Before 1500: Adal, Ajuran and the coast

Ruined towns of the Somali interior: Abasa, Amud, Fardowsa and the stone towns of the trade routes

Dozens of ruined stone towns lie across the Somali interior, most of them in Awdal, Woqooyi Galbeed, Togdheer, Sanaag and the Nugaal valley. They have mosques, houses of stone and mud mortar, wells, walls and cemeteries, and finds of Chinese porcelain, Yemeni pottery, glass and Arab coins. Archaeologists link most of them to the trade routes between the coast and the highlands, and to the Muslim states of Ifat and Adal, with dated material from about the 12th to the 16th century and older finds at some sites. Few have been excavated.

Last checked 2026-11-12 10 sources Reading time: 12 min

In short

Across the Somali interior stand the ruins of stone towns with mosques, houses, wells and cemeteries. Local people point out about 70 of them. They were first written about by Richard Burton in 1854, surveyed by a British officer in the 1930s and visited by a few teams since, but only a handful have been dug (Mire 2015, African Archaeological Review 32). This page gathers what each of those visits recorded.

Most of the dated material is from the 12th to the 16th century, the age of the trade routes from Zeila and Berbera to the highlands and of the sultanates of Ifat and Adal. Some sites hold older finds.

Who recorded them

Richard Burton, 1854. On 9 December 1854, on his way to Harar, Burton visited two ruined towns in the hills west of Hargeisa, Darbiyah Kola ("Kola's fort") and Aububah. At Darbiyah Kola he saw houses, wells "filled with rubbish", a palace "with its walls of stone and clay intersected by layers of woodwork", and a roofless mosque with "twelve square pillars of rude masonry" and an arched mihrab: "for a mile and a half in length all was ruins—ruins—ruins". His guides told him the two towns had fought each other until both were destroyed, about 300 years before. At Aububah he saw the domed tomb of Shaykh Aububah, two mosques, a small minaret and a large graveyard, and heard of a third town called Ahammed. Burton called them "remains of Galla settlements" (Burton, First Footsteps in East Africa, 1856, Project Gutenberg ebook 6886). See First Footsteps in East Africa.

A. T. Curle, 1930–1934. Alexander Curle, a British political officer marking the border with Ethiopia, reported about twenty ruined towns in the journal Antiquity (Curle, "The Ruined Towns of Somaliland", Antiquity 11, 1937, pp. 315–327). As summarised by Fauvelle-Aymar and his colleagues (Espaces musulmans de la Corne de l'Afrique au Moyen Âge, 2011, ch. 2, §§2, 21, translated here):

  • Thirteen of the towns formed a dense group across south-western Somaliland and the Ethiopian Haud, between 1,500 and 2,000 metres above sea level.
  • The four largest were Abasa, Amud, Au Bare and Gogesa. Each had about 200 houses, with walls still standing up to 5 metres high. The mosques were the best preserved buildings.
  • Curle dated the towns to the 15th and 16th centuries. He collected "several thousand items", among them glazed pottery, Chinese celadon of the Song and Ming dynasties (12th–15th centuries), Arab and Egyptian coins, and glass. The collection went to the British Museum.
  • He dug only one test pit, in a refuse heap at Amud, which went down 1.37 metres through ash, bone and rubbish.

Neville Chittick, 1975. The British-Somali expedition led by Chittick visited Abasa and Zeila. At Abasa it found houses "moderately well built of fairly regular but uncut stone set in mud mortar… mostly small, consisting of one or two rooms" (quoted in Fauvelle-Aymar et al. 2011, §22).

The French survey, 2001. Fauvelle-Aymar, Hirsch, Bernard and Champagne spent two week-long missions in Somaliland in February and April 2001, visiting Zeila, Abasa, Amud, Sheikh and Berbera with the support of the Somaliland authorities. They handed what they collected to the Ministry of Education and Culture (Fauvelle-Aymar et al. 2011, §3 and note 4).

Sada Mire and Somaliland's Department of Archaeology, from 2007. Mire's survey maps about 40 ruined towns, pre-Islamic and Islamic, of the "closer to 70" reported by local people. She finds clusters in the Sheekh–Hargeisa–Burco triangle, where "some of the earliest ruined towns in Somaliland are located", and around Boorama in the west (Mire 2015).

The Spanish team, 2015–2020. A Spanish archaeological project led by Alfredo González-Ruibal and Jorge de Torres Rodríguez, working with Somaliland archaeologists, studied eleven sites in central Somaliland between Berbera and Hargeisa (González-Ruibal et al., "Exploring long distance trade in Somaliland (AD 1000–1900)", Azania 52, 2017, pp. 135–172; de Torres Rodríguez, González-Ruibal, Franco Fernández, Martínez Barrio and Gutiérrez de León Juberías, "Towns in a sea of nomads: territory and trade in Central Somaliland during the Medieval period", in Shaping Cultural Landscapes, Sidestone Press, 2022, pp. 255–274). Their results are set out under "Central Somaliland" below.

The Nugaal valley, 2021. Said M-Shidad Hussein reported several newly found ruined towns in the Nugaal valley, three of them relatively large. He notes that they lie "in the same region where the first interior city in Somalia was reported in 1154" and suggests one may be that city (Hussein, "Ruined towns in Nugaal: a forgotten medieval civilisation in interior Somalia", Azania 56.2, 2021, pp. 250–271, abstract).

The sites

  • Abasa (Abbasa), near the village of Bon in Awdal, about 1,536 metres up. A town of several hectares on a plain crossed by a seasonal river. Curle measured its mosque at about 60 by 54 feet, with "twelve unevenly placed pillars of varying forms", some round and some cross-shaped, a brushwood roof and a mihrab of "four recessed rounded arches". Chittick found that the mihrab arch had fallen between 1934 and 1975. The 2001 team found cowrie shells, some pierced, in an open space that may have been a market (Fauvelle-Aymar et al. 2011, §§21–26). Mire calls Abasa "one of the best preserved ruined towns of Somaliland", now "covered with cacti which protects it from looters" (Mire 2015). The 2018 survey of Abasa and two nearby sites is on Abasa (Darbiyah Kola).
  • Amud (Camuud), about 1,504 metres up on the slopes of a limestone plateau in Awdal. Several hundred houses, densely built. Curle described a mosque with a pointed-arch mihrab and "an antechamber containing a storage well 8 feet deep and lined with plaster to hold the water for ablutions". The 2001 team found the mosque by its well, and a cemetery of at least a hundred Muslim graves without inscriptions. The site was then being used as a quarry for public works (Fauvelle-Aymar et al. 2011, §§27–30).
  • Sheikh (Fardowsa), beside the town of Sheekh, about 60 km south of Berbera. The 2001 French team saw about 10 hectares, with a possible town wall, graves, house foundations and a deep well. Local people gave its medieval name as Ferdussa. The team saw pottery, glass bangles, cowries and an Arab coin, and dated several pieces to the 14th and 15th centuries (Fauvelle-Aymar et al. 2011, §§32–34). The Spanish team measures the town at about 35 hectares (see below). Mire reports Sasanian pottery of the 8th century and Chinese ceramics of the 13th to 17th centuries from the surface (Mire 2015).
  • Berbera. A cemetery about 3 km long and 20 to 50 metres wide runs along the shore, with probably several thousand graves, some marked by a single standing stone (Fauvelle-Aymar et al. 2011, §§36–37).
  • In Sanaag and the east, Mire lists Maduuna, God Caanood near Ceel Afweyn, and Shimbiris. She links inland towns in the east to the ports of Heis (Xiis), Maydh and Laasqoray (Mire 2015).
  • In Awdal and Selel, Mire lists Cad Cad, Old Amuud, Old Jidhi, Derbiga and Qoorgaab near Boorama (same source). The towns between Borama and Hargeisa, among them Kabab, Gogeysa and Cad Cad, are on Kabab and the medieval towns of the Gebiley region.
  • Aw-Barkhadle, near Hargeisa, is a ruined town and burial ground that Mire describes as a pre-Islamic and Islamic sacred centre. See Ancient burials, standing stones and towns.

Central Somaliland: Fardowsa, Qalcadda, Biyo Gure and Bagan

The full 2024 report on Fardowsa is summarised on Fardowsa: the medieval trading town at Sheekh.

The Spanish project found six permanent settlements in its area. All are built in the same way, with square or rectangular houses of flat stones bound with mud, and none has a defensive wall or a street plan. Most date from the 15th and 16th centuries; Fardowsa, the largest, probably goes back to the 12th century (de Torres Rodríguez et al. 2022, p. 265):

  • Fardowsa covered "around 35 ha" at the foot of a pass from the Ogo mountains down to the coastal plain, and served as a resupply stop for caravans. In 2020 the team cleared about 600 square metres and found two large houses of identical plan around a courtyard, which they read as the compound of "an extended, wealthy family". Finds include pottery and glass from China, South-East Asia, Persia, Yemen and the Middle East, and camel bones in refuse pits. The team writes that this points to "the existence of privileged groups in the settlements, something not documented before in the medieval archaeology of Somaliland". Fardowsa was abandoned towards the end of the 16th century; there are almost no imports from the 17th century (p. 265).
  • Qalcadda, near the Jerato Pass on the main route from Berbera to Hargeisa, is a caravan station: a square building with rooms around a central courtyard, beside a rectangular fort with round corner towers. Its name comes from the Arabic qal'a, "fortified place". A radiocarbon date from one of its four successive lime floors gives 1486–1646 AD. The team suggests that the cost of building and garrisoning it makes it more likely "a state initiative" of the Adal period than a private venture (pp. 268–269).
  • Biyo Gure, a village of about 30 buildings on the coastal plain, held many imports, among them Chinese porcelain, Martaban and celadon wares and Yemeni pottery of the 15th and 16th centuries. It was destroyed by fire; the team writes that who attacked it, "perhaps nomadic raids, Portuguese soldiers, or locals", is unknown (pp. 265–268).
  • Bagan (6 hectares) and Gidheys (2 hectares) are villages of loosely grouped houses with very few imports (p. 269).

The team describes trade on this coast as tied to the seasons. The north-east monsoon made sailing easy from about October to April, the same months when herders returned to the coast and its wells. Trading meetings were also "places for social and political interactions" for the nomadic communities (pp. 256–257). They see the Portuguese raid on Berbera in 1518, when the fleet found the town "completely empty of people and goods", as one of the first of the events that ended this trade system (p. 255).

Trade and dates

Nicholas Tait and Timothy Insoll write that the Somaliland town sites "are linked to the Ifat and Adal Sultanates and seem to have operated between the late thirteenth and the late sixteenth centuries". Imported goods seen at many of them include Chinese celadon and blue-and-white porcelain, Yemeni "black-on-yellow" pottery, glass beads and glass bangles (Tait and Insoll 2021, p. 438).

Mire writes that the material from the Somali region runs "from Tang Dynasty to Ming Dynasty China" and that "the people of this region were part of the Silk Road trade". She notes that ancient Christian burials are often found inside early Islamic ruined towns, and takes this as a sign that "the areas that blossomed as towns under Islam were most likely also significant pre-Islamic centres" (Mire 2015).

Harlaa: the excavated town at the other end of the route

The best-excavated town on these routes lies just over the border, at Harlaa, about 40 km north-west of Harar in eastern Ethiopia. A team from the University of Exeter and Addis Ababa University dug there over five seasons from 2015 (Insoll et al., "Material cosmopolitanism: the entrepot of Harlaa as an Islamic gateway to eastern Ethiopia", Antiquity 95, 2021, pp. 487–507):

  • The authors report twenty-seven radiocarbon dates, which they read as showing the town was lived in from the mid-6th to the 15th century, with a peak of trade and craft in the 12th and 13th centuries.
  • Its mosque existed by the mid-12th century. Arabic inscriptions found there include one dated 657 AH (1259–1260) and one with a partial date of the 440s AH (1048–1057).
  • Pottery came from Yemen, Iran, India and China, including 160 sherds of Chinese and South-East Asian wares. Shells from the Red Sea were worked into jewellery there (Insoll 2021, Journal of African Archaeology 19).
  • Some pottery rims at Harlaa are decorated like those of the Somaliland towns, and "many of the finer decorated rims at Harlaa… are potentially imports from Somaliland". A burnished sherd from Curle's Somaliland collection in the British Museum is "visually identical" to a Harlaa ware (Tait and Insoll 2021, pp. 438–439).
  • The authors identify Harlaa with Hubat, capital of a small Harla sultanate tied to Ifat. They credit Said M-Shidad Hussein with the suggestion that the Somali word hoobat and the Arabic hubut, both meaning to come down a slope from high ground, describe Harlaa's position between highland and lowland (Insoll et al. 2021, p. 504).

Who built them: who says what

  • Burton's guides, 1854, told him the towns of Darbiyah Kola and Aububah had been "Galla" (Oromo) settlements, and Burton wrote the same (Burton 1856).
  • In eastern Ethiopia, ruined stone towns and tombs are commonly called "Harla" and ascribed by the people living there today to "legendary ancient giants" (Insoll et al. 2021, p. 488).
  • Sada Mire argues that the spread of the towns across the whole country "challenges the notion that Arabs had outposts there", and writes that it suggests the kingdoms of Ifat and Adal "not only were settled people, but also were part of what might be viewed as an early Islamic Empire in the Horn of Africa" (Mire 2015).
  • Tait and Insoll link the Somaliland town sites to the Ifat and Adal sultanates (2021, p. 438).

Threats

The 2001 team found the Abasa mosque much reduced since Curle's photographs, Amud being quarried for building stone, and the Sheikh site dug by treasure hunters and threatened by the growth of the modern town (Fauvelle-Aymar et al. 2011, §§24, 30, 34). Mire photographed looted and destroyed sites across Somaliland, among them a looted cairn in Togdheer, and calls decorated stelae "one of the most threatened artefacts in this region" (Mire 2015).

Sources

  1. Mire 2015, African Archaeological Review 32 · link.springer.com
  2. Burton, First Footsteps in East Africa, 1856, Project Gutenberg ebook 6886 · gutenberg.org
  3. Curle, "The Ruined Towns of Somaliland", Antiquity 11, 1937, pp. 315–327 · doi.org
  4. Espaces musulmans de la Corne de l'Afrique au Moyen Âge, 2011, ch. 2, §§2, 21 · books.openedition.org
  5. González-Ruibal et al., "Exploring long distance trade in Somaliland (AD 1000–1900)", Azania 52, 2017, pp. 135–172 · doi.org
  6. de Torres Rodríguez, González-Ruibal, Franco Fernández, Martínez Barrio and Gutiérrez de León Juberías, "Towns in a sea of nomads: territory and trade in Central Somaliland during the Medieval period", in Shaping Cultural Landscapes, Sidestone Press, 2022, pp. 255–274 · ancientportsantiques.com
  7. Hussein, "Ruined towns in Nugaal: a forgotten medieval civilisation in interior Somalia", Azania 56.2, 2021, pp. 250–271 · doi.org
  8. Tait and Insoll 2021, p. 438 · doi.org
  9. Insoll et al., "Material cosmopolitanism: the entrepot of Harlaa as an Islamic gateway to eastern Ethiopia", Antiquity 95, 2021, pp. 487–507 · doi.org
  10. Insoll 2021, Journal of African Archaeology 19 · doi.org

Toosoo!

Calankii is made by Toosoo, which works for one united, prosperous Somalia.

Can you help with Ruined towns of the Somali interior: Abasa, Amud, Fardowsa and the stone towns of the trade routes? Tell us what you can give.

I want to help with this
What you can give

Only the core team sees your answer. We never share or publish it. Privacy · or email hello@calankii.com.

All the ways to help

Where it happened

Same era

Keep reading