Medieval Swahili Forts and City Walls
------------------------------------------------
The Portuguese describe Mogadishu with a wall around.
(On the first return voyage of Vasco Da Gama 1499) [Magadoxo]
We found ourselves off a large town, with houses of several stories, big palaces in its center, and four towers around it. (Álvaro Velho 1499).
(In 1507): From this place of Brava went Tristao da Cunha to the city of Magadaxo inhabited by Moors, ………… and he saw a lot of people by the walls and around them, ……… (Chronica d'el-rei D. Manuel by Góis, Damião de, 1502-1574).
The Portuguese description in 1502 of Kilwa (Quiloa) was: (see my webpage on End of the Middle Ages view of Kilwa by the Portuguese).
The city is large and is of good buildings of stone and mortar with terraces, and the houses have much wood works. The city comes down to the shore, and is entirely surrounded by a wall and towers, within which there may be twelve thousand inhabitants. (Martin Fernandez de Figueroa 1505-1511)
PATE.
After passing Melinde, and going towards India, they cross the Gulf (because the coast trends inwards) towards the Red Sea, and on the coast there is a town called Pate, and further on there is another town of the Moors, called Lamon; all these trade with the Gentiles of the country, and they are strongly-walled towns of stone and whitewash, because at times they have to fight with the Gentiles, who live in the interior of the country. (Duarte Barbosa 1514)
Note: Archaeology did not find a wall in Lamu.
BRAVA.
Leaving these places (Pate and Lamu), further on along the coast is a town of the Moors, well walled, and built of good houses of stone and whitewash, which is called Brava. It has not got a king; it is governed by its elders,' they being honoured and respectable persons. It is a place of trade, which has already been destroyed by the Portuguese, with great slaughter of the inhabitants, of whom many were made captives, and great riches in gold, silver, and other merchandise were taken here, and those who escaped fled into the country, and after the place was destroyed they returned to people it. (Duarte Barbosa 1514)
[1506] (This is the eyewitness account of Hans Mayer)
(Mombasa) This is a very big city that lies on an island a little more than a league and a half to two leagues round, The city is on the highest point, and all of it is built on rocks and the side facing the sea is not walled, being very high. And part of the island has a wall that is about the height of a fortress.
The houses are after the fashion of Kilwa and the greater part are of three stories and all covered with plaster.
Taken from: Swahili pre-modern warfare and violence in the Indian Ocean by Stephane Pradines 2020.
P13
The first Swahili city walls (‘enceintes’) are mentioned as from the 10th century, during the formative period of Swahili architecture. In 945-46, Buzurg ibn Shahriyar gave accounts of raids from Malagasy pirates (see my webpage Buzurg ibn Shahriyar 955). The Waq Waqs (Malagasy) arrived on the East African Coast in about a thousand small crafts and attacked the coastal cities from Sofala to Pemba, where the City of Qanbalu resisted their attack. According to Buzurg b. Shahriyar, the City of Qanbalu was built in the middle of an estuary and surrounded by a city wall which gave it the appearance of a castle. A fort, K.l.n.ku, is mentioned on a Fatimid map of the Indian Ocean which dates back to the first half of the 11th century (see my webpage Kitab Ghara’ib al-funun 1050). In the 13th Century, Yakut mentions the city of Mknblu (Mkumbuu). This later reference means we can place Qanbalu on the site of Ras Mkumbuu in the Pemba Island.
P16
Sanje ya Kati (small island just S of Kilwa) was a fortified entrepôt which was protected by a quadrangular wall measuring 100m by 340m, which could be precisely identified with regards to the east and north flanks. With a tower at each corner of the city. The north-east tower is the most significant with a wall of nearly 3m still preserved. With regard to their construction techniques, the houses, the mosque and the fortifications form a homogenous architectural complex from the 11th to the 12th century.
P22
Plan of the town wall of Songo Mnara, 15th c. Kilwa bay, Tanzania. Source: Pradines, 2005. The city of Songo Mnana was founded at the end of the 14th c. and this very small agglomeration, just 4.4 hectares, rarely goes beyond the city perimeter imposed by its city wall. The site was abandoned in the second half of the 16th century.
Almost all of the stone houses are concentrated in the city wall (Pradines and Blanchard, 2005, pp. 25-80). The city wall measures only 50cm thick, and this wall continues on the west, north and east sides. Only the south section is more difficult to interpret as it includes houses in its outline. This element is very common in urban fortifications on the Swahili coast as it saves material and time, with the blind façade of the stone houses forming an effective city wall.
Note: the Portuguese also noted the walled city:
…………… and they came to a land called Kilwa, which the viceroy had already conquered, on which island lies Songo (Songo Mnara) which is a walled town two leagues from Kilwa, Songosongo (Island halfway between Kilwa and Mafia), six leagues away lies Manfia, a fertile and cool land, fifteen leagues to the north of the island lies Tomagunda, (Taka Ungu in the Book of The Zanj) (Martin Fernandez de Figueroa 1505-1511).
P21-23
The rectilinear façade is quite rare and the Swahili city walls have a zigzag design with numerous projections. Some saw-tooth outlines are not connected to the houses or to urbanisation. This is the case with the first city wall of Gedi, which dates from the 15th century, where the zigzags replace the towers and cover the adjacent flanks. ………………
This large city wall (of Gedi) is made up of segments of rectilinear walls measuring an average of 100m long. The city wall is not generally disturbed by inclusive elements that are integral to its structure. The first line of fortification has been plotted on a zone that is uninhabited or peripheral to the 14th -century city. This first fortification was built at the beginning of the 15th century (Kirkman, 1973, pp. 3-8). The fortification was built in a single phase. It lies on an old substratum of the end of the 14th century and is characterised by the inhabitants to the north and a necropolis to the south. The visible outline of the large city wall, therefore, was neither altered nor modified.
The interior city wall is from the 16thc.; it is smaller but more complex than the preceding one.
Taken from: Fortifications et urbanisation en Afrique orientale by Stéphane Pradines 2004.
Somalia
P284
Bur Gao; Site 2 has a 15th-16th century wall. Site 1 and 3 have late walls.
Myaandi is 300 m from the beach, its wall is very poorly preserved with a height of only 70 cm in places. It surrounds a small hill with stone buildings. According to surface material, the city dates back to the 14th-15th century.
Kenya
P288
She Jafari is a small town of 4 hectares, including 3 hectares within the walls. One of its two mosques and all the stone houses are enclosed in a wall 40 to 50 cm wide and 2 m high. T. Wilson's surveys indicate that the centre of She Jafari was occupied in the 15th century, but the mosque outside seems to date back to the 14th century. This observation recalls the urban plan that we studied in Gedi in 1999. The new town of the 15th century is built within a wall that excludes the old Friday mosque.
P292
For James Kirkman, Ungwana would be the town of Hoja mentioned by the Portuguese. But for Mark Horton, it would be the city of Shaka which disappears from the Portuguese sources in 1637 while the establishment of Hoja is known until 1678. We think that 50 years of difference do not allow such an affirmation, especially since there is an archaeological site at Ras Shaka. The Portuguese attacked Hoja in 1505 and noticed that the city was fortified only on the land side and open on the sea side. The site of Ungwana seems entirely fortified, but the inhabitants were able to build part of the wall after the attack of 1505.
Tanzania
P306
The small island of Mtambwe Mkuu is completely covered with archaeological material over an area of 215 hectares. The ancient city, occupied from the 9th to the 11th century, is 22 hectares in itself. The enclosure does not completely encircle the site; it is pierced to the south by a gate that gave access to Pemba during low tides. In 1933, Buchanan clearly saw this wall to the south of the island; it has almost disappeared today.
Near Wete, the city of Mkia wa N'gombe, occupied from the 11th to the 14th century, has ruins of stone houses and a surrounding wall.
P307
In the 13th century, Yakut cites Unguja to designate the island of Zanzibar. The site of Unguja Ukuu was found by Mark Horton in the southwest of the island. The city of Unguja Ukuu was protected by a wall. Ahmed Juma's excavations revealed occupation from the end of the 8th to the 11th century.
Kizimkazi is located in the extreme southwest of the island. Dating from the 12th to the 15th century, the city is defended by a wall with a square tower to the south. According to local tradition, the mosque is the work of a foreign architect. The mihrab has an inscription dated 1107; its style is similar to Persian and Indian mihrabs of the same period.
Mozambique
P323
Somana is located 100km north of the island of Mozambique. The site was discovered in 1983 by Ricardo Duarte. This 13th-15th century settlement is the largest site of this period. It is located on a small island, called Gomene, which is protected by a 2.5 m high wall. The bottom of the wall is composed of irregular coral limestone rubble bound with lime mortar that form a sort of foundation sole; the upper part is made of well-cut rectangular blocks covered with plaster. Halfway up, the wall is sometimes pierced with small rectangular holes measuring 10cm by 30cm. The interior of the enclosure included buildings made of perishable materials, a cistern and a central building that is not a mosque, and which seems to have been used as a dwelling and warehouse. The doors of the building are decorated with moldings and medallions in cut stone, the motifs used are reminiscent of those of the palace of Husuni Kubwa and the portals of the mosque of Fakhr al-Din in Mogadishu.
Madagascar
P324
The city of Mahilaka is located at the northern end of the island, in the bay of Ampasindava. Founded in the 10th or 11th century, the establishment succumbed at the end of the 15th century under the blows of the Itongomaro dynasty. The enclosure of Mahilaka protects an area of 60 hectares. A large structure, improperly called fort by Pierre Vérin and Chantal Radimilahy, is built near the main mosque in the center of the city. This fortified residence, 30m long, has its own trapezoidal enclosure of 3.8 ha. The surrounding wall is 1.05m or 60cm thick depending on the location, or one to two cubits. Its maximum height reaches 3.7m high, with a triangular section of 60cm at the top. A wall, which remains 40cm high, divides the enclosure into two parts. During his excavations, Pierre Vérin unearthed numerous forge debris, perhaps it was a metallurgical workshop. This ensemble of domestic and artisanal character was occupied from the 11th to the 15th century. We interpret it as a boma, that is to say an area reserved for the ruler of the city, close to the great mosque. This enclosure surrounds the palace and serves both as the boundary of the royal domain and as an element of protection.
Comoros
Taken from: Acoua-Agnala M’kiri (Mayotte - 976) Archéologie d’une localité médiévale (11e au 15e siècles EC), entre Afrique et Madagascar; Martial Pauly 2013.
In Acoua, a Ki-bushi (Malagasy) village situated on the northwest coast of Mayotte Island (Comoro Archipelago). Between the 12th and the 15th centuries AD, Acoua Agnala Mkiri was a thriving place playing a part in the western Indian Ocean trade. It was a port of call for ships sailing from Madagascar to the Swahili Coast. Its central area including stone-town was 2.5ha and was built in the 15th century but surrounded by a stone-wall from the 12th century. It was not a Swahili settlement as till now the people speak a Malagasy dialect.