection was the subject of a monograph in 1993,⁷ but nobody did fieldwork at the site again until 2017‑2018, when two surveys were conducted independently. One of them was carried out by Carl Phillips on behalf of a construction company aiming to build a new road leading to a future harbour.⁸ The other survey was conducted by our team, as part of a long-term project on long-distance trade in Somaliland, from the 1st to the 20th centuries AD.⁹ Fieldwork continued in 2019, during which we carried out excavations in four cairns,¹⁰ but was interrupted in 2020, due to tropical storm Pawan destroying the road leading to Xiis. At the moment of writing, plans are underway to resume work in 2021. The main aim of this chapter is to present the imported materials documented during recent archaeological work at the site. We will focus on our finds from the 2018 and 2019 field seasons, and will use G. Révoil’s and C. Phillips’ materials for comparison.

The site, or rather sites, of Xiis are located on the coast of the Sanaag region, eastern Somaliland. For its location, topography and archaeological finds, Xiis has been traditionally identified with one of the “far-side ports” of the Periplus of the Erythraean Sea, either Mundu or Mosyllon. Most authors tend to agree that the best match would be Mundu, as the Periplus describes an island close to the coast, identical to the one existing in Xiis:¹¹ the rocky isle lies less than 200 meters off the shore. The text referring to Mundu in the Periplus is the following:

From Mâlao [Berbera] it is two runs to the port of trade of Mundu, where vessels moor fairly safely at the island that lies very near the shore. This place offers a market for the aforementioned and, similarly, from it is exported the aforementioned merchandise [myrrh, frankincense, cassia, and slaves] plus the incense called mokrotou. The traders who live here are rather hard bargainers.¹²

The main site is a large necropolis consisting in over 300 cairns of different shapes and dimensions and extending for 1,800 meters along the coast, between the aforementioned isle and the mouth of the wadi El Usbale (fig. 1). The greatest concentration of tombs lies in an area of around 800×250 meters extending between the coast and a rocky plateau 145 meters above the mean sea level and known as Ma’ajilayn. On the plateau itself, there is a second site, where a diversity of imported materials has been found, including Roman amphorae and South Arabian containers, but no structures. Some amphorae were already collected here by G. Révoil.¹³ There is a third site, located in the rocky isle mentioned above. The isle is very steep and only a small, flatter area to the south could have been used for habitation. Here, a rectangular platform of 15 meters made of coral stones was documented and two groups of stone terraces nearby, perhaps to adapt the terrain for setting up huts or tents. The place yielded plain, wheel-made and handmade pottery and amphora, but very little in the way of fine wares (two fragments of turquoise-glazed pottery) or glass (one pushed-up base in green glass). This is the only site that seems to have been visited in medieval and post-medieval times, as proved by some pottery sherds and a medieval or post-medieval polychrome glass bangle.

We conducted a systematic survey of the necropolis, in which we numbered, georeferenced and documented each cairn individually and recorded all surface finds in their surroundings. The entire necropolis was mapped using a drone and total station. Imported materials were found in association with 80 cairns, Parthian turquoise or green glazed ware being the most common find (46 of the tombs), followed by glass (N=25), amphorae (N=19) and undecorated wheel-made wares (N=15). Seven cairns yielded stone or glass beads. The rest of the site was quite devoid of materials, except from the south-central area, at the feet of Ma’ajilayn hill, where we recorded an artefact scatter with Egyptian fine ware, Roman amphorae, millefiori shards, and glass inlays.¹⁴ In 2019, four tombs were also excavated, two intact, one partially looted and one that had been completely looted. One of the intact tombs and the one that had been partially looted yielded no funerary goods whatsoever, only the remains of one individual each, whereas in the other intact grave we found a skeleton of a pre-adult individual with a large Roman glass bottle dated to the 3rd century AD, iron anklets and a necklace with glass and bronze beads. While we have been unable to infer the biological sex based on the osteological remains, the artefacts to which the skeleton are associated have been traditionally linked in the Horn to female individuals. Attempts to date the bones of the tombs with no associated materials were unsuccessful as none of the bones had any collagen left and their chronology remains uncertain. The looted tomb, which has been described in detail elsewhere,¹⁵ yielded a rich assemblage of artefacts from the 1st century AD, including millefiori glass, a Roman ribbed glass bowl, Parthian glazed pottery, terra sigillata, Egyptian fine ware, and decorated glass inlays that were part of a box (see below).

The collection of imported wares and glass described here comes mainly from the 2018 survey, as the finds from 2019 are still under study, so we will provide only some general comments on this latter material. The finds from 2018 come from two areas: the central sector of the necropolis and Ma’ajilayn hill. As a complement to our research, we have the materials collected by C. Phillips,¹⁶ which confirm our own data and observations. The items collected by G. Révoil, mostly glass, will be mentioned for reference. Imported materials make the immense majority of our finds. In the necropolis, we have only found a sherd of undiagnostic handmade pottery that seems to have been produced in the region. In the island, there were six sherds of handmade pottery, grey-black inside and brown-red outside, with abundant granitic temper. This represents less than 1% of the entire documented assemblage.

Amphorae are one of the most common imported elements in Xiis, despite their scant representation in the material gathered by G. Révoil. Both our survey and C. Phillips’ delivered an important number of them.¹⁷ The most common types are Italian Dressel 2‑4, mostly from Campania, with characteristic “black sand fabric” (fig. 2: 1, 8), but there are also other kinds of containers with slightly different fabrics, also with volcanic sand (fig. 2: 3, 10) and other Italian productions with coarse igneous rock inclusions (fig. 2: 7). Within the Dressel 2‑4 family, we have an individual of probable Cilician origin (fig. 2: 2) and a rim (fig. 2: 4) with fabric containing quartz, mica, some limestone and a little iron oxide compatible with the Rhodian 1 fabric of Aegean Dressel 2‑4.¹⁸ We have been unable to see C. Phillips’ materials, but his descriptions lead us to believe that there is also a variety of provenances, including Aegean and Cilician examples.

Gaulish amphorae are represented by a rim and a handle, possibly of the Gauloise 4 type (fig. 2: 5, 9) with its typical calcareous fabric and whose production is to be located in southern Gaul, between Languedoc and Provence.¹⁹ Lastly, there is an important volume of undiagnostic Egyptian amphorae, whose presence in the site is corroborated by a rim of the type AE 3‑1.5 (fig. 2: 6).²⁰ G. Révoil also recovered several fragments of amphorae,²¹ including a Campanian Dressel 2‑4.²² The amphorae documented by C. Phillips in his 2017 survey coincide with ours; in his collection, Italian (mainly Campanian) amphorae prevail, along with undetermined productions of Dressel 2‑4, Rhodian and Gaulish amphorae.²³ The only novelty is a massive conical base of Agora M54 type,²⁴ which he documented in Ma’ajilayn.

Several sherds from the same vessel of Italian terra sigillata were found in the south-central cairn zone. The vessel (part of a small hemispheric bowl) could belong to the Conspectus 34.1 type, like the two pieces from the Révoil collection (fig. 4: 1‑2),²⁵ which were also recovered in the area of the necropolis.²⁶ Maurice Picon believed that these pieces may come from Arezzo and our fragments have indeed a fabric compatible with those workshops. Another type of sigillata than has been documented is Eastern Sigillata A (ES A) from the area of Antioch (Syria). A base of a vessel was discovered during the 2018 survey at the feet of the Ma’ajilayn hill (fig. 4: 1‑3) and two other fragments were found by G. Révoil in the necropolis.²⁷ The excavation of Tomb 153 in 2019 yielded fragments belonging to two different bowls of Italian sigillata.

Fragments from at least two bowls with red polished surfaces (fig. 4: 4) were found in the necropolis, one of them in a looted grave (Tomb 153). We initially mistook them for Indian red polished ware.²⁸ The pottery has micaceous fabric with abundant inclusions of very small size, some opaque quartz and a high amount of small black inclusions. The walls are coated with burnished red to brown slip. Wheel marks are visible on the surface. Both the base and the surface treatment are compatible with an origin in an Egyptian workshop on the Nile.²⁹ Similar dishes and bowls have been found in Hafun West³⁰ in Puntland, as well as in several Red Sea sites, such as Coptos and Myos Hormos,³¹ dated to the 1st century AD. However, although the Egyptian origin is beyond doubt, the vessels from Xiis are of fine and very well levigated fabric, with no vegetal temper, unlike the pieces from Hafun,³² which likely come from another workshop on the Nile.³³

Common and cooking pottery is also present in Xiis, although in lesser quantities than amphorae and glazed wares. Within the cooking wares, there are some bowls with polished, red-slipped interior (fig. 4: 6) and burnt traces on the outer surface. The fabric is coarse and has fine muscovite, a little of biotite, some dark mineral inclusions and shells. The presence of slip and the grey core, typical of irregular firing, makes it similar to the productions of the Nile Valley, although shells are uncommon there. Other possibilities should not be ruled out, such as the Persian Gulf, where a shell-tempered production has been documented in imitation of Indian wares.³⁴ A more certain Egyptian origin can be assigned to another red-slipped bowl (fig. 4: 5). The alluvial, micaceous fabric, with microfossils and a grey core are typical of the fine cooking wares of the Theban region.³⁵ We also have two jugs or bottles with alluvial fabric, typical of Egyptian productions (fig. 4: 7‑8). This kind of vessels traveled along the Red Sea from production centers on the Nile, as evinced by their presence in Ras Hafun³⁶ and Kanē, in Yemen.³⁷ Also from Egypt come several fragments that may be identified as kegs (siga in Arabic), cylindrical containers with thin walls, red fabric and grey surfaces. These vessels have been documented in the Kharga Oasis from the Ptolemaic period onwards³⁸ and are common in contexts dated as late as the early 2nd millennium AD along the Hadramawt coast,³⁹ but their production and export has been documented. It would not be strange to find it in Xiis. We have been unable to find parallels for a wheel-made pot with fabric similar to Cilician productions and yellow slip, but we cannot rule out other possible sources, except Aegean, which it is clearly not (fig. 4: 9). Finally, there is a fragment from a local/regional, handmade cooking pot (fig. 4: 10), with granitic temper (quartz, mica and feldspar) and dark fabric. We have found this kind of pottery in indigenous contexts from the 1st millennium AD to the 13th century in the Berbera region.

Mesopotamian glazed ware is very abundant in Xiis, with over a hundred sherds documented so far; during the 2019 survey, 46 cairns delivered this kind of pottery, with a proportion of 1 green to 4 turquoise/blue sherds. In the surveys of 2017 and 2018, however, the ratio was the reverse. The difference in color is in all likelihood original and not the result of degradation. The same cannot be said of a few beige-glazed sherds that might have been either green or turquoise originally. Finds of glazed wares came mostly from artefact scatters not directly associated with specific tombs. The fabrics of green and turquoise or blue-glazed wares are very similar (fig. 5). Perhaps the latter have slightly fine fabrics than the green, although this might be due to an intentional selection of the clay depending on the vessel. Generally speaking, their fabrics are yellow or light brown and sandy with large quantities of very fine mineral temper (mainly quartz and small red inclusions). The origin has to be found in southern Iraq, in the region of Basra.⁴⁰ The pieces of Xiis could be the Parthian forebears of TURQ.4 (green) and TURQ.5 (blue) proposed by Derek Kennet from the medieval contexts of early medieval Ras‑al‑Khaimah.⁴¹ The shapes documented at Xiis are limited to middle-sized jugs or bottles represented by diagnostic fragments of rims, handles, necks and bases (fig. 6: 1‑7). In fewer quantities, we have also documented bowls (fig. 6: 8‑9). The most remarkable element in the assemblage is a green-glazed ovoid vessel with high footing (fig. 6: 10) of which we found 23 sherds. The fabric is different from the rest of the pieces from the assemblage, because alongside fine quartz sand, it has vegetal inclusions (straw), like the South Arabian ovoid jars. A similar base was documented in the necropolis in the 2017 survey. A good parallel was excavated in Berenike.⁴² In our case, the glaze covers both the interior and the exterior of the piece in an irregular way, but the original green color can still be seen. Imports of glazed ware are common in the ports of the Gulf of Oman and the Red Sea and are present from Hellenistic to early medieval contexts.⁴³

An important quantity of South Arabian organic storage jars have been found.⁴⁴ These elongated, ovoid containers are very common throughout the Red Sea, the Arabian coast and India, where they have been recently detected in Pattanam.⁴⁵ Different kinds of fabrics have been described, belonging to different production areas, all of them probably in South Arabia⁴⁶ and sharing straw as the main type of temper. The two most common fabrics, based on data from the coastal sites of Kanē (Yemen) and Sumhuram (Oman) are light red – compact, pinkish and with lighter color outside –⁴⁷ and yellow-greenish, which is more porous than the other.⁴⁸ In Xiis we have identified several individuals with these fabrics; light red (fig. 7: 1‑2) being more common than yellow-greenish, of which we only have two individuals. We have also recovered organic storage jars of two other types. One of them is beige and has a very porous, micaceous fabric, with abundant straw (fig. 7: 3‑5), probably a variant of the yellow fabric, but more micaceous. This type of fabric probably comes from South Arabia as well and it has been documented (marginally in comparison with more common fabrics) in Sumhuram, Pattanam and Kanē.⁴⁹ Two high footings have also been documented in a totally different fabric, with inclusions of igneous rock alongside the typical organic temper (fig. 7: 6‑7). This temper has not been documented, as far as we can tell, in other sites of the Red Sea. Based on the morphology of the bases, it seems that the vessels were also elongated and ovoid. The presence of igneous temper may indicate an origin in southern Yemen or Oman, where basalt deposits are common.⁵⁰ The only rim that we have been able to locate (fig. 7: 1) is of the light red type and probably Variant A of the typology proposed by Vittoria Buffa.⁵¹

A small fragment of a rim from a white alabaster vessel was found, which is the second to be found in Xiis (fig. 8: 1). G. Révoil collected the base of an amphora-shaped unguentarium.⁵² The rim diameter (ca 8 cm) of our piece, however, indicates that it does not belong to this kind of vessel. In any case, the thin walls and the everted rim (forming a mushroom-shaped mouth) suggest that it was probably also a perfume container, of larger size and globular shape. It likely came from Egypt, where the majority of alabaster workshops are found, although another origin cannot be completely ruled out.

Glass is very abundant at Xiis, particularly in the necropolis, where it was part of the funerary offerings of the tombs. Glass was found in the surveys of 2018 and 2019, but also during excavations of the latter field season and therefore in context. Glass makes the greater part of the Révoil collection.⁵³ Its provenance is the workshops of Alexandria, which exported it throughout the eastern Mediterranean and the Red Sea. It was transported first by river to Coptos, then overland to Myos Hormos (or Berenike) and finally by sea toward sub-Saharan Africa, Arabia and India.⁵⁴ However, other origins in the Mediterranean are also possible, including Rome itself, or even in the Red Sea: the Periplus mentions Muza as a place of production of polychrome glass.⁵⁵ We can divide the assemblage into three groups: monochrome and mosaic vessels; inlays, plaques and tiles; and glass beads. The assemblage is extremely similar to G. Révoil’s, as all these groups are present in his collection.⁵⁶

Several glass vessels have been found during survey and excavation. The most remarkable piece is a complete flagon with ribbed handle in green glass (fig. 8: 2) coming from Tomb 120 and associated, as mentioned above, to the remains of a pre-adult individual. Glass bottles have been found in Early Roman contexts in Egypt, like Myos Hormos,⁵⁷ but quadrangular shapes are much more common.⁵⁸ An identical parallel, however, comes from the Royal Ontario Museum. John W. Hayes proposes a chronology of the 3rd century AD and an origin in Syria or Cyprus.⁵⁹ Apart from the complete piece, the neck of another, smaller one was found, also in green color (fig. 8: 4) and comparable to a vessel recovered by G. Révoil.⁶⁰

Open monochrome vessels include a pillar molded bowl in translucent blue-green glass (fig. 8: 3), similar to one recovered by G. Révoil.⁶¹ This is doubtless the most popular of the glass vessels in the Red Sea during the first half of the 1st century AD.⁶² It has been documented in the Omani coast, in Khor Rori,⁶³ and as far away as Afghanistan and India.⁶⁴

During the survey of the necropolis, a bowl in olive green glass was found with a flagon-like glass appliqué (fig. 8: 5), the rim of a jug in green glass (fig. 8: 6), with parallels in Myos Hormos,⁶⁵ a plain base in blue glass, probably from a bottle (fig. 8: 7), and a base with kick-up and rolled ring footing in clear glass (fig. 8: 8). All these vessels can be generally dated during the 1st century AD, particularly in the first half. The dating was already suggested for the materials from the Révoil collection by E. Marianne Stern,⁶⁶ who also noted that some pieces had a later chronology (2nd century AD).

Finally, the necropolis also yielded some shards of mosaic glass. These include several fragments from an unguentarium in dark blue opaque glass with trailed on marvered yellow threads (fig. 8: 9) and several fragments of millefiori (fig. 8: 10‑11) with good parallels in the Révoil collection⁶⁷ and other Red Sea sites, such as Myos Hormos.⁶⁸ The unguentarium, with marvered and dragged decoration, has to be dated at the beginning of the 1st century AD, as it is a typically Hellenistic decoration that disappears during that century.⁶⁹

Also, very common in Xiis, and always in association with tombs, are the inlays, plaques and tiles in glass, which were embedded in other objects, probably wooden boxes. This kind of inlays represents an important part of the materials gathered by G. Révoil, many of which are identical to those retrieved by us in 2018 and 2019. The most important lot comes from Tomb 153, which includes several fragments decorated with a classical palmette and acanthus chain (fig. 8: 14‑16), two opaque white strips (fig. 8: 20‑21), an inlay of glossy opaque red glass (fig. 8: 19) and two polychrome inlays – red and black (fig. 8: 17‑18). From this same grave come two fragments of glass tile, one of them decorated with serpentine patterns in opaque glass (fig. 8: 13) and another one with wheel rosettes (fig. 8: 12). The best parallels for these pieces outside Somaliland come from the royal tombs of Meroe, where identical inlays have been found in contexts dated to the mid-1st century AD, a chronology that is consistent with the rest of the glass assemblage.⁷⁰ The same chronology can be assigned to the fragment of pillar moulded bowl from the same tomb. Both inlays and tiles recovered during our fieldwork have direct parallels in the Révoil collection,⁷¹ where the over-representation of glass confirms that the materials came from tombs.

A necklace containing several faience beads, probably from Egypt, was found in Tomb 120. Elsewhere in North-East Africa, this kind of beads has been found in funerary contexts, as in the Nubian necropolis of Sedeinga.⁷² There are different types: the most common are melon-shaped, like those from Tomb 120, disc-shaped and cylindrical like those found in the 2018 survey. Less common are glass beads in different colours and shapes, such as the yellow and green melon-shaped beads from Cairn 58 and the blue one from the necklace of Tomb 120 (fig. 9). Some of these beads are akin to so-called Indo-Pacific models.

Non-glass beads have also been found at Xiis, including the bronze beads from the necklace of Tomb 120, which appeared in combination with faience and stone beads. Carnelian and agate beads are among the most typical and widespread elements in Indo-Roman trade. Although they are traditionally believed to come from India, Egyptian products have also been detected.⁷³ Whatever their origin, carnelian beads are present in virtually all sites along the Red Sea. In Xiis we have some barrel and globular-shaped examples and one with truncated-cone shape found in Tomb 14, along with two cylindrical agate beads. Beads made of shells have also been attested, although they are not as common in Xiis as in other necropolises of North-East Africa.⁷⁴ Mollusk shell beads associated with tombs have been found in both the 2018 and 2019 surveys. None exists in the Révoil collection, probably because they were not deemed worth keeping.

Materials from fieldwork in Xiis between 2017 and 2019 provide a chronology situated between the early 1st century and the 3rd century AD. Campanian Dressel 2‑4 amphorae with black sand fabric should be dated mainly during the 1st century AD.⁷⁵ The Egyptian AE 3‑1.5 is dated to the 2nd century AD and could attain the 3rd.⁷⁶ The remaining Italian Dressel 2‑4, the Cilician and Rhodian amphorae and the Gaulish 4 can be dated between the end of the 1st century BC and the 3rd century AD. Similar assemblages, with a comparable chronology, have been found in the Red Sea region, including Myos Hormos, Berenike, Kanē, Sumhuram and Ras Hafun. The Italian terra sigillata has a 1st century AD chronology. The three Cons. 34.1 cups retrieved by G. Révoil are dated between the reigns of Tiberius and Claudius,⁷⁷ that is, during the first half of the 1st century AD. As for the Egyptian plain and red-slipped wares, they have been found in contexts of the 1st‑2nd centuries AD, such as Hafun, Coptos and Myos Hormos.⁷⁸ Other materials providing dates are the organic storage jars, whose chronology has been established between the 2nd century BC and the 4th‑5th centuries AD.⁷⁹ Nevertheless, the excavations of Kanē and Sumhuram have offered a more precise chronology, with the light red fabric dated to the second half of the 1st century BC to late 1st century AD and the yellow-green fabric during the 2nd‑5th centuries AD, although both can appear in small proportions in other periods.⁸⁰ Both types are present in Xiis, but the former is much more common, again corroborating the 1st century AD as the apex in long-distance trade at the site.

Finally, both the glass from the Révoil collection and from our own fieldwork can be dated between the 1st and 3rd centuries AD, with the earlier materials being far more common. Thus, pillar moulded bowls are very common from the 1st century BC to the late 1st century AD and a similar chronology is furnished by the abundant glass inlays. The cylindrical bottle from Tomb 120, instead, is best dated around the 3rd century AD, which evinces the continuation of trade in Xiis at a later date. Instead, there is a total absence of materials dated from the 4th century onwards, meaning that, unless other evidence is retrieved in new excavations of intact tombs, long-distance trade in Xiis came to an end at some point around the 3rd century AD. This is similar to the situation in Myos Hormos, but different from most other Red Sea ports, such as Berenike, Kanē or Adulis,⁸¹ which continue active until the 7th century. The lack of infrastructure or permanent settlement associated with the ports of the Somali coast may explain that they shifted easily, a situation that we have documented for the Middle Ages. The absence of Axumite materials may be also related to the end of Xiis around the 3rd century AD, as the involvement of Axum in the Red Sea increased from the 3rd century onwards.

The fact that the assemblage of Xiis is similar to other Red Sea sites that were active during the first phase of the Indo-Roman trade⁸² is proof of the strong integration of Somaliland in the system and the participation of its communities in transregional consumption trends. This includes the use of wine, which plays an important role in long-distance trade not only in the Mediterranean but also in South Arabia and Mesopotamia, as evinced by the abundance of amphorae. Containers produced in South Arabia (such as the organic storage jars) and Mesopotamia (like one of our glazed jars) may have also been used to transport wine; Muza, according to the Periplus, produced “wheat in moderate quantity and wine in greater”⁸³ and some of the storage jars have traces of bitumen. At the same time, there are some dissonances with other Red Sea sites. Although contacts with India are attested through carnelian and agate beads, whose most likely provenance is South Asia, we have been unable to find Indian pottery so far, which is common in other Red Sea sites, including Hafun in Somalia and Kanē and Sumhuram in South Arabia. This is perhaps related to the fact that most of our finds come from a necropolis.

Our fieldwork has also enabled a better understanding of the spatial organization of the site of Xiis. Glass, glazed wares and fine pottery are in most cases associated with tombs, which means that they were part of the funerary offerings; this was certainly the case in two of the graves that we excavated, with glass in situ (Tomb 153 and Tomb 120). In Ma’ajilayn hill the finds are less diverse; there are many South Arabian organic storage jars and cooking pottery, but few glazed wares and no other fine wares or glass was recorded. The hilltop was probably a temporary settlement area, perhaps the place where the indigenous people awaited the arrival of the foreign merchants. The extraordinary visibility from the hill would support this interpretation. The nomads would live there during the period in which they resided near the coast. As for the island, there is some handmade pottery and plain wheel-made wares, but few amphorae or fine wares, which means that the place was not used for unloading goods. It might have been a temporary settlement area, perhaps for the merchants. Many of the finds that are not directly associated with tombs come from a specific area at the feet of Ma’ajilayn hill. Here we found glass, glazed wares, terra sigillata, Egyptian wares and amphorae. Possible interpretations are a temporary settlement, a meeting place between merchants and locals, or a space for rituals conducted in relation to the tombs. Banquets, pilgrimages and other rituals related to funerary spaces are historically known among the pastoralists of the Horn.⁸⁴ The interpretation of all three non-funerary sites as temporary is based on the lack of permanent structures, the scarcity of artefacts associated with daily life (such as grinding stones and domestic pottery) and the absence of drinkable water. If there was a permanent presence of merchants in Xiis we have failed to find convincing evidence of their existence thus far.

Our knowledge of the people buried in Xiis is still limited. We can assert with certainty that they were nomads, as they were buried in the same kind of graves that appear along the routes leading to the interior of Somaliland and that are still used today by transhumant pastoralists. The absence of local handmade pottery is also indicative of their nomadism, as they were probably using baskets and gourds – as most mobile pastoralists until recently. In addition, no permanent settlements from this period have been found either in the coast or the interior of Somaliland. The elusive nature of nomadic populations in Antiquity makes them a difficult, though fascinating, topic of study.⁸⁵