FUNCTIONS AND HISTORICAL SIGNIFICANCE OF SWAHILI WALL NICHES

‫عدد ‪ ، (٣٨( ۱٩‬يوليو ‪۲۰۲٤‬‬ Nouvelles Chroniques du manuscrit au Yémen 19 (Ancienne série 38) Juillet 2024 (prochain numéro janvier 2025) S. Pradines & O. Onézime Swahili wall niches FUNCTIONS AND HISTORICAL SIGNIFICANCE OF SWAHILI WALL NICHES Stéphane Pradines (The Aga Khan University, Institute for the Study of Muslim Civilisations) & Olivier Onézime (Géomètre/topographe/photogrammètre) Abstract The Swahilis people are an African maritime culture based on long-distance trade in the Indian Ocean. One physical characteristic of Swahili stone houses is the presence of wall niches, especially in the palaces from the 18th century. These niches were used to store important objects of the household such as imported ceramics and precious manuscripts, mostly Qurʾans and family archives. Our study starts by our new archaeological investigations on the Ujumbe palace in Mutsamudu on Anjouan Island (Nzwani). The large walls covered by a series of stucco niches, from Mutsamudu and Domoni in the Comoros are very similar to the stucco niches in the Lamu Archipelago in Kenya and Bajuni Archipelago in Somalia. We will trace the history of Swahili wall niches starting from the 18th century houses in Comoros and Lamu and going backward until to the 12th century and the first Swahili stone houses in Kilwa. Then, we will explain the political and social significances of these wall niches in Swahili culture in relationship with imported ceramics and manuscripts. Finally, we will explore the architectural kinship between East Africa and Arabia, especially with Zabīd and Mocha in Yemen, and the links with Indian communities from Sindh and Gujarat, showing some possible origin of this architecture directly from the Gulf and Abbasid Iraq. Résumé Fonctions et importance historique des niches murales swahili Les Swahilis forment une culture maritime africaine fondée sur le commerce à longue distance dans l’océan Indien. Une caractéristique physique des maisons swahilies est la présence de niches murales, notamment dans les palais du xviiie s. Ces niches servaient à garder les objets prestigieux de la maison tels que des céramiques importées et des manuscrits précieux, principalement des corans et des archives familiales. Notre étude débute par de nouvelles investigations archéologiques sur le palais de l’Ujumbe à Mutsamudu sur l’île d’Anjouan (Nzwani). Les grands murs recouverts d’une série de niches en stuc de Mutsamudu et Domoni, aux Comores, ressemblent beaucoup aux niches en stuc de l’archipel de Lamu au Kenya et de l’archipel de Bajuni en Somalie. Nous retracerons l’histoire des niches murales swahilies depuis les maisons du xviiie s. aux Comores et à Lamu, en remontant jusqu’au xiie s. et aux premières maisons en pierre swahilies à Kilwa. Ensuite, nous tenterons de comprendre les significations politiques et sociales de ces niches murales dans la culture swahilie en relation avec les céramiques et manuscrits importés. Enfin, nous explorerons la parenté architecturale des niches entre l’Afrique de l’Est et l’Arabie, notamment avec Zabīd et Mocha au Yémen, les liens avec les communautés indiennes du Sindh et du Gujerat, montrant une origine possible de cette architecture directement dans le Golfe et l’Irak abbaside. ‫اخلالصة‬ ‫الكوات) يف املنازل السواحلية‬ َ ( ‫الوظائف والمهية التارخيية للنوافذ اجلدارية‬ ‫ اإحدى اخلصائص الفيايئية‬.‫الشعب السواحييل ثقافة حبرية أفريقية تقوم عىل التجارة ملسافات طويةل يف احمليط الهندي‬ ‫ مت اس تخدام‬.‫ خاصة يف القصور اليت تعود اإىل القرن الثامن عرش‬،‫للمنازل احلجرية السواحلية يه وجود منافذ جدارية‬ nCmY 19 (Juillet 2024) 137 S. Pradines & O. Onézime Swahili wall niches ‫ ومعظمها من القرأن الكرمي‬،‫هذه املنافذ لتخزين الغراض املزنلية املهمة مثل اخلزف املس تورد واخملطوطات المثينة‬ .)‫ تبدأ دراستنا بتحقيقاتنا الثرية اجلديدة يف قرص أوجوميب يف موتسامودو يف جزيرة أجنوان (نزواين‬.‫وأرش يفات الرسة‬ ‫ تش به اإىل حد كبري‬،‫ من موتسامودو ودوموين يف جزر القمر‬،‫اجلدران الكبرية املغطاة بسلسةل من الكوات اجلصية‬ ‫ سوف نتتبع اترخي منافذ اجلدار السواحيلية‬.‫الكوات اجلصية يف أرخبيل لمو يف كينيا وأرخبيل ابجوين يف الصومال‬ ‫بدءا من منازل القرن الثامن عرش يف جزر القمر ولمو والعودة اإىل القرن الثاين عرش وأول املنازل احلجرية السواحيلية‬ ً ‫ سنرشح المهية الس ياس ية والاجامتعية لهذه املنافذ اجلدارية يف الثقافة السواحلية وعالقهتا ابخلزف‬،‫ بعد ذكل‬.‫يف كيلوا‬ ‫ وخاصة مع زبيد‬،‫ سوف نس تكشف العالقة املعامرية بي رشق أفريقيا واجلزيرة العربية‬،‫ وأخريا‬.‫واخملطوطات املس توردة‬ ‫ مما يدل عىل بعض الصل احملمتل لهذه الهندسة‬،‫ والروابط مع اجملمتعات الهندية من الس ند وجوجارات‬،‫واخملا يف المين‬ .‫املعامرية مبارشة من اخلليج والعراق العبايس‬ Keywords Swahili — niche — architecture — ceramic— Indian Ocean — Comoros — Yemen — Oman — Sindh — Gujarat — Irak — Iran Mots-clés swahili — niche — architecture — céramique — océan Indien — Comores — Yémen — Oman — Sindh — Gujerat — Iraq — Iran ‫لكامت رائس ية‬ ‫ اخلزف — املطيط الهندي — جزر القمر — المين — عامن‬،‫سواحيل — املشاكة — عمل العامرة — السرياميك‬ ‫— الس ند — جوجارات — العراق — ايران‬ I. Introduction Mainly located in Kenya, Tanzania, but also in Somalia, Comoros and Mozambique, the Swahilis form an extremely original African maritime culture based on longdistance trade in the Indian Ocean [Fig. 1]. Wa-Swahili, or “those of the shore”, refers to the inhabitants of the coast of East Africa. Arabo-Persian geographers called Zanj, the inhabitants of the coasts of East Africa1. In fact, the word Swahili covers a complex reality: it does not designate a population, but a culture, made up of several Swahilispeaking groups. This common cultural background encompasses populations of different origins, some Nilotic, other Bantu. On the other hand, these coastal populations share the same social organization, the same architecture, the same language, Kiswahili, and the same religion, Islam. This culture, on the periphery of Asian and African worlds, occupied a position extremely favorable to the development of an original coastal culture based on trade relations and the spread of Islam. 1 L.-M. Devic, Le Pays des Zendjs, 1975 [1883], p. 15; C.H. Stigand, The Land of Zinj, 1913 [reed. 1966]. nCmY 19 (Juillet 2024) 138 S. Pradines & O. Onézime Swahili wall niches Pre-Islamic settlements identified on the East African coast demonstrate a continuity of human occupation since at least the Stone Age until our days. These ancient settlements were active players of the international trade in the Indian Ocean during the Roman and Sassanian periods. African ports or emporiums were mentioned during the Antiquity2. No kingdom or empire federated the cities of the East African coast. Each port was independent and formed a city-state which controlled a territory. This system of thalassocracies has been described in the Periplus of the Eritrean Sea who mentions the political autonomy of the cities of the African coast from the beginning of the first millennium3. The contact with Abbasid sailors and the islamization of the coast represents a decisive factor in the evolution of Swahili town planning and architecture. More settlements were founded on the littoral of East Africa, such as the city of Qanbalu4. It was during this period that coral was introduced as a building material. The technology of sea coral in Swahili architecture appeared in East Africa between the ninth and the 10th century in Lamu and Comoros archipelagos5. The blocks of coral porites were collected at a low tide. During this period, only little uncut blocks of coral were used, bounded by a red clay mortar. Between the eleventh and the 12th century the techniques of construction changed radically and became extremely elaborate. The principal characteristic of this period was the exclusive use of sea coral well cut into little quadrangular blocks of horizontal courses with lime mortar to build the walls and niches such as the ones of the mosques of Kizimkazi on Zanzibar Island and the houses in Sanje ya Kati. Blocks of sea coral were carved in little bricks to produce mouldings and arches of the early Swahili niches. The dry laid assembly (dry stone) was used to build the mouldings of the frame of the niches. There was no need of mortar or just very little6. The 13th century was a period of transition in which Swahili masons abandoned the use of blocks of sea coral to build their walls and preferred ashlar blocks of fossil coral limestone. Sea coral, however, continued to be used until the 17th century, but would only be used for fine sculpture, particularly for the mihrabs, door jambs and window pillars, rope mouldings, friezes, inscriptions, and wall niches. Between the 13th and the 15th century, the development of Swahili city-states, such as Kilwa, Mombasa, Malindi, Pate and Mogadishu, has always been driven by commercial competition and political rivalry for control of a region7. The economy was at the centre of Swahili society, dependent on long-distance trade. The coastal cities formed an ethnic and cultural interface between the Africans of the highlands and A. Juma, “The Swahili and the Mediterranean worlds: pottery of the late Roman period from Zanzibar”, 1996. 3 G.S.P. Freeman-Grenville, The East African Coast: Select documents from the first to the earlier nineteenth century, 1962a, pp. 1–7. 4 S. Pradines, Fortifications et urbanisation en Afrique orientale, 2004, p. 27. 5 M. Horton, Shanga, 1996; S. Pradines, “Islamic Archaeology in the Comoros”, 2019. 6 P. Garlake, The Early Islamic Architecture of the East African Coast, 1966, pp. 42–48; S. Pradines, Historic mosques in sub-Saharan Africa, 2022, pp. 174–177, 212–213. 7 J. De Barros, Decadas Da Ásia, 1552 [1945], p. 21. 2 nCmY 19 (Juillet 2024) 139 S. Pradines & O. Onézime Swahili wall niches the merchants of the Indian Ocean. The Swahili cities were above all positioned at the mouth of rivers or islands forming obligatory passage points for ships8. The Swahilis first exported raw materials: ivory, turtle scales, gold, rock crystal9. In return, Swahilis imported manufactured goods produced in large quantities on eastern markets: fabrics, embroideries, silks, Indian beads, Persian ceramics, Chinese porcelains, and glassware; these objects were used as currency. In addition to income from trade and fishing, the Swahili cities had plantations enabling them to meet their needs such as coconut and rice10. One physical characteristic of Swahili stone houses is the presence of wall niches. These niches were used to store important objects of the household such as imported ceramics and precious manuscripts, mostly Qurʾans and family archives. Our study starts by new archaeological investigations on the Ujumbe palace in Mutsamudu in Anjouan (Nzwani). The large walls covered by a series of stucco niches, from Mutsamudu and Domoni in the Comoros are very similar to the stucco niches in the Lamu Archipelago in Kenya and Bajuni Archipelago in Somalia. In a second part, we will trace the history of Swahili wall niches starting from the 18th century and going backward until to the 12th century and the first Swahili stone houses. In the third part of this research, we will explain the political and social significances of these wall niches in Swahili culture in relationship with imported ceramics and manuscripts. In the fourth part, we will explore the architectural kinship between East Africa and Arabia, especially with Zabīd and Mocha in Yemen, the links with Indian communities from Sindh and Gujarat, showing some possible origin of this architecture directly from the Gulf, especially Abbasid Iraq and Buyid Iran. Our paper is first based on archaeology and architecture, but also on Arabic primary sources, descriptions of European travellers, and studies from anthropologists, to understand the role and function of mural niches in Swahili society. II. Eighteen century Swahili niches from the Comoros and Lamu archipelagos In 2022, in collaboration with the CNDRS and the Charity “Patrimoine des Comores”, we started archaeological investigations in Mutsamudu, the capital city of Anjouan Island (Nzwani) in Comoros. According to oral tradition, the city of Mutsamudu was founded in the late 15th century11. The Sultan Mohamed bin Sultan Hassan decided to 8 S. Pradines, Fortifications et urbanisation en Afrique orientale, 2004, pp. 29, 87; A.D. Jama, The Origins and Development of Mogadishu AD 1000 to 1850, 1996, p. 28. 9 L.-M. Devic, Le Pays des Zendjs, 1975 [1883], pp. 160–166, 179–188; G.S.P. Freeman-Grenville, The East African Coast: Select documents from the first to the earlier nineteenth century, 1962a, p. 152. 10 N. Chittick, “Indian relations with East Africa before the arrival of the Portuguese”, 1980, p. 19. 11 All our dates are given in common era or Gregorian calendar. nCmY 19 (Juillet 2024) 140 S. Pradines & O. Onézime Swahili wall niches build a new city called Mutsamudu in 148212. Later in 1492, a palace was built by people from Pate as residence for the Wali (Governor), the Sultan, and his family when they were visiting the town [Fig. 2]. The tradition reports that the builders were Nabahani people from Pate from the al-Nabhan tribe of Omani descent13. The work started in 1492 and ended in 1530. Unfortunately, in the current state of the research and our recent excavations in Mutsamudu, no archaeological evidence supports these early dates—no ceramics or visible historical monuments predate the 17th century. In the 18th century, Mutsamudu was an important stopover for English, French and Portuguese ships travelling to India especially from 1727 to 1788. However, by the late 18th century, the insecurity and violence were continuous on the Eastern coast of Anjouan Island with raids from Malagasy pirates14. The capital city of Domoni was attacked several times and sacked in 1790. For all these reasons the Sultan decided to transfer his capital from Domoni to Mutsamudu in 1792. The palace or “Ujumbe” of Mutsamudu was the sultan of Anjouan’s residence from 1803 to 1909. “Ujumbe” is the Comorian word for palace. In fact, it is interesting to note that the so-called palace is called a palace because it was the residence of the Sultan but many stone houses with similar plans and decoration still exist in Mutsamudu, another Ujumbe exist too in Domoni the previous and former capital of the Island. Three other large stone houses or palaces have been recorded in Domoni: Pangani Darini and Toiyfa15. In Mutsamudu, twelve palaces have been identified plus a few others not listed16, in total possibly around 20 major stone houses, including the following: Singani, Mpangahari, Said Ahmed Zaki, Barakani, Stambul, Bertal, Dumani, Rochani, Swaaniani, Kastwantwani, Wemani and Oubweni. All the palaces or stone houses of Anjouan date from the 18th century and are contemporary with those of Lamu and Pate. In the early 18th century, Comorian stone houses (so-called “palaces”) had a front courtyard with trees and a shaded veranda with pillars. This hypostyle courtyard was the social space of the house, for family, guests, and daily-life activities. During the 18th century this open-air courtyard and veranda disappeared with the construction of a first floor for extended family. The open-air courtyard was closed by a wooden paint- H. Ben Said Mohammed, Les Sharifs dans l’histoire des Comores, 2015; A. Ben Said Mohammed, “Exposé sur l’histoire de la vile de Mutsamudu”, 2022 and A. Bourhane, “Mutsamudu et Anjouan aux temps des sultans”, 2006. 13 S. Pradines, Fortifications et urbanisation en Afrique orientale, 2004, p. 318. 14 J.-C. Hebert, “Documents sur les razzias malgaches aux Îles Comores et sur la côte orientale africaine (1790-1820)”, 1983; J. Martin, Comores : quatre îles entre pirates et planteurs, 1983. 15 S. Hirschi & C. Nafa, Sultanats historiques des Comores. Recueil de relevés du patrimoine architectural et urbain, 2014, pp. 107–120. 16 S. Hirschi & C. Nafa, Sultanats historiques des Comores. Recueil de relevés du patrimoine architectural et urbain, 2014, pp. 85–99. 12 nCmY 19 (Juillet 2024) 141 S. Pradines & O. Onézime Swahili wall niches ed ceiling and large skylights17. The courtyard was once in front, ended up in the center of the house, and it was renamed “Shandza Hari” the central room of the house. A new masculine space appears on the facade overlooking the street, the “Ukumbi”. This space was originally a simple vestibule, then became a small reception room open to the street. The second and the third spaces at the back of the house are called the “Shandza” and the “Ndani Bweni”, a bedroom for the wife. Sometimes attached to it, a fourth room the “Poro” is a storage space or a spare bedroom for guests. Decorative wall niches called “Ziloho” are mainly located in the Shandza Hari and the Ndani [Fig. 3]. The niches and imported porcelains, the painted Qurʾanic inscriptions on the ceilings of the Ukumbi and the Shandza Hari acted as talismanic protection for the occupants of the house [Fig. 4]. The painted Quʾranic blessings and the “wafaku”, magic squares and talismans, with similar inscriptions are visible in the Indian Ocean, notably in the Maldives, India, Yemen but also at Harar in Ethiopia and Siyu in the Lamu Archipelago18. Stucco wall niches are also present in the ablution area [Fig. 5]. Ceramics displayed in niches acted as charms against witchcraft to protect women against sterility and diseases19. It is interesting to note that the Ndani was also protected by niches but not by the painted Qurʾanic inscriptions. It seems that this Qurʾanic or Islamic ritual protection was reserved for masculine spaces. The Comorian house has a gendered division of spaces with two reception areas and intimate spaces, one for men and guests and the other for women. The Ndani was a feminine place for purification and rituals of passage, from births and weddings (Harusi) to funerals. The Harusi is an important ritual, during the wedding when the groom steps into the Ndani and lifts the veil to see his wife’s face20. Our research in Mutsamudu focused on the Ujumbe palace with a study of the architecture to identify the main construction phases of the building. The core of the Ujumbe is visible to the northeast [Fig. 2]. This phase probably dates from the late 17th century. The house was then enlarged at the beginning of the 18th century; then the first floor was added in the 18th century. The building was completed in 1786. It was modified several times from the late 17th to the late 20th century. The two floors correspond to two family units with distinctive entrances, reception halls, rooms, bathrooms, and toilets. The Ujumbe is a classic and standard Swahili stone house like those in Lamu and Pate in Kenya, with a series of three narrow, parallel rooms. This very old plan dates to the 12th century as visible in Sanje ya Kati, Kilwa Archipelago. These parallel rooms are also observed in Shanga, Lamu Archipelago in the 14th century and in Gedi, Kenya in the 15th century. S. Blanchy, “La maison urbaine, cadre de production du statut et du genre à Anjouan (Comores), XVIIe-XIXe siècles”, 2022, pp. 7–10. 18 A. Rouaud, “L’émigration yéménite”, 1984; and J. Campo, forthcoming. 19 S. Blanchy, “La maison urbaine, cadre de production du statut et du genre à Anjouan (Comores), XVIIe-XIXe siècles”, 2022, pp. 16–25. 20 S. Blanchy, “La maison urbaine, cadre de production du statut et du genre à Anjouan (Comores), XVIIe-XIXe siècles”, 2022, pp. 12–14, 31–38. 17 nCmY 19 (Juillet 2024) 142 S. Pradines & O. Onézime Swahili wall niches The Ujumbe of Mutsamudu has many stuccos and moulded decorations, mainly niches, located in the reception rooms, the main bedrooms, and the bathrooms [Fig. 6]. The most impressive is certainly the large panel of niches of the upper floor. The first floor forms an independent house or apartment with access by a monumental staircase on the street side. On the first floor, room 16 is decorated with a panel of niches on the north wall [Fig. 7]. The panel is framed by a moulded band before the ceiling cornice and is made up of 15 squares framed by mouldings with medallions decorated with star-shaped interlacing in the centre. Two medallions are missing above the door, so in all there were probably 17 squares. The drawings in the medallions are not very visible because of the lime renders which have covered and abraded the moulded decorations. Two large squares with medallions frame the entire frieze. Below this frieze are 6 series of bands of niches with different types of arcades. There are first the flat rectangular niches with arcature composed of double triangles, then a series of semicircular arch niches, then a series of niches with polylobed arches, and finally a series of large rectangular niches surmounted by three small semicircular arches. These niches are framed by smaller niches surmounted by stepped or corbelled arches. Next comes a series of flat rectangular niches with an arcade made up of double triangles, then a series of simple rectangular flat-bottomed blind niches. The sides of the mouldings of the door are framed by niches which descend to ground level. The symmetrical composition of the niche panel is focused on the door which is the central element. The door is surmounted by a painted lintel with an Arabic inscription. A series of 6 niches is fitted in each door door jamb. There is greater variation in the types of arches of the ground floor niches in room 4 or in the courtyard niches (Space 8) [Fig. 8]. These niches are certainly older and made by a master mason who had a greater artistic repertoire than the mason who decorated the upper floor. The panel of niches in room 4 has 7 friezes plus a missing band at the level of the plinth, therefore probably 8 friezes of niches originally. The wall panel is composed of a series of flat rectangular niches with single or double triangular arches, a series of pointed arch niches with three mouldings, a series of niches with two semicircular arches or a single arch, the junction between the two arches forms a pendant decorated with a moulded cross. Then comes a series of niches surmounted by three mini round arches and a series of moulded or corbelled stepped niches. The stucco niches of the Ujumbe of Mutsamudu are exceptional and characteristic of the Swahili culture [Fig. 9]. The stucco niches of Ujumbe are made on masonry base of lime mortar, sand, sea coral stone and sometimes basalt. Plaster had been used on the East African coasts since at least the 10th century, but the 18th century marks the apogee of stucco and plaster in Swahili architecture21. The friezes and the decorations carved in the stucco were accomplished performances from the start and were derived essentially from the wood carving applied to the doors of Swahili houses. The chain motifs on the mouldings recall elements of navigation just as the corded mould- 21 S. Pradines, Historic mosques in sub-Saharan Africa, 2022, pp. 215–217. nCmY 19 (Juillet 2024) 143 S. Pradines & O. Onézime Swahili wall niches ings cut in coral stone recall the mooring cables of boats, like the Gujarati marble carvings. The slight variations in the representations of floral motifs in the same building proves that the decorations were not moulded serially. The decorative elements were engraved in the stucco.22 First, the craftsmen designed a modular grid divided into squares and triangles. Using these basic elements, they could create geometrical patterns (stars, friezes with curvilinear motifs, zigzags, waves, etc.), or a stylised vegetation (lotuses, flowers, arabesques, palmettes, etc.). The modular grid acted as a rigorous framework for the artist’s imagination and allowed an almost faithful reproduction of each engraving. All the stucco decorations were then covered with a fine and brilliant cream of white lime. The niches in the Ujumbe of Mutsamudu have different size, shapes and types of arches with moulded and chiselled decorations [Fig. 10]. Some are real utilitarian niches with a depth that can accommodate objects: ceramics, lamps, small weapons such as daggers... The lowest ones rectangular and flat could accommodate Qurʾans, books, and archives23. Some blind niches are purely decorative [Fig. 11]. The niches cover the whole of certain walls, this treatment being reserved for the most prestigious rooms of the house: the reception room, bedroom of the master of the house. Niches are also present in the bathrooms and toilets. It seems that many niches in the palace were backfilled at the end of the 19th century or during the 20th century. The new occupants no longer have interest in or use for these decorations; they wanted to fill these niches to have flat walls without openings. The courtyard niches correspond to another house destroyed during the construction of the concrete-built storerooms and the courtyard [Fig. 12]. This house was probably in ruins when these stores were built. Similar palaces and wall niches are still preserved in the Lamu Archipelago, Kenya, and the Bajuni Archipelago in Somalia [Fig. 13], with prestigious examples in Lamu and Pate towns24. The presence of these niches in Northern Kenya, and Comoros, shows very strong religious, economic, cultural ties and kinship between these archipelagos, themselves strongly connected to the South of the Arabian Peninsula, Yemen, and Oman. The stone houses in Pate and Lamu towns were extensively studied and published by architects and anthropologists, especially Usam Ghaidan, James de Vere Allen and Linda W. Donley-Reid. These 18th century stone houses are still in use today and most of the vocabulary that researchers use to describe Swahili domestic architecture is based on the Kiswahili spoken in Lamu Archipelago; some variations occur between the Bajuni Archipelago in Somalia and Comoros25. In front of the house in Lamu and Pate, a narrow porch, called the “Daka”, is framed by long stone benches called “Baraza” for welcoming guests. In Yemen, the U. Ghaidan, “Swahili plasterwork”, 1973. E. Ho, The graves of Tarim, 2006. 24 J. de V. Allen, “The Swahili house: cultural and ritual concepts underlying its plan and structure”, 1979, p. 1. 25 U. Ghaidan, Lamu. A study in conservation, 1976, pp. 14–19, 51–60. 22 23 nCmY 19 (Juillet 2024) 144 S. Pradines & O. Onézime Swahili wall niches “Deka” is also a stone bench outside the house and near the main entrance26. From a bent entrance hall called “Tekani”27, the visitor arrives in an open rectangular or square central courtyard, the “Kiwanda”28. The “kiwanda” is sometimes covered by the late addition when an upper storey was added. The “Sabule” is a guestroom with bathroom ensuite29. This guest room for visiting merchants is an extension of the exterior world in the private space. The “Sabule” is generally attached to the central courtyard, Kiwanda. The other side of the courtyard, in front of the main entrance, leads on series of three parallel rectangular rooms exactly like in Comoros. The core of the Lamu house is made up of these three transverse rectangular rooms placed in a row and connected by a central passage (“msana wa-tini, msana wa-yuu, ndani”). The first rectangular space is the Lower chamber, “Msana wa Tini”, an open room or veranda overlooking the courtyard30. The second room is the Upper chamber, “Msana wa Yuu”, a more private reception room for special guests, and used as sleeping area31. The third and most private room is the “Ndani”, considered as parents’ bedroom with an ensuite bathroom and toilet32. By intimacy gradients, the “Ndani” is the most sacred place of the house, the harem. Located in the heart of the house, the “Ndani” is also a ceremonial room. Presented as a room for women seclusion, it is more a space for women activities. The room is used for birth and “Harusi kuu” the great weddings33. Sometimes a storeroom, “Nyumba ya-Kati”34 is attached to the “Ndani”. The “Nyumba ya Kati” or “back room” is also used for funerals, laying out and washing the dead bodies. In Lamu Archipelago, the walls of the second and third rooms, “Msana wa Yuu and Ndani”, are covered by a rich stucco ornamentation with mouldings and niches35. The stucco motives are very diverse chains, zigzag, fluted patterns, crosses. The harem wall niches are called “Kidaka, Zidaka, or Vidaka”36. The stucco works in the Ndani are C. Hardy-Guilbert & G. Ducatez, “Al-Shihr, Yémen, porte du Hadramawt sur l’océan Indien”, 2004, p. 129. 27 J. de V. Allen, “The Swahili house: cultural and ritual concepts underlying its plan and structure”, 1979, p. 7. 28 U. Ghaidan, Lamu. A study of the swahili town, 1975, pp. 43–50. 29 J. de V. Allen, “The Swahili house: cultural and ritual concepts underlying its plan and structure”, 1979, p. 22. 30 U. Ghaidan Lamu. A study of the swahili town, 1975, pp. 43–50. 31 J. de V. Allen, “The Swahili house: cultural and ritual concepts underlying its plan and structure”, 1979, p. 9; U. Ghaidan, Lamu. A study of the swahili town, 1975, pp. 43–50. 32 L.W. Donley-Reid, “A structuring structure: the Swahili house”, 1990b. 33 J. de V. Allen, “The Swahili house: cultural and ritual concepts underlying its plan and structure”, 1979, pp. 16–21. 34 M. Horton & J. Middleton, The Swahili, 2000, p. 117. 35 J. de V. Allen, “Swahili culture reconsidered: some historical implications of the material culture of the Northern Kenya Coast in the Eighteenth and Nineteenth centuries”, 1974; U. Ghaidan, “Swahili plasterwork”, 1973. 36 J. de V. Allen, “The Swahili house: cultural and ritual concepts underlying its plan and structure”, 1979, p. 11. 26 nCmY 19 (Juillet 2024) 145 S. Pradines & O. Onézime Swahili wall niches executed during a single ritual occasion the great wedding and the revelation of the bride37. The niches play an important social role and connect intimately the stucco work completed for this time in harem niches to the owners of the house. Mostly the niches displayed imported porcelain plates and bowls38. They were also used to display imported bronze artifacts, and manuscripts, such as Qurʾans. Porcelain dishes placed or inserted in these niches are symbols of high social status, but also protective charms against evil spirits. These objects also symbolize an established order and give ritual value to the domestic space. The Zidakas have an arch comparable to that of the mihrabs of the local mosques. However, in religious buildings such as the mosques and tombs, the ceramics are directly embedded in the masonry of the qibla wall and the niche of the mihrab. The bathroom is invariably placed at the back of the house or near the central courtyard. It is divided into two small spaces; one has a pit for the latrines and the other a sink with one or two water pots. The bathroom is also decorated with plasterwork and niches39. According to Donley-Reid, the use of niches in wall ornamentation reflects the role of India in the genesis of the Swahili domestic architecture, especially in Gujarat and Kutch40. The niches inside the Gujarati and Kutchi houses obey to a notion of balance of spatial harmony, a regularity of space. This balance is based on a symmetry of architectural elements41. The stucco niches of Indian bohra houses look like miniature mihrabs42, these alcoves, called “Gokala”, are often associated with magic squares containing the name of Allah. The niches of the Indian house were used to house oil lamps, precious ceramics, or incense burners on either side of the front door to protect the occupants43. These niches acted as ritual purifiers, and they served mainly as receptacles for valuable items, and the niches were themselves decorative and protective elements44. For Donley-Reid, the origin of stucco niches in 18th century Lamu comes from India45, especially from the city of Surat when the Bohra, Shiʿa Ismaili J. de V. Allen, “The Swahili house: cultural and ritual concepts underlying its plan and structure”, 1979, p. 17–20. 38 U. Ghaidan, Lamu. A study in conservation, 1976, pp. 5–7. 39 J. de V. Allen, “Swahili ornament: a study of the decoration of the 18th century plasterwork and carved doors in the Lamu region”, 1973a; id., “A further note on Swahili ornament”, 1973b. 40 L.W. Donley-Reid, “Symbolic meaning within the traditional Hindu and muslim houses of Gujarat (India) and Lamu (Kenya)”, 1991, p. 78. 41 L.W. Donley-Reid, “Symbolic meaning within the traditional Hindu and muslim houses of Gujarat (India) and Lamu (Kenya)”, 1991, p. 81. 42 E. Lambourn, “A self-conscious art? Seeing micro-architecture in Sultanate South-Asia”, 2010. 43 L.W. Donley-Reid, “Symbolic meaning within the traditional Hindu and muslim houses of Gujarat (India) and Lamu (Kenya)”, 1991, p. 77. 44 R. Lewcock, “Architectural connections between Africa and parts of the Indian Ocean littoral”, 1976, pp. 17–20. 45 L.W. Donley-Reid, “Symbolic meaning within the traditional Hindu and muslim houses of Gujarat (India) and Lamu (Kenya)”, 1991, p. 78; S. Pradines, “L’influence indienne dans l’architecture swahili”, 1999, pp. 112–113. 37 nCmY 19 (Juillet 2024) 146 S. Pradines & O. Onézime Swahili wall niches traders and merchants were extremely active in the Indian Ocean46. Indian houses have separate commercial areas and strictly private domestic spaces. In Gujarat, the front of the house was used to welcome visitors, customers, and merchants. The women live at the back of the building, which has an independent exit from the main entrance on the front. In East Africa, it seems that the apparent absence of market structures and bazar influenced the tradition of using domestic houses for trade and commercial activities47. The spatial organisation was similar between Swahili and Gujarati houses for the 18th centuries48. This tripartite plan is also visible in Hadrami dwellings of Yemen which had very close ties with East Africa and India. It is extremely interesting to note that the best and latest examples of Swahili niches are to be found in 18th century urban centres and not much in 19th century cities such as Zanzibar stone town. It shows that the origin of these niches is first local African, and they are not connected to the late re-Arabisation of the coast during the Zanzibari Sultanate and the Omani rule49. And I agree with Sandy Prita Meier when she wrote that the Swahili stuccowork was an older fashion already disappearing in the early 19th century50. All the large houses and palaces built in Zanzibar during the 19th century have a plan completely different from the traditional Swahili stone houses51. And the Swahili houses from 19th century were more inspired by Omani and Gujarati models52. Whereas Zanzibar is always presented as the “pearl of Swahili culture”, in fact, this period marks a revival but also the end of Swahili traditional architecture. III. Archaeology of Swahili wall niches from the 12th century to the 17th century Based on our archaeological fieldwork and the existing literature on Swahili architecture, we were able to trace backward the evolution and history of Swahili wall niches from the 12th to the 17th century. We will demonstrate this continuity with the Swahili wall niches from the 18th century examples described previously to the earliest examples from the medieval period. L.W. Donley-Reid, “A structuring structure: the Swahili house”, 1990b; A. Mitra, Report on house types and village settlement patterns in India, 1961; K.S.K. Dongerkery, Interior Decoration in India, 1973. 47 T. Gensheimer, “House as marketplace. Swahili merchant houses and their urban context in the later Middle Ages”, 2018a, pp. 122–125. 48 M. Desai, “Traditional Bohra Dwellings of Gujarat, India: Architectural Response to Cultural Ethos”, 2023, pp. 255–258. 49 J.C. Anene, “The Omani Empire and its Impact on East African Societes”, 1966; C. Le Cour Grandmaison, “Parenté, migrations, alliances. Les réseaux omani en Afrique orientale et centrale”, 1991, pp. 163–177 ; id., “L'héritage arabe, XVIIIe et XIXe siècle”, 1998. 50 S.P. Meier, Swahili port cities: the architecture of elsewhere, 2016, pp. 152–153. 51 D. Rhodes, C. Breen & W. Forsythe, “Zanzibar: A Nineteenth-Century Landscape of the Omani Elite”, 2015. 52 A. Sheriff, “Mosques, Merchants and landowners in Zanzibar Stone Town”, 1995, p. 97. 46 nCmY 19 (Juillet 2024) 147 S. Pradines & O. Onézime Swahili wall niches Pre-18th century Swahili houses had wall niches, these niches were generally single and simple squares or rectangle openings sometimes surmounted by pointed arches. Most of the time, decorated walls had one or two separated niches, generally located in the middle of the wall, around 100 to 120 cm above the floor level. Niches were also present in door frame and door jamb, always arranged in symmetrical pairs. These Swahili decorative niches were made of cut and carved sea coral bricks. This building tradition started around the late 11th century until the end of the 17th century, where sea coral niches were replaced by stucco niches. The ancient capital of Mafia, Kua, is located on the small island of Juani to the south-west of Mafia Island at the mouth of the Rufiji River in Tanzania53. The site is mentioned as Kahua on the map of al-Idrisi in 115454. In 2018, we recorded different phases of occupation on the site from the late 11th to early 18th century. The main and principal occupation phase in Kua occurred during the 16th–17th centuries in many areas of the site. Most of the buildings, mosques, tombs, and houses date from this period. Kua was a large Swahili site, and it was during the Portuguese era that the town saw its greatest expansion with the creation of various compounds. The site started to decline during the 18th century and many people left the site of Kua and the Island of Juani for the Island of Chole. The major characteristic of Kua is the presence of large stone houses55 with a very peculiar plan called “mirror” or “twin” houses56. In fact, this plan reflects extended family living under the same roof with two different households. During the 18th century, the Swahili house plan will change for a first floor designated for extended family instead of a horizontal extension or “twin” houses. The 17th century houses in Kua were decorated with single niches built in the middle of some walls [Fig. 14]. Nine different types of niches were recorded in Kua, the most common being simple rectangles or square niches with round arches, generally one niche per wall [Fig. 15]57. The niches were made of mouldings of fresh sea coral bounded by lime mortar and covered by a thick stucco. The exception is House 3, which has a wall panel with friezes of three rows of 20 stucco niches with pointed arches and dates to the very late 17th century [Fig. 16]. Kua is really a site of transition between single sea coral niches with simple shapes from the medieval period to elaborate stucco wall niches with complicated designs from the modern period and the 18th century. Songo Mnara was a medieval town located on the eastern coast of the eponymous island, in the bay of Kilwa, South of Tanzania. According to our excavations in 53 G.S.P. Freeman Grenville, The Medieval History of the Coast of Tanganyika, 1962b, pp. 211–215; P. Garlake, The Early Islamic Architecture of the East African Coast, 1966, pp. 164–168 and 198. 54 S. Pradines, “Kua, Tanzania, excavation report 2018”, 2020. 55 A. Christie, “Structures and Settlement Organization at Kua Ruins, Juani (Mafia Archipelago, Tanzania)”, 2019, p. 257. 56 P. Garlake, The Early Islamic Architecture of the East African Coast, 1966, pp. 108–109; S. Pradines, “Swahili domestic architecture, between warehouses and palaces”, forthcoming. 57 A. Christie, “Structures and Settlement Organization at Kua Ruins, Juani (Mafia Archipelago, Tanzania)”, 2019, pp. 254–256. nCmY 19 (Juillet 2024) 148 S. Pradines & O. Onézime Swahili wall niches 2004, all the architectural remains visible in Songo Mnara are from the 15th century58. We have identified 36 stone houses, plus the so-called palatial complex59. The houses in Songo Mnara have monumental doors with stairs leading to central sunken courtyards and three rectangular parallel rooms60. Deprived of light and narrow, the rooms were to serve only as bedrooms or warehouses; everything suggests that social life took place mainly outside in the courtyards; as for the meals, they were probably taken under the verandas. Only the back part of the house was private, and the harem was reserved for the family61. The biggest house of the site, the so-called “palace” of Songo Mnara had a large sunken courtyard surrounded by a portico. The large doors opening onto the courtyard are framed with beautiful mouldings and sea coral niches [Fig. 17]62. The Kenyan site of Gedi, near the present town of Malindi, is an extremely interesting site for the study of Swahili niches63. According to our excavations from 1999 to 2003, most of the stone houses visible date back to the end of the 14th or early 15th century64. The plans of the houses present narrow and elongated rooms, all parallel except one which forms a side corridor and a courtyard attached to the house. Some rooms have symmetrically arranged single niches in the walls. The niches are generally located in the middle of the walls. They are organised symmetrically facing each other’s. Sometimes niches are present in door jambs such as house HO31. The bathroom and toilets are also decorated with single wall niches, such as houses HO 10 and HO 7 [Fig. 18]. The niches in bathroom were used to store freshwater pots for ablutions and cleaning65. This type of niche has a cup and a small hole drilled in its base. Three types of arches were used for these niches, simple straight lintel, round arch, and pointed arch without central voussoir. The arches and mouldings around the niches were made of carved blocks of fresh coral. The palace of Gedi is a large building characterized by a monumental gate of 5 m high framed by benches and surmounted by a series of small square niches [Fig. 19], the entrance is leading to a platform with a S. Pradines & P. Blanchard, “Kilwa al-Mulûk”, 2005. S. Pradines & P. Blanchard, “Kilwa al-Mulûk”, 2005; P. Garlake, The Early Islamic Architecture of the East African Coast, 1966, pp. 118–119, figs 73–74. 60 P. Garlake, The Early Islamic Architecture of the East African Coast, 1966, pp. 92–93, 195. 61 S. Pradines, Gedi, une cité portuaire swahilie, 2010, pp. 111–112, 125–134. 62 T.H. Wilson, “Swahili funerary architecture of the North Kenya coast”, 1979, pp. 7–16; T. Gensheimer, “Swahili houses”, 2018b; M. Baumanova & L. Smejda, “Space as material culture: residential stone buildings on the precolonial Swahili coast in comparative perspective”, 2018; J. de V. Allen, “The Swahili house: cultural and ritual concepts underlying its plan and structure”, 1979, p. 24. 63 S. Pradines, Gedi, une cité portuaire swahilie, 2010, pp. 159–185. 64 S. Pradines, Gedi, une cité portuaire swahilie, 2010, pp. 111–157, 201–210. 65 J. de V. Allen, “The Swahili house: cultural and ritual concepts underlying its plan and structure”, 1979, p. 25. 58 59 nCmY 19 (Juillet 2024) 149 S. Pradines & O. Onézime Swahili wall niches sunken courtyard below66. The rectangular courtyard runs parallel to the main body of the palace, it is bordered by platforms to the east, west and south. A row of steps in the middle of the southern platform gives access to the main apartments of the palace. The main wing of the palace is preceded by a reception room, with a facade decorated with niches and pegs holes to hang silks and embroideries [Fig. 20]67. The court of audience, the main wing in the south and the western wing were built at the beginning of the 15th century68. In Tanzania, Kilwa Kisiwani was the largest Swahili metropolis before the arrival of Europeans in the Indian Ocean69. According to the Chronicles of Kilwa, the city was founded by refugees from Shiraz between 957 and 98570. Later, the city of Kilwa was endowed with a palace-caravanserai, the “great castle” “Husuni Kubwa”. The palace is located on a promontory in the east of the island and dated to the 14th century thanks to a lapidary inscription in sea coral, dedicated to the glory of Ḥasan b. Sulaymān, who reigned over Kilwa from 1320 to 133371. One of the most characteristic elements of the palace is a large, sunken audience courtyard, surrounded by monumental steps and a wall decorated with a multitude of small square niches of 30 cm by 30 cm in size [Fig. 21]. The great palace of Kilwa is interesting because it is one of the oldest examples of Swahili domestic architecture with series of wall niches. Another 14th century example of Swahili niches with single round arches in domestic architecture is visible in Shanga, Lamu Archipelago, Kenya72. Swahili coral stone houses were already in use as early as the late 11th to the middle of the 12th century. As we excavated in 2005–2006 a stone house from this period in Sanje ya Kati, in Kilwa Archipelago, Tanzania. The heyday of the city of Sanje ya Kati and most of the constructions date back to 1050–1150 AD. The city was definitively abandoned in the early 13th century. The house excavated in Sanje is composed of three parallel and narrow spaces [Fig. 22]. During our excavations of the eastern part of the two long rooms, we found seven large Chinese jars73. Some of them were still in situ with the base still fixed on the ground floor and they demonstrate that these rooms were clearly storage rooms. The back of the house consisted of two rooms 3.2 m wide, a small storeroom to the southwest and a rectangular space to the southeast, possibly a bedroom. The creation of stone houses that serve as habitats and stor- 66 J. Kirkman, Gedi, the palace, 1963, pp. 6–7; S. Pradines, Gedi, une cité portuaire swahilie, 2010, pp. 159– 161, 178–180. 67 R. Lewcock, “Architectural connections between Africa and parts of the Indian Ocean littoral”, 1976, pp. 17–20. 68 P. Garlake, The Early Islamic Architecture of the East African Coast, 1966, p. 96. 69 S. Pradines & P. Blanchard, “Kilwa al-Mulûk”, 2005. 70 G.S.P. Freeman-Grenville, The Medieval History of the Coast of Tanganyika, 1962b, p. 89. 71 S. Pradines, Fortifications et urbanisation en Afrique orientale, 2004, pp. 111–112. 72 M. Horton, Shanga, 1996, pp. 41–44; M. Horton & J. Middleton, The Swahili, 2000, p. 118. 73 S. Pradines, “L’île de Sanjé ya Kati (Kilwa, Tanzanie)”, 2009. nCmY 19 (Juillet 2024) 150 S. Pradines & O. Onézime Swahili wall niches age areas illustrate the emergence of a merchant elite. This very functional architecture transforms the house into a warehouse (storeroom) before being a place of habitat. These commercial buildings generated domestic architecture specific to the Swahili coast. It was long time believed that this organization of space was an Indian influence from Gujarat74. In fact, the plan of the Swahili houses is from a much older Islamic architecture tradition that probably came from the Gulf or the Sindh. The masonry techniques were extremely elaborate with marine coral ashlars cut into a square jointed with a quality lime mortar with regular horizontal courses75. Unfortunately, the walls of the house were preserved up to 1 m and it is generally above this level that wall niches were built. However, because of the tripartite plan of this house identical to 15th–18th century models, it is strongly probable that niches were in use in Swahili domestic architecture since this period. To support my proposal that wall niches existed in Swahili houses at least since the 12th century, it is important to connect building technologies with interior decorations, but also to understand Swahili architecture by including religious monuments in our study as niches appear frequently in Swahili mosques and tombs such as Mwana [Fig. 23]. Swahili monumental tombs were built in the center of cities near prestigious public places such as the great mosque, the palace, certain neighborhood mosques or certain large stone houses of notables. These tombs marked the attachment of a family or clan to a district of the city. They were social references for the community, they acted as a place of memory which reinforced the role of ancestors, the prestige and authority of clans and ruling families within the city. The pillar tombs are the most symptomatic funerary monuments on the Swahili coast76. These constructions are rectangular paneled platforms to which a column or pillar has been added, located to the East of the grave above the head of the buried body. The column can be from 1 m to 14 m high and can have a round, square, polygonal, or fluted section with vertical moldings of triangular profile as in Gedi or Lamu [Fig. 24]. The main pilar tomb in Gedi has six small square niches similar in size to the ones above the main entrance of the palace and the niches in Husuni Kubwa in Kilwa [Fig. 25]. The oldest Swahili tomb with a dated inscription is at Barawa in Somalia, dating back to 1104–1105 AD77. At Shanga, a fluted column tomb is dated to the late 13th century by associated levels of occupation. This dating is confirmed by a Longquan (Northern Song) celadon bowl from the same period and inserted in the masonry of the burial78. A pillar tomb from L.W. Donley-Reid, “Symbolic meaning within the traditional Hindu and muslim houses of Gujarat (India) and Lamu (Kenya)”, 1991; S. Pradines, “L’influence indienne dans l’architecture swahili”, 1999. 75 M. Horton, Shanga, 1996, pp. 26–28, fig. 5. 76 U. Schlegelmilch, Typologie und datierung der sogenannten Pillar tombs, 1982; R. Wilding, “Panels, Pillars and Posterity: ancient tombs on the North Kenya coast”, 1988. 77 G.S.P. Freeman-Grenville, The Medieval History of the Coast of Tanganyika, 1962b, pp. 98–122. 78 M. Horton, Shanga, 1996, p. 71. 74 nCmY 19 (Juillet 2024) 151 S. Pradines & O. Onézime Swahili wall niches Kiunga in Kenya is decorated with 14th century ceramic material79. The pillar tombs of Ras Mkumbuu, on the island of Pemba, also date from the 14th century. The large pillar tombs of Malindi are more recent and date from the city’s heyday in the late 15th or 16th century. The Kunduchi cemetery in Tanzania brings together numerous later pillar tombs [Fig. 26], certainly from the 17th and 18th century like the small column of Mambrui north of Malindi [Fig. 27]. A sort of hood at the top of the steles has been interpreted for the Swahili tombs as a phallus to validate the African nature of these buildings. In fact, there are many examples of pillar tombs throughout the Muslim world, notably in Fatimid Egypt and Seljuk Iran, and later in Ottoman Anatolia80. The bulge in the upper part of the pillar represents a stylized turban, this headgear was once hung at the top of the tombs [Fig. 28]. Among the Swahilis, the column symbolizes the presence of an important buried person, generally a member of the Ungwana caste or notable of the city. It seems that the greater the fame of the deceased, the larger the pillar. This prestige could be acquired by a pilgrimage to Mecca. The tombs of sheriffs and saints attracted many pilgrims like the tomb of Sheikh Fakihi Mansuru in Takwa. There is an explicit sexual division of these monuments, the pillar tombs were reserved for men while many more recent domed tombs were often attributed to Swahili princesses. Mausoleums topped with a conical or bulbous dome appeared in the 17th century. This style of burial is very well represented in the Lamu Archipelago, notably in the cities of Siyu and Pate, but also in the Comoros, notably in Mutsamudu and Tsingoni [Fig. 29]. The inner walls sometimes have niches for storing incense burners. These mausoleums are funerary enclosures where entire families are grouped together. The facades of the large tombs and funerary enclosures were decorated with niches and recessed panels framed by coral stones; most of the time these niches were frames for ceramic inlays such as in Siyu on Pate Island [Fig. 30]81. The multiplication of niches symbolizes the Ndani, the ritual room in the Swahili house where many rites of passage take place82. Pillar tombs were also decorated with ceramics inserted into the masonry. The inlaid dishes or bowls are very useful for a relative chronology of these architectural structures because they are mainly pieces imported from China, celadons, or blue and white porcelains such as pilar tombs in Gedi, Mambrui and Kunduchi. The ornamentation of the tombs with ceramics recalls the Waungwana origin of the deceased, individuals of high social status in Swahili society. Swahili mosques were most of the time commissioned by private individuals, some mosques, such as Kilwa and Mogadishu, were built by local Sultans83. Mosques were also a way to display power and social influence. Wall niches are present in Swa- T.H. Wilson, “Swahili funerary architecture of the North Kenya coast”, 1979. S. Pradines, “Rituels funéraires swahili”, 2000. 81 S. Pradines, Historic mosques in sub-Saharan Africa, 2022, pp. 207–218. 82 T.H. Wilson, “Swahili funerary architecture of the North Kenya coast”, 1979. 83 S. Pradines, Historic mosques in sub-Saharan Africa, 2022, pp. 232–237, 249–254. 79 80 nCmY 19 (Juillet 2024) 152 S. Pradines & O. Onézime Swahili wall niches hili mosques, and these niches could be utilitarian or decorative84. Single wall niches were built in the walls of the prayer room to host documents, mainly Qurʾans and manuscripts [Fig. 31]. Very small niches were used to shelve small incense burners and oil lamps [Fig. 32]. Wall niches are frequently present on the frame of mihrab, generally in pair both side of the opening of the niche [Fig. 33]. The recess of mihrabs were sometimes decorated with a frieze of small niches made of sea coral stone like in Songo Mnara for the 15th century [Fig. 34]. The earliest example known in Kizimkazi in Zanzibar Island is dated from 1107 AD85. In Swahili mosques, the inlaying of ceramics, in the qibla wall, mihrab niches and sometimes the ceilings were also very common. Neville Chittick distinguished several categories of incrusted ceramics in Kilwa: Persian bowls with turquoise blue glaze and black silhouettes with floral or palmette motifs. Chinese blue and white porcelain bowls, and large Chinese blue and white porcelains and green celadon dishes86. The 14th century mihrab in Kua shows a combination of small niches decorated with blue and white porcelains and Persian turquoise bowls [Fig. 35]. At the time when Swahili trade reached its apogee, between the 14th and the 15th century, inlay ceramics multiplied on the tympanum, the lintel, and the architrave of the mihrabs and sometimes even in the niche. This tradition was to persist until the end of the 18th century. Twelve ceramics decorated the squinches and the jambs of the mihrab of the mosque (ǧamiʿa) of Chwaka, thirteen blue and white porcelain were inserted in the tympanums and the intrados of the mihrab of the ǧamiʿa of Gedi [Fig. 36], seventeen in the mihrab of Mafui, while the mosque of Tundwa held the record with over fifty ceramics inlaid in the wall of the qibla [Fig. 37]. The ceramics inserted in the masonry of mosques and tombs were usual potteries, to the opposite of the glazed tiles produced for the architecture. These ceramics, small bowls, or large plates were inlaid on the circumference of the mihrabs, or in the coating of the cupolas surmounting the prayer hall such as the small mosque in Kilwa87 [Fig. 38]. So when were Asian ceramics introduced into the Swahili architecture? Between the 11th and the 14th century there was a veritable craze for plates and cups inlaid in the coating of religious or civic buildings in the East. This Oriental fashion, known as that of the bacino88, spread across the entire Mediterranean basin in contact with Muslim countries—from Spain, to the south of France and to northern Italy (Tuscany and Venice). It led to a public display of the wealth of a town or an individual since the inlay ceramics came from far off, often from a Muslim country89. This custom seems to have originated in the East and can be explained by the fact that the poly- 84 S. Pradines, Historic mosques in sub-Saharan Africa, 2022, pp. 180, pp. 208–212. S. Pradines, Historic mosques in sub-Saharan Africa, 2022, pp. 219–222, 238–242. 86 N. Chittick, Kilwa: an Islamic Trading City on the East African Coast, 1974, pp. 161–166, 306–308. 87 N. Chittick, Kilwa: an Islamic Trading City on the East African Coast, 1974, p. 36. 88 O. Watson, Ceramics from Islamic Lands, 2004. 89 S. Pradines, “Le mihrab swahili”, 2003. 85 nCmY 19 (Juillet 2024) 153 S. Pradines & O. Onézime Swahili wall niches chromic Islamic pottery of this period was of the highest quality and could serve as a decorative element. This decorative tradition took root in East Africa in the 13th century90. The fashion of buildings covered in ceramics must have come from Seljuq Anatolia (1219–1237), as we see in the town of Konya. Later, under the Timurids in the 15th century, covering tiles in ceramic reached their apogee. Certain Swahili cities, such as Mogadishu, were strongly influenced by Seljuq architecture.91 The use of inlay ceramics in the mosques of the East African coast was a part of this fashion and of a cultural movement peculiar to the Turkic-Arabic and Mediterranean world.92 This use of bowls and plates of Chinese blue and white porcelain as a decorative element for the qiblas or the mihrabs was also known in Oman93 in the mosques of the 16th century such as the mosque of al-ʿAyn in Manah of 1505 and those of the town of Nizwa,94 alShawadhna dating from 1529, al-Sharja of 1518 and al-Jinah of 1519.95 Chinese Ming porcelains were also inlaid in the mihrab of the mosque of al-Manah, dating from 150596. The fashion of inlay ceramics in the mihrabs, however, seems to have been earlier in East Africa than in Oman and was on the rise. And yet there was a far earlier substratum of influence of Omani mihrabs going back to the Nabahani kings. Ruba Kana’an suggested that the Omani mihrabs were influenced by Fatimid Egypt and Yemen.97 The decorative repertoire, certainly, is sometimes similar, but the dates are very different, and one could object that it was the Fatimids who were influenced by the Abbasid repertory. Another scholar, Soumyen Bandyopadhyay thinks that this type of mihrab was introduced into Oman in the 10th century by way of Sohar98. The great port of Sohar was visited by al-Muqaddisi who described the mihrab of the mosque decorated by polychrome earthenware. In this case the Swahili mosques and mihrabs would have been influenced by two waves of migrants and travellers, first under the Buyids between 945 and 1055, then under the Seljuqs between 1038 and 1118. The ports of Fars and Siraf played a far from negligible part in the exportation of these models, first to Oman and then to Africa. 90 S. Pradines, Historic mosques in sub-Saharan Africa, 2022, pp. 208–217. P. Garlake, The Early Islamic Architecture of the East African Coast, 1966, p. 114. 92 S.P. Meier, “Chinese Porcelain and Muslim Port Cities: Mercantile Materiality in Coastal East Africa”, 2015; V.-S. Schulz, “Artistic Dynamics across the Seas”, 2018, pp. 191–194. 93 M. Kervran, “Mihrab/s omanais du XVIe siècle : un curieux exemple de conservatisme de l’art du stuc iranien des époques seldjouqide et mongole”, 1996, pp. 126–127. 94 P.M. Costa, Historic Mosques and Shrines of Oman, 2001, p. 223. 95 M. Kervran, “Mihrab/s omanais du XVIe siècle : un curieux exemple de conservatisme de l’art du stuc iranien des époques seldjouqide et mongole”, 1996, pp. 126–127 ; P.M. Costa, Historic Mosques and Shrines of Oman, 2001, p. 59. 96 P.M. Costa, Historic Mosques and Shrines of Oman, 2001, pp. 73, 223. 97 R. Kana’an, “The carved stucco mihrab of Oman, forms, styles and influences”, 2008, pp. 251–257. 98 S. Bandyopadhyay, “From Another World! A possible Bûyid origin of the decorated Mihrab of central Oman”, 2008, pp. 374–379 ; M. Kervran, “Mihrab/s omanais du XVIe siècle : un curieux exemple de conservatisme de l’art du stuc iranien des époques seldjouqide et mongole”, 1996, p. 112. 91 nCmY 19 (Juillet 2024) 154 S. Pradines & O. Onézime Swahili wall niches Although we do not have many examples of ceramic inlays in Swahili domestic architecture, some examples exist such as the barrel vault ceiling of a room in the socalled “house of the secretary” in Songo Mnara, dated from the 15th century. The palace of Songo Mnara was also decorated by ceramics inserted in the vaults of the portico around the main sunken courtyard99. Finally, the great house, near the Friday Mosque in Kilwa was decorated with inserted Iranian and Chinese bowls in the vaults of the main rooms. Completely overlooked in the literature, one final point needs to be clarified in this chapter: it is the relationship between objects and architecture in early Swahili settlements from the 9th–10th century. The idea to display imported ceramics and wealth was probably much older than the Swahili architectural remains still visible today. One specific characteristic of Swahili material culture was the importation of foreign manufactured objects, textiles, glass, and ceramics. The importation of luxurious ceramics precedes the first coral stone houses, and these artefacts were probably displayed in traditional Swahili wattle and daub houses. For the Abbasid period, most of the imported ceramics found in Ironi Be (Dembeni, Mayotte) were coming from the Gulf, mainly Iraq and Iran and from China100. If some of these objects were maybe used in ritual and festive activities101, these ceramics were first symbols of wealth and power of the Swahili elite. The Persian ceramics were also symbols of the pedigree of the most important Swahili families who claimed Shirazi origins, they also connected the owners of the ceramics to Muslim and Asian material cultures and to a certain extend to Islam as new religion of the African elites. These ceramics were certainly displayed in the first Swahili houses exactly as Chinese porcelains and celadons were displayed in the 15th and 18th century houses. These objects represented the intercultural and economic links between the Swahili elite and other cultures of the Indian Ocean. Unfortunately, because these ceramics were displayed in niches, they are not present anymore, they were displaced by the owners of the houses when they left as these imported objects were of great value. What is extremely important here is to recognize the intime relationship between Swahili niches and imported ceramics. More than that, the Swahili niches help us to understand the role of the inserted ceramics in mosques and mausoleums. The inserted ceramics in religious buildings acted exactly as the ceramics in domestic context, they provided ritual protection and social status to the owner of a tomb or the patron of a mosque. In short, imported ceramics, inserted in buildings, or just displayed in niches tell us the same thing: the power of the Swahili elite and rich merchants, the Waungwana. 99 P. Garlake, The Early Islamic Architecture of the East African Coast, 1966, p. 37. S. Pradines, “Islamic Archaeology in the Comoros”, 2019. 101 J. Fleisher, “Rituals of Consumption and the Politics of Feasting on the Eastern African Coast, AD 700–1500”, 2010. 100 nCmY 19 (Juillet 2024) 155 S. Pradines & O. Onézime Swahili wall niches IV. Functions of Swahili niches: kinship, economic and politic According to Donley-Reid, niches in Swahili stone houses display expensive objects and imported goods such as Chinese porcelain plates [Fig. 39], they acted as talismans to attract the eyes of the visitors and to protect from negative thought and evil eye and bad spirits, jinn102. Such as the painted inscriptions on the beams of the 18th century Comorian palaces, with verses of the Qurʾan were divine blessings and protective charms for the owner of the house, his family, and visitors [Fig. 40]103. Talismanic calligraphy was common in Muslim houses such as Ottoman houses in Cairo104. During the 19th century, Swahili stopped to build niches in their houses, porcelain plates and imported objects such as glassware, weapons (swords and muskets) were displayed directly on plain walls [Fig. 41]105. As mentioned by Elizabeth Lambourn, the aesthetic beauty of Indian micro architecture has the strength to bound and to frame the space where “small” can be apprehended, this world conscious perception gives a feeling of dominance and possession106. Swahili niches obey to the same logic where the imagination is contented. More than dominance and possession, I prefer to say that these niches were used to create an idealistic and protective cosmopolitan cultural cocoon. As rightly wrote by Meier, “people give meaning to objects by arranging them in relationship to other things”. And she adds “Swahili coast visual culture emphasizes overseas connections”107. I totally agree with Meier, but I do not think that it is specific of the Swahili and their cosmopolitan nature. It is true that displaying these objects shows the cosmopolitanism of the Swahili, but to me, it was first to display the wealth and power of the elite. The proof is that Yemeni and Ethiopians displayed imported objects and had similar niches in their houses and palaces. Moreover, arranging Chinese plates in manors and castles was a fashion in Europe and it was called “Chinoiserie” in the 18th century England and France. It was also the case earlier with the city state of Venice displaying Middle Eastern textiles and ceramics. Later this fashion was called the “Turquerie” with display of Ottoman rugs and Iznik ceramics. Chinese porcelains were collected and displayed all over the world, for example in Iran the Chini Khaneh house of porcelains in the Shaykh Safi al-Din shrine in Ardabil108. For me, the most important is not to develop a specific Swahili visual culture separated from other world fashions and tastes from Europe to India; but more to integrate the L.W. Donley-Reid, “The power of Swahili porcelain, beads, and pottery”, 1990a, pp. 49–50; S.P. Meier, Swahili port cities: the architecture of elsewhere, 2016, pp. 154–155. 103 L.W. Donley-Reid, “A structuring structure: the Swahili house”, 1990b. 104 J. Campo, “‘This blessed place’: the talismanic significance of house inscriptions in Ottoman Cairo”, 2022, pp. 132–154. 105 S.P. Meier, “Objects on the edge. Swahili coast logics of display”, 2009, pp. 11–12. 106 E. Lambourn, “A self-conscious art? Seeing micro-architecture in Sultanate South-Asia”, 2010, pp. 150– 152. 107 S.P. Meier, Swahili port cities: the architecture of elsewhere, 2016, pp. 139–143. 108 E. Cooke, Global objects: toward a connected art history, 2022, pp. 171–172. 102 nCmY 19 (Juillet 2024) 156 S. Pradines & O. Onézime Swahili wall niches Swahili culture into an Indian Ocean multicultural diversity, beyond African stereotypes. Ceramics displayed in niches had aesthetic and symbolic significance; they represented beauty, civilisation, protection, and the cosmopolite status of the Wangwana109. The Wangwana visual culture celebrates mercantile oceanic trade and cultural exchanges. The interior space of the rich patrician Swahili houses was decorated with Chinese porcelains110. Staging cosmopolitanism with decorative objects was bringing to the family a direct connection to overseas social and cultural networks111. Layered assemblages of prestige objects had an important role in coastal African identity112. The display of Eastern and Western material culture exhibits the transcultural identity of the Swahili elite113. As mentioned by Meier, in Kiswahili, “Uwezo” means dignity, honour, power and will”. When objects have “Uwezo” as adjective, it is because they matter in Swahili social life. These objects and sometimes micro architecture have a direct effect on people, for example fine objects and porcelains are called “Mapambo ya nyumba”, the decoration of the house114. The Uwezo also shows the consumer religious affiliation and social mobility, for example by being Muslim from prestigious descent such as Shirazi, Yemeni, and Omani. Ceramics and niches were also used on religious buildings again to demonstrate the power of the Waungwana. Coral and stucco niches were key elements of Swahili Culture and played a social role and support the identity of the local elite. The stone house had an extremely important role in Swahili society, it was a place for trade as a warehouse115, a place for family, a place of power that represented the important linages of a city or a town, the notables or Waungawana. The Waungwana (sing. Mwungwana) were the elite of the Swahilis; they owned stone houses, boats, cattle, agricultural estates “Mbara”, and coconut plantations, “Mashamba”. These rich families of African origin strictly controlled the activities of Arab, Persian, and Indian merchants. The Waungwana hold economic and military power and regulate the activities that govern Swahili society. The Waungwana were free-born individuals, patricians and citizens living in stone houses in the center of town. The Mwungwana citizen has a legal status, while the city dweller simply has the right to live in the city. By intermarrying with the daughters of Waungwana, Arab and 109 M. Horton & J. Middleton, The Swahili, 2000, p. 112; S.P. Meier, Swahili port cities: the architecture of elsewhere, 2016, pp. 144–150. 110 S. Pradines, “L’Afrique noire et la Chine. La céramique importée : symbole du pouvoir des marchands swahili”, 2003. 111 S.P. Meier, “Objects on the edge. Swahili coast logics of display”, 2009, pp. 8–10. 112 L.W. Donley-Reid, “Life in the swahili town house reveals the symbolic meaning of spaces and artefact assemblages”, 1987. 113 S.P. Meier, “Objects on the edge. Swahili coast logics of display”, 2009, pp. 11–12. 114 S.P. Meier, Swahili port cities: the architecture of elsewhere, 2016, pp. 152–153. 115 S. Wynne-Jones, “The Public Life of the Swahili Stone house, 14th–15th centuries AD”, 2013, pp. 761– 772. nCmY 19 (Juillet 2024) 157 S. Pradines & O. Onézime Swahili wall niches Indo-Persian traders obtained the status of local dignitaries116. This status allowed them to carry out their commercial transactions under better conditions and to establish a family warehouse117. Kinship networks appear as the way Swahili groups were able to integrate foreigners, such as the Hadramis and the Omanis118. The term Mwungwana is in opposition to that of Mtwana, the captive, dependent man, or slave, an individual who does not have or has lost his freedom. The Washenzi and Wanyika refer to Cushitic pastoralists and Bantu farmers living in the hinterland. The people who live outside the city walls, in the bush. Swahili society expresses this permanent duality between Arabo-Persians and Africans, patricians and city dwellers, Muslims, and non-Muslims, rich and poor, freemen and slaves119. Swahili cities were managed by the Waungwana, a council of elders, sometimes with an elected leader called ‘Sultan’. The Mwungwana system was based on an assembly of notables elected from the most respectable and oldest clans of the Swahili city. The Mwungwana council’s members belonged to the descendant of the oldest families and founders of the city. This municipal council manages the city and its territory; it does not obey any higher authority. Examples of the Mwungwana system are numerous and ancient. In the 13th century, Yāqūt noted that Mogadishu had no sultan; affairs were settled by an assembly of families of notables120. The Chronicles of Pate describe the council of elders of Manda121. In 1331, Ibn Baṭṭūṭa writes that Mogadishu is managed by a council of elders headed by a Sheikh. In Barawa, this council was composed of twelve notables and was observed by Charles Guillain and Georges Révoil in the 19th century122. This organization is found in Siyu and Pate which were ruled in the 19th century by a sultan and a council of elders123. The “Sultanate” was also based on the Mwungwana council. The difference lies in the election of the main leader to head this council. For example, the city of Mogadishu was ruled by a Sultan who was elected by an oligarchic council. The members of this council were themselves drawn from the highest and most representative clans of each district of the city. The ruler was often chosen from families of Sheriffs who had legitimacy in the genealogy and Muslim scholars of the city. It is ultimately difficult to say whether the Sultanate, 116 F. Le Guennec-Coppens & D. Parkin, Autorité et pouvoir chez les Swahili, 1998, p. 10. L.W. Donley-Reid, “A structuring structure: the Swahili house”, 1990b, pp. 123–124. 118 F. Le Guennec-Coppens, “Qui épouse-t-on chez les Hadrami d’Afrique orientale ? Les réseaux d’alliances”, 1991; F. Constantin, “Condition swahili et identité politique”, 1989, p. 342. 119 M. Horton & J. Middleton, The Swahili, 2000. 120 M. Horton, Shanga, 1996, p. 425. 121 G.S.P. Freeman-Grenville, The East African Coast: Select documents from the first to the earlier nineteenth century, 1962a, p. 248. 122 S. Pradines, Fortifications et urbanisation en Afrique orientale, 2004, pp. 340–343; G.S.P. FreemanGrenville, The East African Coast: Select documents from the first to the earlier nineteenth century, 1962a, p. 78. 123 W.H. Brown, History of Siyu: the development and decline of a Swahili town on the northern Kenya coast, 1985. 117 nCmY 19 (Juillet 2024) 158 S. Pradines & O. Onézime Swahili wall niches called sometimes “Shirazi system” was the adoption of a Persian or an Arab model, probably Yemeni, or whether it is the evolution of the local political system of the Swahili city-states. This political organization seems to appear in the 13th century and develops mainly in southern Kenya, Tanzania, and the Comoros. The palace became the place of residence of the sultan, like the Husuni Kubwa of Kilwa, the Yumbe of Pate or the Ujumbe of Mutsamudu. The Swahili palace represented the place of power where meetings of elders take place. The strong characteristic of Swahili society was the adaptation and integration of new elites and foreigners. The travelers and merchants, first Persians then Arabs and Indians, came to settle in Africa and mix with the local populations. These foreigners brought with them their cultural traditions which was integrated in the founding myths of the largest African city-states of the coast. The lineage is the most important element among the Swahili elites. A Swahili notable must have prestigious ancestors, Shirazi, Hadrami, Omani or Gujarati. These origins strengthen the local authority of the elites and provide them with privileged commercial links with other economic partners in the Indian Ocean. As Swahili society is very hierarchical, the prestige of everyone depends on his lineage and ancestors124. For example, in the late 17th century, Shirazi rulers of Anjouan started to marry their daughters to Mahdali rulers of Pate. These Mahdali claimed to be descendants of Yemeni sharifs from the Hadramawt. The newcomers settled in Domoni and Mutsamudu, hometowns of their Comorian spouses. These marital agreements allowed Comorian rulers to be connected to one of the most powerful Swahili City-State from the coast at that time, the city of Pate in the Lamu Archipelago. These new Indian Ocean networks were extremely useful to Comorian leaders for economic and religious reasons. Religious, because the sharifian linage from the Hadramawt, from the family of the Prophet, re-enforced their prestige in the Muslim community. Economic, because Pate was resisting to the Portuguese hegemony on the coast, the city had strong and independent commercial links with Yemen and India. Later Pate was also resisting to the Omani control, this last stand was dreadful to Pate as the power shifted to Zanzibar during the 19th century. The new elite forged in Anjouan were called the “Makabaila”. There was a social cleavage between “Makabaila”, the “noble men” and city people, as opposed to the “Wamatsaha”, “the men of the grass”, the rural people. The slaves were called the “Wazalia”. From the 17th to the 19th century, the Makabaila, were a social dominant group of Muslim scholars, merchants, and landowners125. Makabaila can be translated as “nobles” and it has a similar meaning to Waungwana on the Swahili coast. The large stone houses of the main cities of Domoni and Mutsamudu were owned by the Makabaila. The stone houses of the Makabaila express this social division between urban R. Pouwels, “Tenth Century Settlement of the East African Coast: The Case for Qarmatian/Ismaʿili Connections”, 1974; id., “Oral Historiography and the Problem of the Shirâzi in East African Coastal History”, 1984. 125 S. Blanchy, “Anjouan (Comores), un nœud dans les réseaux de l’océan Indien. Émergence et rôle d’une société urbaine lettrée et marchande (XVIIe-XXe siècle)”, 2015, pp. 4–13. 124 nCmY 19 (Juillet 2024) 159 S. Pradines & O. Onézime Swahili wall niches versus rural, and rich versus poor. It was precisely during the late 17th and early 18th century that new decoration techniques were introduced in the houses, with stucco niches and wooden painted ceilings. Among the Swahili there is a division of habitat which is first social before being ethnic. The stone buildings in the heart of the city are reserved for the Waungwana, representing the Swahili elite126. Only wealthy merchants and noble families from the ruling lineages have the possibility of living in stone houses127. The stone houses are emblematic of the power of the Waungwana, the noble famillies, and represent the very notion of the Swahili ruling class. The stone house symbolizes the higher social status of Swahili merchants and rulers. The stone houses of the Waungwana or leading families could be rebuilt but never sold128. In Anjouan Island, the stone houses belong to the women through the matrilocal marital residence rule. In short, the husband comes to live with the wife to whom the house is transmitted by heritance129. Called the “Manyahuli” or maternal inheritance, it allows transmission of land by women. The same ritual was observed in Lamu. For the wedding of the daughter of a patrician, new lime pits were dug and new stucco niches decorations on the house were planned for the ceremony. It is the “Fola la Kuwaza”, feast of the stucco. The art of lime stuccowork is called “Uwezo wa Niumba”, the talent of the house130. The wall niches and the imported ceramics that they contained represented these different levels of the Swahili society: trade, wealth, power, family and social status. Due to limited urban space and growing families, the houses grew vertically, and more often one floor was added over time. An extended family shared the same house on two floors. A case of sisters living above each other is reported in Anjouan131. Contrary to an “idée reçue”, the upper floors are not assigned to guests in the 18th century two-storey houses in Lamu, Pate and Anjouan. The social division of space is not organized vertically but horizontally, exactly like in Swahili medieval houses. Each floor is as an independent family unit, and the whole house belongs to the extended family. V. Wall niches and Indian Ocean networks In Ethiopia, some similarities exist between Harari and Swahili niches, and they betray Yemeni and Indian influences. The walls of the Harari house, Gidir Gar, have eleven niches called “Taqet” [Fig. 42]132. Five niches are built in the wall opposite the main en- 126 F. Le Guennec-Coppens & D. Parkin, Autorité et pouvoir chez les Swahili, 1998. F. Siravo & A. Pulver, Planning Lamu: conservation of an East African Seaport, 1986. 128 J. de V. Allen, “The Swahili house: cultural and ritual concepts underlying its plan and structure”, 1979, p. 5. 129 J.-L. Guébourg, Espaces et pouvoirs en Grande Comore, 1996. 130 S.P. Meier, Swahili port cities: the architecture of elsewhere, 2016, pp. 152–153. 131 S. Blanchy, “La maison urbaine, cadre de production du statut et du genre à Anjouan (Comores), XVIIe-XIXe siècles”, 2022, pp. 3, 17. 132 E.-D. Hecht, “The city of Harar and the traditional Harar house”, 1982, pp. 59–61. 127 nCmY 19 (Juillet 2024) 160 S. Pradines & O. Onézime Swahili wall niches trance. Two long rectangular niches are utilized for storing books and Qurʾans. The small triangular niches are used to display porcelains and nice imported objects. The lower niches are used to store shoes and incense burners. Niches with bi- or trifoliate arches were used to exhibit glassware, Chinaware or lamps, and they are dated from 18th and 19th century. Carved lintels of Harari or neighbouring Argobba houses with carved decorations, some like those of Swahili houses such as mooring ropes and turtle motifs, show influences from the Swahili world, Yemen, and India133. The doors of Harari houses were generally done by Indian woodcarvers134. In 2003, I observed stucco niches in Fasil Ghebbi or Fasilädäs castle in Gondar, they were very similar Swahili niches. The palatial compound of Fasil Ghebbi was the capital of Ethiopia from 1636 to the late 18th century. A Yemeni Ambassador, Ḥassan b. Aḥmad al-Ḥaymī who visited Gondar in 1648, wrote that the main architect of the castle was an Indian stonemason called ʿAbd al-Karīm, a Banyan from Diu in Gujarat. According to Portuguese sources, the use of lime mortar and plaster was introduced in Ethiopia in 1621 by a Gujarati mason, probably the same man135. Indian artistic influences are well documented in Yemen, Ethiopia, Somalia, Lamu, and Comoros for the late period, 17th–18th century. For example, Indian Ocean Qurʾans show incredible similarities between the Swahili coast, Yemen, Ethiopia, and India, notably the use of black and red colors and motives136. In Yemen, niches with stucco mouldings and polylobed arches are found in Zabīd and Moka, cities strongly influenced by the Indian architecture of Surat and confirmed by historical sources137. The Yemeni houses have also long rectangular parallel rooms decorated with niches made of fired bricks and stucco. The four walls are covered with stucco niches, especially in the main entrance and the liwān. There is an exact symmetry in the decoration of the walls138. Most of the ceilings of these rooms are painted with black and red motives and inscriptions139. The houses in Mocha and Zabīd have wall panels of nine niches, three niches on three levels140. The niches are still used to display ceramics and porcelains. The niches of the house in Zabīd have a flat plain back; they were used as storage and display cases for porcelains. Some niches R. Wilding, “Harari domestic architecture”, 1976. E.-D. Hecht, “The city of Harar and the traditional Harar house”, 1982, pp. 62–63. 135 R.K.P. Pankhurst, History of Ethiopian towns from the Middle Ages to the early nineteenth century, 1982, p. 109; R. Shaalini, The Castle of Emperor Fasilädäs, 2001. 136 A. Bang, “The Ḥaḍramaut in Lamu. The manuscript collection of the Riyadha mosque of Lamu, Kenya”, 2014; Z. Hirji, “The Siyu Qurʾans”, 2019; S. Mirza, “Developing the Harari Mushaf”, 2023; id., “The Qurʾan Illuminations”, 2024. 137 P. Bonnenfant, “La marque de l’Inde à Zabîd”, 2000; A. Das Gupta, Indian Merchants, and the Decline of Surat c. 1700–1750, 1994. 138 P. Bonnenfant, Les maisons de Zabîd, 2008, pp. 175–198, 199–211. 139 P. Bonnenfant, Les maisons de Zabîd, 2008, pp. 223–241. 140 P. Bonnenfant, Les maisons de Zabîd, 2008, pp. 36–41. 133 134 nCmY 19 (Juillet 2024) 161 S. Pradines & O. Onézime Swahili wall niches had a function to support incense burners and oil lamps141. The symmetry of the Yemeni wall niches creates a harmony around a vertical axis passing through the center of the wall. The idea is to create an aesthetic and a visual comfort to the eyes. This spatial and visual beauty is like a protective cocoon for the owner of the house, his family, and his guest. Stucco niches are also reported in the Hadramawt, the Tihāma plain in Sabya and also in Farasān Islands142. The Yemeni cities of Zabīd and Sanaa attracted Hindu merchants from Gujarat the Banyans since the 16th century. These merchants developed a very flourishing trade especially with textile during the 17th century. The port of Moka became the most important harbour in the Red Sea during this period, especially with the coffee trade coming from Ethiopia143. In 1616, the Indian community was massive in the region with 3,000 Banyans working mainly in banking industry, silver and goldsmiths, and craft with wood carvers, stone carvers, and masons. In the 18th century, Indians are reported in all the Arabian Peninsula, especially those originated from Diu and Surat port-cities from Gujarat144. In Yemeni houses, the best niches are in the main reception room to impress guests, to the opposite the best niches of the Swahili houses are in the Ndani, the most private part of the house. The niches on the lateral door jambs are similar in Swahili, Yemeni and Gujarati architecture. Paul Bonnenfant proposed that Yemeni houses plans and decorations have been influenced by Indian architecture, especially from the Gujarat, the Kutch, and the Bohra houses from the city of Siddhapur. Walls with friezes of niches are common in Gujarat, but also in Rajasthan, and Panjab, for example in Sikandra in Agra, the Red Fort of Delhi [Fig. 43] or the Shalimar gardens in Lahore [Fig. 44]. Bonnenfant proposed that Indians brought a new style of stucco decoration in Yemeni trading cities around the 17th century. However Arab traders and Yemeni craftsmen have also been reported in Gujarat and Kerala and it is possible that these cultural influences were multidirectional. In fact, wall niches in the Indian Ocean are more documented on the Swahili coast than anywhere else. Swahili niches predate Indian Bohra niches and Yemeni niches from the Ottoman era. One can object that Swahili niches are much older than the 18th century. Therefore, the comparison with late modern Bohra houses might not be so relevant. If we focus on stucco wall panels design, we must seriously envisage the possibility of a Swahili influence towards the Arabian Peninsula and Gujarat. Even if I was in favour of Donley-Reid’s proposal about Indian/Gujarati influences twenty-five year ago145, I am not satisfied today with this explanation. Swahili niches seem to predate the African relationships with Gujarat 141 P. Bonnenfant, Les maisons de Zabîd, 2008, pp. 314–316. G. King, The Traditional architecture of Saudi Arabia, 1998, pp. 10–11, 56–60, 70–71. 143 M. Tuchscherer, “Les échanges commerciaux entre les rives africaine et arabe de l’espace mer Rougegolfe d’Aden aux seizième et dix-septième siècles”, 2004. 144 P. Bonnenfant, Les maisons de Zabîd, 2008, pp. 317–320; K.N. Chaudhuri, Trade and Civilization in the Indian Ocean. An Economic History from the Rise of Islam to 1750, 1989; M. Pearson, Merchants and Rulers in Gujarat, 1976; E. Alpers, “Gujarat and the trade of East Africa c. 1500–1800”, 1976. 145 S. Pradines, “L’influence indienne dans l’architecture swahili”, 1999. 142 nCmY 19 (Juillet 2024) 162 S. Pradines & O. Onézime Swahili wall niches and 17th century Bohra houses. There are a continuity of the Swahili wall niches from the 12th century to the 18th century146. Most of the archaeologists working in East Africa mention the relationships between the Swahili with Arabia, the Gulf and India. But they failed to provide a detailed and coherent narrative about the intercultural connections, material, and historical evidence across the Indian Ocean. So, from where did this plan originate? Oman in the Batinah, and several towns such as Ibri, Nizwa and Sur have ancient traditions of stucco art and designs with mouldings, niches, and several similar plaster decorations. Similar stucco niches are visible all over the Muslim world, from Madinat al Zahra in Spain to Kufa in Iraq. In Samarra, the decoration with marble and stuccoes, plus the codified hierarchy of spaces express the status and court privileges of the guests and the owners of these houses and palaces147. The Dār al-ʿImāra in Kufa was built during the Umayyad period and restored later by the Abbasids, it shows a strong influence from the Sassanian Persian architecture148. The plan of the Swahili houses shows similarities with the palace in Ukhaidir dated from 774, and stucco niches in Sāmarrāʾ with the rooms in the Dār al-Ḫilāfa palace (836)149, and the palaces and houses in al-Mutawakkiliyya and those of al Muʿtasim, Sur Isa house and Balkuwara150. Indian five lobes’ niches present also some similarities with the Abbasid stucco niches from palaces and houses in Sāmarrāʾ and Raqqa [Fig. 45]151. A miniature shows the house of wisdom “Beit al-Hekmeh”, the Abbasid library of Baghdad, and the books are stored in niches like those of the Swahili houses [Fig. 46]. Concerning architectural maritime influences in early Swahili architecture, we should first look at the Fars and Sindh from the 10th to the 12th century. The Abbasid stucco niches should be kept in mind as prototypes circulating from Basra in the Gulf up to Sindh. According to Swahili oral traditions, the “Wadebuli”, “people from Debul”, were great builders, and arrived in East Africa between the 9th and 12th century. There is an on-going discussion about the origins of the “Wadebuli and Wadiba”: for some researchers they are just myths like the Shirazi, for others the “Debuli” were connected to people from India and the “Diba” people from the Maldives and Sri Lanka or even Austronesians152. The problem is that the Debuli were not correctly identified. They were not people from the Bahamani port of Dabhol in Maharashtra (between Mumbai 146 M. Horton, Shanga, 1996, pp. 40–62, 58–60; M. Horton & J. Middleton, The Swahili, 2000, pp. 116–119. A. Northedge, “Abbasid Earth Architecture and Decoration at Samarra, Iraq”, 2004, pp. 8–12; M. Saba, “A restricted gaze: the ornament of the main caliphal palace of Samarra”, 2015, p. 187. 148 U. Siegel, “The architecture of the Abbasid residence at ar-Raqqa/ar-Rafiqa”, 2017, pp. 76–80. 149 R. Lewcock, “Architectural connections between Africa and parts of the Indian Ocean littoral”, 1976, pp. 17–20. 150 A. Northedge, The Historical Topography of Samarra, 2005, pp. 111–118. 151 P. Bonnenfant, Les maisons de Zabîd, 2008, pp. 314–316. 152 J. de V. Allen, “The Shirazi problem in East African Coastal History”, 1982; J. Gray, “The Wadebuli and the Wadiba”, 1954; M. Walsh, “The Diba, Debuli and related traditions of the East African coast”, 2010, pp. 454–465. 147 nCmY 19 (Juillet 2024) 163 S. Pradines & O. Onézime Swahili wall niches and Goa). The Wadebuli were most probably people from Daybul (Banbhore) in the Sindh delta and they are connected to early Muslim migrant communities in East Africa153. The early port of Daybul (today’s Banbhore) in the Sindh, played an important commercial role between India and Africa especially under the Ghaznavids (977–1186) and the Ghurids (879–1215). House with tripartite parallel rooms has been excavated in Banbhore and is dated from the 9th–12th century154. According to Alka Patel, possible Buddhist influences are visible in Indo-Persian architecture especially the niches with trefoil arches155. However, polylobed arches are also present in Sasanian architecture, such as the façade of Taq-I Kisra156. A series of Turkic and Persianate dynasties were certainly at the origin of the boom of urbanisation and architecture on the Swahili coast157. Especially under the Buyids of Fars (934–1062) as this period coincided very well with the so-called Shirazi in East Africa and the imports of Garrus wares or hatched sgraffito in the 11th and 12th centuries. Initially Shiʿa Zaydi, Buyids became Shiʿa Twelver in 941, the presence of Shiʿa communities is also attested in East Africa. As mentioned by Alka Patel, the Great Seljuks (1037–1194) created a cultural bridge between Anatolia into the Iranian world and these had repercussions on the Indian Ocean coasts158. As proposed by Soumyen Bandyopadhyay, some stucco designs have been introduced in Oman during the Buyid dynasty from Iraqi ports and Iranian ports from the Fars. The period 934– 1062 fits well with the role of Persians in Yemeni architecture for this period according to Guy Ducatez and the Swahili oral traditions for the same period159. In Siraf, the 9th– 11th c. houses of the residential quarters show an architectural continuity from the Sasanian to Buyid periods160. The houses with central courtyards are dated from the Abbasid and Buyid periods, 9th–11th century. The plan of Buyid houses in Siraf presents some similarities with the Swahili houses. The Iranian port of Siraf also played a central role in the Swahili historical traditions between the 10th and 12th century. Before being Shiʿa Twelver in 941, the Buyids were Shiʿa Zaydi, like in Yemen. It is extremely important to note the importance of Shiʿa communities in the early islamisation of the East African coast and the introduction of new building material such as fresh sea coral and plaster161. It is quite possible that Buyids were at the origins of the diffusion of late Abbasid architectural models in the Indian Ocean. This cultural 153 S. Pradines, Historic mosques in sub-Saharan Africa, 2022, pp. 227–231. S. Mantellini, “Investigating the core of the urban asset of the site”, 2019, pp. 72–85. 155 A. Patel, Iran to India: The Shansabanis of Afghanistan, c. 1145–1190, 2022, pp. 293–328. 156 A. Northedge, personal communication. 157 S. Pradines, Historic mosques in sub-Saharan Africa, 2022; id., “Swahili domestic architecture, between warehouses and palaces”, forthcoming; id., “The trees from the sea, coral stone architecture”, forthcoming. 158 A. Patel, Iran to India: The Shansabanis of Afghanistan, c. 1145–1190, 2022, pp. 280–283. 159 S. Pradines, “L’île de Sanjé ya Kati (Kilwa, Tanzanie)”, 2009. 160 D. Whitehouse, Siraf, history, topography and environment, 2009, pp. 34–42. 161 S. Pradines, Historic mosques in sub-Saharan Africa, 2022, pp. 227–232. 154 nCmY 19 (Juillet 2024) 164 S. Pradines & O. Onézime Swahili wall niches movement was pursued later in East Africa by the Great Seljuks (1037–1194), notably in Mogadishu162. VI. Conclusion Previously, the genesis of the Swahili architecture was summarized in a theoretical, and simplistic way, as pro-local development, and African evolution163 against partisans of Indian Ocean exogenous influences164. Indeed, Swahili architecture is the product of a constant African acculturation, marked by exogenous influences from the Indian Ocean worlds165. Arabs, Turks, Persians, Indians have all played a role in the genesis of the Swahili architecture. Swahili niches show some possible Abbasid influences from the Gulf and later from the Sindh. And it is probably from the Abbasid empire that this plan originated and was disseminated in the Indian Ocean. However, the evolution of the Swahili niches from the 12th to the 18th century shows some continuity and the Swahili had a major role in the diffusion of wall niches design in the Indian Ocean. Swahili niches and imported ceramics were elements of prestige reserved for the merchant elite166. The house was the central place of trade and exchange, but also a protected and sacred area. Swahili families and important lineages were represented by stone houses, pillar tombs and mosques. The Waungwana displayed their wealth in public and private architecture with imported Chinese and Persian ceramics framed in sumptuous coral or stucco niches. The study of Swahili architecture demonstrates a unique expression of an African culture turned towards the Indian Ocean, with complex multidirectional cultural relationships between Africa, India, the Arabian Peninsula, and the Gulf. Swahili stone houses were used to create and support trading networks in the Indian Ocean by creating familial bounds between foreign merchants and local women, daughters of African rulers. All these information’s are saying the same thing: the Swahili were and are maritime and cosmopolitan in nature. Greetings We would like to thank all the National organisations and NGO’s who supported our work all these years, especially the Tanzanian Museums and Antiquities, the National Museums of Kenya, the World Monument Fund (WMF), the Unesco, the Charity “Nat- 162 S. Pradines, Historic mosques in sub-Saharan Africa, 2022, pp. 232–238. M. Horton & J. Middleton, The Swahili, 2000, p. 119. 164 L.W. Donley-Reid, “A structuring structure: the Swahili house”, 1990b; A. Sheriff, “The historicity of the Shirazi tradition along the East African coast”, 2001, p. 76; S. Pradines, “L’influence indienne dans l’architecture swahili”, 1999, pp. 109, 115, S. Pradines, Fortifications et urbanisation en Afrique orientale, 2004, pp. 50–51. 165 S. Pradines, “L’île de Sanjé ya Kati (Kilwa, Tanzanie)”, 2009, pp. 67–71. 166 S. Pradines, “L’île de Sanjé ya Kati (Kilwa, Tanzanie)”, 2009, pp. 19–23; N. Um, The merchant houses of Mocha, 2009, pp. 135–161. 163 nCmY 19 (Juillet 2024) 165 S. Pradines & O. Onézime Swahili wall niches uralistes de Mayotte”, the Conseil Général of Mayotte, the DRAC “Océan Indien”, the CNDRS of Comoros and the Charity “Patrimoine des Comores”. Bibliography Allen, James de Vere. 1973a. “Swahili ornament: a study of the decoration of the 18th century plasterwork and carved doors in the Lamu region”, Art and Archaeology Research Papers 3, pp. 1–14. _____ . 1973b. “A further note on Swahili ornament”, Art and Archaeology Research Papers 4, pp. 87–92. _____ . 1974. “Swahili culture reconsidered: some historical implications of the material culture of the Northern Kenya Coast in the Eighteenth and Nineteenth centuries”, Azania 9, pp. 105–138. _____ . 1979. “The Swahili house: cultural and ritual concepts underlying its plan and structure”, Art and Archaeology Research Papers 10, pp. 1–32. _____ . 1982. “The Shirazi problem in East African Coastal History”, Paideuma 28, pp. 9–27. Alpers, Edward. 1976. “Gujarat and the trade of East Africa c. 1500–1800”, The International Journal of African Historical Studies 9–1, pp. 22–44. Anene, Joseph C. 1966. “The Omani Empire and its Impact on East African Societes”, in: id. & Godfrey N. Brown (eds), Africa in the nineteenth and twentieth Century, Ibadan, Ibadan University Press, pp. 440–457. De Barros, João. 1552 [1945]. Decadas Da Ásia, dos feitos qie os Portuguezes fizeram no descubrimento, e conquista dos mares, e terras do Oriente, Lisbonne, Ministério das colonias, 2 vols. Bandyopadhyay, Soumyen. 2008. “From Another World! A possible Bûyid origin of the decorated Mihrab of central Oman”, in: Eric Olijdam & Richard H. Spoor (eds), Intercultural relations between South and Southwest Asia, Oxford, Archaeopress, ser. “BAR International”, 1826, “Society for Arabian Studies Monographs”, 7, pp. 372–382. Bang, Anne. 2014. “The Ḥaḍramaut in Lamu. The manuscript collection of the Riyadha mosque of Lamu, Kenya”, in: Anne Regourd (guest ed.), Manuscripts of Yemen, circulation of ideas and models, special issue of Journal of Islamic Manuscripts 5/2, pp. 125–153. Baumanova, Monika & Ladislav Šmejda. 2018. “Space as material culture: residential stone buildings on the precolonial Swahili coast in comparative perspective”, South African Archaeological Bulletin 73/208, pp. 82–92. Ben Said Mohammed, Hashim. 2015. Les Sharifs dans l’histoire des Comores, Moroni, KomEdit. _____ . 2022. Exposé sur l’histoire de la vile de Mutsamudu. nCmY 19 (Juillet 2024) 166 S. Pradines & O. Onézime Swahili wall niches Blanchy, Sophie. 2015. “Anjouan (Comores), un nœud dans les réseaux de l’océan Indien. Émergence et rôle d’une société urbaine lettrée et marchande (XVIIeXXe siècle)”, Afriques 6, pp. 1–43. En ligne, https://journals.openedition.org/afriques/1817 _____ . 2022. “La maison urbaine, cadre de production du statut et du genre à Anjouan (Comores), XVIIe-XIXe siècles”, Afriques 13, pp. 1–57. En ligne, https://journals.openedition.org/afriques/3530 Bonnenfant, Paul. 2000. “La marque de l’Inde à Zabîd”, Chroniques yéménites 8, pp. 31– 57. _____ . 2008. Les maisons de Zabîd, Paris, Maisonneuve & Larose. Bourhane, Ahmed. 2006. “Mutsamudu et Anjouan aux temps des sultans”, TAREHI. Revue d’histoire et d’archéologie 14, pp. 23–34. Brown, W. Howard. 1985. “History of Siyu: the development and decline of a swahili town on the northern Kenya coast”, Ph.D., Bloomington, Indiana University. Campo, Juan. 2022. “‘This blessed place’: the talismanic significance of house inscriptions in Ottoman Cairo”, in: Marcia Probert & Petra Sijpesteijn (eds), Amulets and talismans of the Middle East and North Africa in context, Leiden, Brill, pp. 126–161. Charton, É. 1855. “Iles Comores”, Le magazin pittoresque, pp. 99–101, 131–133, 196–198, 259–261. Chaudhuri, Kirti Narayan. 1989. Trade and Civilization in the Indian Ocean. An Economic History from the Rise of Islam to 1750, Cambridge, Cambridge University Press. [1st ed. 1985]. Chittick, Neville. 1974. Kilwa: an Islamic Trading City on the East African Coast, Nairobi, The British Institute in East Africa, 2 vols. _____ . 1980. “Indian relations with East Africa before the arrival of the Portuguese”, Journal of the Royal Asiatic Society 2, pp. 117–127. Christie, Annalisa. 2019. “Structures and Settlement Organization at Kua Ruins, Juani (Mafia Archipelago, Tanzania)”, African Archaeological Review 36, pp. 249–269. Constantin, François. 1989. “Condition swahili et identité politique”, in: Jean-Pierre Chrétien & Gérard Prunier (dir.), Les ethnies ont une histoire, Paris, Karthala, pp. 337–355. Cooke, Edward. 2022. Global objects: toward a connected art history, Princeton, Princeton University Press. Costa, Paolo Maria, with a contribution by E. Baldissera. 2001. Historic Mosques and Shrines of Oman, Oxford, Archaeopress, ser. “BAR International”, 938. Das Gupta, Ashin. 1994. Indian Merchants, and the Decline of Surat c. 1700–1750, New Delhi, Manohar. Desai, Miki. 2023. “Traditional Bohra Dwellings of Gujarat, India: Architectural Response to Cultural Ethos”, in: Stéphane Pradines & Farouk Topan (eds), Muslim nCmY 19 (Juillet 2024) 167 S. Pradines & O. Onézime Swahili wall niches cultures of the Indian Ocean, Edinburgh, Edinburgh University Press, pp. 246– 270. Devic, Lucien-Marcel. 1975. Le Pays des Zendjs ou la Côte Orientale d’Afrique au MoyenAge, Amsterdam, Oriental Press. [1st ed. 1883]. Dongerkery, Kamala Sunderrao Kulkarni. 1973. Interior Decoration in India. Past and Present, Bombay, Taraporevala. Donley-Reid, Linda W. 1987. “Life in the swahili town house reveals the symbolic meaning of spaces and artefact assemblages”, African Archaeological Review 5, pp. 181–192. _____ . 1990a. “The power of Swahili porcelain, beads, and pottery”, in: Alice B. Kehoe & Sarah M. Nelson (eds), Power of observation: Alternative Views in Archaeology, Washington, American Anthropological Association, Archaeological Papers of the American Anthropological Association 2/1, pp. 47–59. _____ . 1990b. “A structuring structure: the Swahili house”, in: Susan Kent (ed.), Domestic architecture and the use of space, New Directions in Archaeology, Cambridge, Cambridge University Press, pp. 114–126. _____ . 1991. “Symbolic meaning within the traditional Hindu and muslim houses of Gujarat (India) and Lamu (Kenya)”, in: Brian Durrans & T. Richard Blurton (eds), The Cultural Heritage of the Indian Village, London, British Museum, ser. “British Museum London: Occasional Paper”, 47, pp. 75–89. Fleisher, Jeffrey. 2010. “Rituals of Consumption and the Politics of Feasting on the Eastern African Coast, AD 700–1500”, Journal of World Prehistory 23–24, pp. 195– 217. Freeman-Grenville, Greville Stewart Parker. 1962a. The East African Coast: Select documents from the first to the earlier nineteenth century, Oxford, Clarendon Press. _____ . 1962b. The Medieval History of the Coast of Tanganyika (with special reference to recent archaeological discoveries), London, Oxford University Press. _____ . 1965. The French at Kilwa Island. An Episode in Eighteenth-century East African History, Oxford, Clarendon Press. Garlake, Peter. 1966. The Early Islamic Architecture of the East African Coast, Nairobi, The British Institute in East Africa. Gensheimer, Tom. 2018a. “House as marketplace. Swahili merchant houses and their urban context in the later Middle Ages”, in: Patrick Haughey (ed.), A History of architecture and trade, London, Routledge, Chap. 5, pp. 115–127. _____ . 2018b. “Swahili houses”, in: Stephanie Wynne-Jones & Adria Laviolette (eds), The Swahili World, London, Routledge, pp. 500–511. Ghaidan, Usam. 1973. “Swahili plasterwork”, African Arts 6/2, pp. 46–49. _____ . 1975. Lamu. A study of the swahili town, Nairobi, East African Literature Bureau. _____ . 1976. Lamu. A study in conservation, Nairobi, East African Literature Bureau. Gray, John. 1954. “The Wadebuli and the Wadiba”, Tanzania Notes and Records 36, Dar es-Salaam, pp. 22–42. nCmY 19 (Juillet 2024) 168 S. Pradines & O. Onézime Swahili wall niches Guébourg, Jean-Louis. 1996. Espaces et pouvoirs en Grande Comore, Paris, L’Harmattan. Hardy-Guilbert, Claire & Guy Ducatez. 2004. “Al-Shihr, Yémen, porte du Hadramawt sur l’océan Indien”, Annales Islamogiques 38, pp. 95–157. Hebert, Jean-Claude. 1983. “Documents sur les razzias malgaches aux Îles Comores et sur la côte orientale africaine (1790-1820). 1ère partie : Les invasions à Mayotte et Anjouan jusqu’en 1807”, Études sur l’océan Indien 3, pp. 5–60. Hecht, Elisabeth-Dorothea. 1982. “The city of Harar and the traditional Harar house”, Journal of Ethiopian Studies 15, pp. 57–78. Hirji, Zulfikar. 2019. “The Siyu Qurʾans: Three Illuminated Qurʾan Manuscripts from Coastal East Africa”, in: id. (ed.), Approaches to the Qurʾan in Sub-Saharan Africa, Oxford, Oxford University Press, pp. 431–471. Hirschi, Susanne & Cherazade Nafa. 2014. Sultanats historiques des Comores. Recueil de relevés du patrimoine architectural et urbain, Rapport Unesco de l’École Nationale Supérieure d’Architecture et du Paysage de Lille. Ho, Eng Seng, 2006. The graves of Tarim: Genealogy and Mobility across the Indian Ocean, Berkeley, University of California Press. Horton, Mark. 1996. Shanga. The archaeology of a Muslim trading community on the coast of East Africa, Nairobi, The British Institute in East Africa. _____ & John Middleton. 2000. The Swahili, Oxford, Blackwell. Jama, Ahmed Dualeh. 1996. The Origins and Development of Mogadishu AD 1000 to 1850: A Study of the Urban Growth Along the Benadir Coast of Southern Somalia, Uppsala, Dept. of Archaeology, Uppsala University, ser. “Studies in African Archaeology”, 12. Juma, Abdurahman. 1996. “The Swahili and the Mediterranean worlds: pottery of the late Roman period from Zanzibar”, Antiquity 70/267, pp. 148–154. Kana’an, Ruba. 2008. “The carved stucco mihrab of Oman, forms, styles and influences”, in: Abdulrahman Al-Salimi, Heinz Gaube, Lorenz Korn & Faysal alHafiyan (eds), Islamic Art of Oman, Muscat, Mazoon Print., Pub. & Advertising, pp. pp. 230–259. Kervran, Monique. 1996. “Mihrab/s omanais du XVIe siècle : un curieux exemple de conservatisme de l’art du stuc iranien des époques seldjouqide et mongole”, Archéologie islamique 6, pp. 109–156. King, Geoffrey. 1998. The Traditional architecture of Saudi Arabia, London, Tauris. Kirkman, James. 1963. Gedi, the palace, La Hague, Mouton. Lambourn, Elizabeth. 2010. “A self-conscious art? Seeing micro-architecture in Sultanate South-Asia”, Muqarnas 27, pp. 121–156. Le Cour Grandmaison, Colette. 1991. “Parenté, migrations, alliances. Les réseaux omani en Afrique orientale et centrale”, in: Françoise Le Guennec-Coppens & Pat Caplan (dir.), Les Swahili entre Afrique et Arabie, Paris, CREDU/Karthala, pp. 163– 177. nCmY 19 (Juillet 2024) 169 S. Pradines & O. Onézime Swahili wall niches _____ . 1998. “L'héritage arabe, XVIIIe et XIXe siècle”, in: id. & Ariel Crozon (dir.), Zanzibar aujourd'hui, Paris, Karthala/Institut français de recherche en Afrique (IFRA), pp. 35–71. Le Guennec-Coppens, Françoise. 1991. “Qui épouse-t-on chez les Hadrami d’Afrique orientale ? Les réseaux d’alliances”, in: id. & Pat Caplan (eds), Les Swahili entre Afrique et Arabie, Paris, CREDU/Karthala, pp. 145–162. _____ & David Parkin (eds). 1998. Autorité et pouvoir chez les Swahili, Paris, Karthala Institut français de recherche en Afrique (IFRA). Lewcock, Ronald. 1976. “Architectural connections between Africa and parts of the Indian Ocean littoral”, Art and Archaeology Research Papers 9, pp. 13–23. Mantellini, Serge. 2019. “Investigating the core of the urban asset of the site: excavation of trench 9”, Sindh Antiquities Journal 5/2, pp. 70–87. Martin, Jean. 1983. Comores : quatre îles entre pirates et planteurs, Paris, L’Harmattan. Meier, Sandy Prita. 2009. “Objects on the edge. Swahili coast logics of display”, African Arts, pp. 8–23. _____ . 2015. “Chinese Porcelain and Muslim Port Cities: Mercantile Materiality in Coastal East Africa”, Art History. Journal of the Association of Art Historians 38/4, pp. 702–717. _____ . 2016. Swahili port cities: the architecture of elsewhere, Bloomington, Indiana university press. Mirza, Sana. 2023. “Developing the Harari Mushaf: The Indian Ocean milieu of Ethiopian scribes”, in: Stéphane Pradines & Farouk Topan (eds), Muslim Cultures of the Indian Ocean, Edinburgh, Edinburgh University Press, pp. 21–39. _____ . 2024. “The Qurʾan Illuminations”, coll. to Anne Regourd, Catalogue des manuscrits du Sherif Harar Municipal Museum. Les corans/Catalogue of the Manuscripts in the Sherif Harar Municipal Museum. The Qurʾans, Paris, Académie des Sciences d’Outre-Mer/Geuthner, ser. “Fontes Historiae Africanae/Sources africaines”, 5, pp. 121-152. Mitra, Asok. 1961. Report on house types and village settlement patterns in India, Census of India, New Delhi, Government of India Publications, vol. 1, Part IV-A (iii). Northedge, Alastair. 2004. “Abbasid Earth Architecture and Decoration at Samarra, Iraq”, in: Leslie Rainer & Angelyn Bass Rivera (eds), The Conservation of Decorated Surfaces on Earthen Architecture: Proceedings from the International Colloquium Organized by the Getty Conservation Institute and the National Park Service, Mesa Verde National Park, Colorado, USA, September 22–25, 2004, Los Angeles, Getty Conservation Institute, pp. 5–14. En ligne, http://hdl.handle.net/10020/gci_pubs/decorated_surfaces_earthen_arch _____ . 2005. The Historical Topography of Samarra, London, British school of archaeology in Iraq/Foundation Max van Berchem, ser. “Samarra Studies”, 1. Pankhurst, Richard Keir Pethick. 1982. History of Ethiopian towns from the Middle Ages to the early nineteenth century, Wiesbaden, Steiner. nCmY 19 (Juillet 2024) 170 S. Pradines & O. Onézime Swahili wall niches Patel, Alka. 2022. Iran to India: The Shansabanis of Afghanistan, c. 1145–1190, Edinburgh, Edinburgh University Press. Pearson, Michael. 1976. Merchants and Rulers in Gujarat. The response to the Portuguese in the sixteenth Century, Berkeley, University of California Press. Pouwels, Randall. 1974. “Tenth Century Settlement of the East African Coast: The Case for Qarmatian/Ismaʿili Connections”, Azania 9, pp. 65–74. _____ . 1984. “Oral Historiography and the Problem of the Shirâzi in East African Coastal History”, History in Africa 11, pp. 236–266. Pradines, Stéphane. 1999. “L’influence indienne dans l’architecture swahili”, Swahili Forum VI, AAP 60, Köln, Institut für Afrikanistik, pp. 103–120. _____ . 2000. “Rituels funéraires swahili : les sépultures islamiques du site de Gedi (Kenya)”, Les Recherches sur le Monde Arabe en France, 1er Colloque des Jeunes Arabisants, Toulouse, Université de Toulouse Le-Mirail, CEMAA-AMAM, pp. 167–193. _____ . 2003. “L’Afrique noire et la Chine. La céramique importée : symbole du pouvoir des marchands swahili”, La grande histoire de la porcelaine chinoise, Paris, Réunion des Musées Nationaux, pp. 35–41. _____ . 2003. “Le mihrab swahili : Evolution d’une architecture islamique en Afrique subsaharienne”, Annales Islamologiques 37, pp. 355–381. _____ . 2004. Fortifications et urbanisation en Afrique orientale, Oxford, Archaeopress, ser. “BAR International”, 1216, “Cambridge Monographs in African Archaeology”, 58, British Archaeological Reports. _____ & Pierre Blanchard. 2005. “Kilwa al-Mulûk. Premier bilan des travaux de conservation-restauration et des fouilles archéologiques dans la baie de Kilwa, Tanzanie”, Annales Islamologiques 39, pp. 25–80. _____ . 2009. “L’île de Sanjé ya Kati (Kilwa, Tanzanie). Un mythe Shirâzi bien reel”, Azania 44/1, pp. 49–73. _____ . 2010. Gedi, une cité portuaire swahilie. Islam médiéval en Afrique orientale, Monographies d’archéologie islamique, Le Caire, Institut Français d’Archéologie Orientale. _____ . 2019. “Islamic Archaeology in the Comoros: The Swahili and the Rock Crystal Trade with the Abbasid and Fatimid Caliphates”, Journal of Islamic Archaeology 6/1, pp. 109–134. _____ . 2020. “Kua, Tanzania, excavation report 2018”, Nyame Akuma 93, pp. 36–45. _____ . 2022. Historic mosques in sub-Saharan Africa, Leiden, Brill. _____ . Forthcoming. “Swahili domestic architecture, between warehouses and palaces”. _____ . Forthcoming. “The trees from the sea, coral stone architecture”. Ranasinghe, Shaalini. 2001. “The Castle of Emperor Fasilädäs: Missionaries, Muslims, and Architecture in Gondär, Ethiopia”, PhD thesis, Columbia University. nCmY 19 (Juillet 2024) 171 S. Pradines & O. Onézime Swahili wall niches Rhodes, Daniel, Colin Breen & Wes Forsythe. 2015. “Zanzibar: A Nineteenth-Century Landscape of the Omani Elite”, International Journal of Historical Archaeology 19/2, pp. 334–355. Rouaud, Alain. 1984. “L’émigration yéménite”, in: Joseph Chelhod (ed.), L’Arabie du sud, Paris, Maisonneuve Larose, vol. 2, pp. 227–250. Saba, Matt. 2015. “A restricted gaze: the ornament of the main caliphal palace of Samarra”, Muqarnas 32, pp. 155–195. Schlegelmilch, U. 1982. “Typologie und datierung der sogenannten Pillar tombs an der ostafrikanischen küste”, MA Munich, Ludwig-Maximilton Universität. Schulz, Vera-Simone. 2018. “Artistic Dynamics across the Seas”, in: Aksay Sarathi (ed.), Early Maritime Cultures. East Africa and the Western Indian Ocean, Papers from a conference held at the University of Wisconsin-Madison (African Studies Program) 23–24 October 2015, with additional contributions, Oxford, Archaeopress, pp. 185– 214. Siegel, Ulrike. 2017. “The architecture of the Abbasid residence at ar-Raqqa/ar-Rafiqa. Aspect of cultural interaction”, in: Lorenz Korn & Martina Müller-Wiener (eds), Central periphery?: art, culture and history of the medieval Jazira (Northern Mesopotamia, 8th–15th centuries), papers of the Conference held at the University of Bamberg, 31 October–2 November, 2012, Wiesbaden, Fritz Thyssen, ser. “Studies in Islamic art and architecture”, 2, pp. 71–96. Sheriff, Abdul. 1995. “Mosques, Merchants and landowners in Zanzibar Stone Town”, in: id. (ed.), The History and Conservation of Zanzibar Stone Town, Nairobi/Oxford, Eastern African Studies/James Currey, pp. 46–66. Sheriff, Abdul. 2001. “The historicity of the Shirazi tradition along the East African coast”, Historical Role of Iranians (Shirazis) in the East African Coast, Nairobi, Embassy of Iran, pp. 21–41. Siravo, Francesco & Ann Pulver. 1986. Planning Lamu: conservation of an East African Seaport, Nairobi, National Museum of Kenya. Stigand, Chauncy Hugh. 1913. The Land of Zinj, being an account of British East Africa, its ancient history and present inhabitants, London, Constable [reed. Frank Cass, 1966]. Tuchscherer, Michel. 2004. “Les échanges commerciaux entre les rives africaine et arabe de l’espace mer Rouge-golfe d’Aden aux seizième et dix-septième siècles”, in: Paul Lunde & Alexandra Porter (eds), Trade and Travel in the Red Sea Region, Oxford, Archaeopress, ser. “BAR international”, 1269, “British Foundation for the Study of Arabia Monographs”, 2, pp. 157–163. Um, Nancy. 2009. The merchant houses of Mocha: trade and architecture in an Indian Ocean port, Washington, University of Washington Press. Walsh, Martin. 2010. “Deep memories or symbolic statements? The Diba, Debuli and related traditions of the East African coast”, in: Chantal Radimilahy & Narivelo Rajaonarimanana (dir.), Civilisations des mondes insulaires : Mélanges en l’honneur du Professeur Claude Allibert, Paris, Karthala, pp. 453–476. nCmY 19 (Juillet 2024) 172 S. Pradines & O. Onézime Swahili wall niches Watson, Oliver. 2004. Ceramics from Islamic Lands: A Catalogue of the al-Sabah Collection, Kuwait, London, Thames & Hudson. Whitehouse, David, Donald Scott Whitcomb & Tony James Wilkinson. 2009. Siraf, history, topography and environment, Oxbow books, British institute of Persian studies. Wilding, Richard. 1976. “Harari domestic architecture”, Art and Archaeology Research Papers 9, pp. 31–37. _____ . 1988. “Panels, Pillars and Posterity: ancient tombs on the North Kenya coast”, Fort Jesus Occasional Papers 6. Wilson, Thomas H. 1979. “Swahili funerary architecture of the North Kenya coast”, Art and Archaeology Research Papers 10, pp. 33–46. Wynne-Jones, Stephanie. 2013. “The Public Life of the Swahili Stone house, 14th–15th nCmY 19 (Juillet 2024) 173 S. Pradines & O. Onézime Swahili wall niches Figures of Swahili Niches Fig. 01. Map the Indian Ocean and the Swahili coast. Credits: S. Pradines & O. Onézime. nCmY 19 (Juillet 2024) 174 S. Pradines & O. Onézime Swahili wall niches Fig. 02. Plan of the main 18th century palace of Mutsamudu, Nzwani Island, Comoros. Credits: O. Onézime. nCmY 19 (Juillet 2024) 175 S. Pradines & O. Onézime Swahili wall niches Fig. 03. Wall niches “Ziloho” in the Shandzahari, Ujumbe of Mutsamudu. Credits: S. Pradines & O. Onézime. nCmY 19 (Juillet 2024) 176 S. Pradines & O. Onézime Swahili wall niches Fig. 04. Talismanic painted ceiling of the Ujumbe of Mutsamudu. Credits: S. Pradines & O. Onézime. nCmY 19 (Juillet 2024) 177 S. Pradines & O. Onézime Swahili wall niches Fig. 05. Stucco wall niches in the ablution area of Shingani palace, Mutsamudu. Credits: S. Pradines & O. Onézime. nCmY 19 (Juillet 2024) 178 S. Pradines & O. Onézime Swahili wall niches Fig. 06. Wall niches “Ziloho” in the door jambs, ground floor, Ujumbe of Mutsamudu. Credits: S. Pradines & O. Onézime. nCmY 19 (Juillet 2024) 179 S. Pradines & O. Onézime Swahili wall niches Fig. 07. Large panel of stucco niches of the ground floor, room 4, Ujumbe of Mutsamudu. Credits: S. Pradines & O. Onézime. nCmY 19 (Juillet 2024) 180 S. Pradines & O. Onézime Swahili wall niches Fig. 08. Wall niches “Ziloho” in the Ndani of the Ujumbe of Mutsamudu. Credits: S. Pradines & O. Onézime. nCmY 19 (Juillet 2024) 181 S. Pradines & O. Onézime Swahili wall niches Fig. 09. Large panel of stucco niches of the first floor, in room 16, Ujumbe of Mutsamudu. Credits: S. Pradines & O. Onézime. nCmY 19 (Juillet 2024) 182 S. Pradines & O. Onézime Swahili wall niches Fig. 10. Different types of stucco niches of the Ujumbe of Mutsamudu. Credits: S. Pradines & O. Onézime. nCmY 19 (Juillet 2024) 183 S. Pradines & O. Onézime Swahili wall niches Fig. 11. Flat-bottomed niches and decorative blind niches, Ujumbe of Mutsamudu. Credits: S. Pradines & O. Onézime. Fig. 12. Niches in the courtyard of the Ujumbe of Mutsamudu. Credits: S. Pradines & O. Onézime. nCmY 19 (Juillet 2024) 184 S. Pradines & O. Onézime Swahili wall niches Fig. 13. Stucco niches in Koyama, Bajuni Archipelago, Somalia. Credits: Diana Powell-Cotton, 1930s, Powell-Cotton Museum PHOTO.2.22.78.2. nCmY 19 (Juillet 2024) 185 S. Pradines & O. Onézime Swahili wall niches Fig. 14. Late 16th–early 17th century house in Kua, Juani Island, Mafia archipelago, Tanzania. Credits: S. Pradines & Zamani Project. nCmY 19 (Juillet 2024) 186 S. Pradines & O. Onézime Swahili wall niches Fig. 15. Late 16th–early 17th century single niches in house, Kua, Juani Island. Credits: S. Pradines & Zamani Project. nCmY 19 (Juillet 2024) 187 S. Pradines & O. Onézime Swahili wall niches Fig. 16. Panel of stucco wall niches in House 3, Kua, late 17th century. Credits: S. Pradines. nCmY 19 (Juillet 2024) 188 S. Pradines & O. Onézime Swahili wall niches Fig. 17. 15th century niches in the palace of Songo Mnara, Tanzania. Credits: S. Pradines & O. Onézime. nCmY 19 (Juillet 2024) 189 S. Pradines & O. Onézime nCmY 19 (Juillet 2024) Swahili wall niches 190 S. Pradines & O. Onézime Swahili wall niches Fig. 18. 15th century niches in houses HO31 and HO5. Gedi, Kenya. Credits: S. Pradines & O. Onézime. nCmY 19 (Juillet 2024) 191 S. Pradines & O. Onézime Swahili wall niches Fig. 19. Small niches above the main entrance to the palace of Gedi, 15th century. Credits: S. Pradines. nCmY 19 (Juillet 2024) 192 S. Pradines & O. Onézime Swahili wall niches Fig. 20. Reception room with niches facing the main courtyard of the palace in Gedi. Credits: S. Pradines & O. Onézime. nCmY 19 (Juillet 2024) 193 S. Pradines & O. Onézime Swahili wall niches Fig. 21. Wall niches in Husuni Kubwa, 1320–1333 AD., Kilwa Kisiwani, Tanzania. Credits: S. Pradines & O. Onézime. nCmY 19 (Juillet 2024) 194 S. Pradines & O. Onézime Swahili wall niches Fig. 22. Coral stone house in Sanje ya Kati, Kilwa archipelago, 1050–1150 AD. Credits: S. Pradines. Fig. 23. Panelled tomb with niches in Mwana, Northern Kenya, late 14th century. Credits: S. Pradines. nCmY 19 (Juillet 2024) 195 S. Pradines & O. Onézime Swahili wall niches Fig. 24. 14th and 15th century Pilar tombs in Gedi. Credits: S. Pradines & O. Onézime. nCmY 19 (Juillet 2024) 196 S. Pradines & O. Onézime nCmY 19 (Juillet 2024) Swahili wall niches 197 S. Pradines & O. Onézime Swahili wall niches Fig. 25. 15th century Pilar tomb Gedi with niches and ceramic inlays. Credits: S. Pradines & O. Onézime. nCmY 19 (Juillet 2024) 198 S. Pradines & O. Onézime Swahili wall niches Fig. 26. 18th century Pilar tomb in Kunduchi, Tanzania. Credits: S. Pradines. nCmY 19 (Juillet 2024) 199 S. Pradines & O. Onézime Swahili wall niches Fig. 27. Chinese blue and white porcelains in Mambrui’s pilar tomb, Kenya. Credits: S. Pradines. nCmY 19 (Juillet 2024) 200 S. Pradines & O. Onézime Swahili wall niches Fig. 28. 17th century pilar tomb in Kua, Juani Island, Mafia archipelago. Credits: S. Pradines & Zamani Project. Fig. 29. 17th–18th century Mausoleums in Tsingoni. Credits: S. Pradines. nCmY 19 (Juillet 2024) 201 S. Pradines & O. Onézime Swahili wall niches Fig. 30. 17th century mausoleum of Mwana Hadiye Msingi in Siyu, Pate Island. Credits: Mohamed Hassan Ali, NMK, Lamu. nCmY 19 (Juillet 2024) 202 S. Pradines & O. Onézime Swahili wall niches Fig. 31. Single niche in the great mosque of Kilwa, 12th–13th century. Credits: S. Pradines. nCmY 19 (Juillet 2024) 203 S. Pradines & O. Onézime Swahili wall niches Fig. 32. Small niches for incense burners or oil lamps, Mnarani of Kilifi, Kenya. Credits: S. Pradines. nCmY 19 (Juillet 2024) 204 S. Pradines & O. Onézime Swahili wall niches Fig. 33. Wall niches on the frame of mihrab, both side of the opening of the niche of the mosque n° 2 of Kaole, Tanzania. Credits: S. Pradines. nCmY 19 (Juillet 2024) 205 S. Pradines & O. Onézime Swahili wall niches Fig. 34. Frieze of small niches in the mihrab of the great mosque of Songo Mnara 15th century. Credits: S. Pradines. nCmY 19 (Juillet 2024) 206 S. Pradines & O. Onézime Swahili wall niches Fig. 35. Persian turquoise blue glaze on black silhouettes and Chinese blue and white porcelains bowls in the late 14th century mihrab of Kua. Credits: S. Pradines. nCmY 19 (Juillet 2024) 207 S. Pradines & O. Onézime nCmY 19 (Juillet 2024) Swahili wall niches 208 S. Pradines & O. Onézime Swahili wall niches Fig. 36. Blue and white porcelain inserted in tympanums and intrados of the mihrab of the ǧamīʿa of Gedi, Kenya, 1425–1450 AD. Credits: S. Pradines & O. Onézime. nCmY 19 (Juillet 2024) 209 S. Pradines & O. Onézime Swahili wall niches Fig. 37. Fifty ceramic plate inlays in the Mihrab of Tundwa. Credits: S. Pradines & I. Almela. nCmY 19 (Juillet 2024) 210 S. Pradines & O. Onézime Swahili wall niches Fig. 38. Persian bowls inlays in the vaults of the small domed mosque in Kilwa Kisiwani. Credits: S. Pradines & O. Onézime. nCmY 19 (Juillet 2024) 211 S. Pradines & O. Onézime Swahili wall niches Fig. 39. Example of 13th century Chinese celadon inserted or exhibited in Swahili buildings in Gedi, Kenya. Credits: S. Pradines. nCmY 19 (Juillet 2024) 212 S. Pradines & O. Onézime Swahili wall niches Fig. 40. Porcelains plates displayed in niches of the Ujumbe during a meeting with the Sultan of Anjouan, 1846. Source: Janet-Lange in Charton, 1855, p. 101. nCmY 19 (Juillet 2024) 213 S. Pradines & O. Onézime Swahili wall niches Fig. 41. Lamu house’s walls decorated with Chinese porcelains and weapons, 1884 AD. Source: Sir John Kirk Archives, reproduced with the permission of the National Library of Scotland. nCmY 19 (Juillet 2024) 214 S. Pradines & O. Onézime Swahili wall niches Fig. 42. Niches called “Taqet” in Harari house Gidir Gar, Ethiopia. Credits: O. Onézime. nCmY 19 (Juillet 2024) 215 S. Pradines & O. Onézime Swahili wall niches Fig. 43. Perfume bottles in flat carved niches probably from the Chini Khana, Gate pavilion or Suraj Bhan ka Bagh, Agra, Sikandra, circa 1620 AD. Credits: Wikimedia Commons, India, Musée Guimet, Paris, inventory number: MA 6775. nCmY 19 (Juillet 2024) 216 S. Pradines & O. Onézime Swahili wall niches Fig. 44. Mughal niches in Shalimar gardens, Lahore. Credits: S. Pradines. Fig. 45. Abbasid stucco niches, palace of Balkuwara, Sāmarrāʾ (Iraq) 1911–1913 by Ernst Herzfeld, Pergamon Museum, Berlin. Credits: Collection Ernst Herzfeld Papers. FSA A.06 04. PF.23.096. Freer Gallery of Art and Arthur M. Sackler Gallery Archives. Gift of Ernst Herzfeld, 1946. nCmY 19 (Juillet 2024) 218 S. Pradines & O. Onézime Swahili wall niches Fig. 46. Miniature showing the house of wisdom “Beit al-Hekmeh”, the Abbasid library of Baghdad and books stored in niches. Credits: Yaḥyā b. Maḥmūd b. Yaḥyā b. Abū al-Ḥasan al-Wāsiṭī in Abū Muḥammad al-Qāsim b. ʿAlī b. Muḥammad b. ʿAlī al-Ḥarīrī al-Baṣrī, Al-maqāmāt, 1237 AD, Bibliothèque nationale de France, ms. Ar. 5847. Public domain. nCmY 19 (Juillet 2024) 219