In 1987, the government of the Maldives first tried to propose a list of 21 historical mosques for nomination on the UNESCO World Heritage list. Then the project was resurrected in 2008 with a reduced list to six monuments selected for the current listing. In January 2017, the UNESCO, the Department of Heritage of the Maldives and the World Monument Fund organised a workshop on the Coral Mosques of the Maldives islands. In November 2017 we organised an archaeological project in the Maldives archipelago with the Aga Khan University in association with the Heritage Department of the Maldives along with the Maldives National Defence Force (MNDF) with the support of the Max van Berchem Foundation and the Unesco New Delhi. Our main goal was to determine scientifically the chronology of some Maldivian coral mosques. This objective responded to our previous workshop.


Trading routes and Islam in the Maldives

Islam was probably introduced to the Maldives by merchants from the Malabar cost in the 7-8th centuries (Forbes, 1983: 47).  The Maldives Islands are mentioned around 850 AD in the Akhbar al-Sin wa l-Hind ca.  They are known as Dîbagât. Abu Musa al-Sirafi mentioned the islands in 980 and in 1154 Al-Idrisi spoke of Dabîgât but later the islands are always called Dîbagât in the Arabic sources (Ducène, 2013: 132-33). It’s only in the 12th century that Islam became the main religion of the Maldives. According to tradition, Maldivians converted to Islam in 1153. But the earliest Arabic script known dates back to 1337. And the process of conversion to Islam might have been much longer than what is described in the local chronicles, for example in 1652 some firmans were promulgated to tax non-Muslims (Forbes, 1983: 48).

Two primary sources date the conversion of the last Buddhist raja in 1153 AD.  The first story, related by Ibn Battuta, mentioned that the first Sultan of the Maldives converted to Islam after having seen the miracles of the magrebi preacher, Yûsuf Abû al-Barakât al-Barbarî (Kalus & Guillot, 2005: 35-36). The people of Male broke the idols and razed the temples to the ground and adopted Islam. The second story can be found in the Maldivian chronicle a Tarik from the 18th c. that mentions a conversion of the Raja following a Persian alim from Tabriz and took the title of Sultan in 1141 AD. The new Sultan sent messengers to convert all the people of the Maldivian islands and built mosques on each island.

The Tarik or Sultans’ chronicles present the genealogy of the Maldivian Sultans from 1141 to 1821. According to these local chronicles written in the 18th century, the conversion to Islam is credited to a Turkish sufi Yûsuf Shams al-Dîn al-Tabrîzî.  According to Kalus & Guillot (2005: 41), it’s still the same Yûsuf but with different nisba, so most probably the same man. During the 15th-16th century the story changed to accommodate the influence of the Rûmî school and new Muslim influences. The nisba of Yûsuf was changed for al-Tabrizi.

According to the Sultans’ chronicles, Hassan Shirazi arrived in the Maldives in 1552 and the dynasty ruled the archipelago until the 18th century. This tradition is very similar to late Shirazi traditions on the East African coast according to local traditions in the Comoros, a Shirazi ruler arrived in Mayotte in the 16th century and funded the mosque of Tsingoni. This late migration has to be connected to the eviction of the Shirazis rulers from Kilwa in the early 13th century, they first moved to Anjouan to move later to Mayotte. It’s possible that these Shirazis move eastward up to Madagascar and to the Maldives (Fig. 1).

mini Pradines figure 01

 Figure 1: The Maldives and the Indian Ocean roads

Mosques and Buddhism

The Maldives archipelago is made up of 1196 islands and 200 are inhabited. The Maldives have 180-200 khutba mosques or Friday mosques so it’s practically one mosque per island (Forbes, 1983: 67-68 and Reynolds, 1984). According to tradition and architectural observations, Maldivian mosques were possibly built on Buddhists buildings. The Buddhist shrines were destroyed and adapted for Islamic use. Bell mentioned a pirivena or monks residence destroyed and converted to a mosque during the 12th century on Gan Island in Addu Seenu Atoll. The process of adapting Buddhist religious buildings for use as mosques occurred widely all over the Maldives.

The first mention in primary sources of the Maldivian mosques can be found in the Rihla of Battuta (1343-44). Battuta didn’t describe the mosques but explained that the people raised their buildings above the ground and they have great skills in dressing the stones (Forbes, 1983: 51). The first accurate description of the Maldivian mosques dates from 1529 with the French traveller Parmentier in Fua Mulaku Atoll. Parmentier gave a very good description of a stone coral mosque and its painted ceiling. He describes beautiful constructions made without mortar (Forbes, 1983: 52). Later another French sailor Pyrard de Laval, in 1602-07, described the mosques in Malé, in a large square surrounded by tombs and wells with steps.  The carpentry was admired and described with precision (Forbes, 1983: 54).

Excavations should explain and reveal the chronology of these complicated buildings. An excavation in the prayer room would allow us to find the square or circular structure built on the platform and demonstrate if we have the foundation of previous buildings from the Buddhist period.


The excavations

Two sites were selected to explain and reveal the chronology of the mosques of the Maldives.

The outcomes expected from our mission were first, to create the first field school of Islamic archaeology of the Maldives, the idea is to introduce and to train the staff of the Heritage department to this specific discipline. Second to provide new scientific data on the coral mosques of the Maldives to facilitate the nomination of the mosques on the World heritage list of the UNESCO. Our objectives were first archaeological and scientific, but also for documentation and conservation purposes to serve the needs of the Maldivian government and international organisations such as the UNESCO and the World Monument Funds. We organised archaeological test pits to provide some dates on these buildings that were never dated scientifically previously and to see if some structures pre-date the mosques and Muslim tombs. One major question was the continuity of the settlements or influences of pre-Islamic cultures such as Buddhist architecture on local Islamic architecture. We worked during one month last November (2017) with one French archaeologist, 3 members from the Heritage department and 3 surveyors form the Maldivian defence forces.


Excavation at Koagannu Cemetery

First, the site of Koagannu Cemetery’s in Hulhumeedhoo (Fandiyaaru) on Addu Atoll which is located in the most southern part of the Maldives (Fig. 2). The excavations took place from 3-13rd November 2017. More specifically, the fieldwork was conducted in Meedhoo city within the Koagannu Cemetery’s enclosure which, according to tradition, is said to be a major archaeological site of the Maldives very well known for its tombs and mosques. But unfortunately the site is not yet listed in the proposal for the UNESCO world heritage list. The Fandiyaru mosque was supposed to have been funded in 1586 and the mosque is considered as the earliest coral stone mosque in the country. A major specificity of this site is the size of the compound, the largest in the Maldives with three other small mosques, fifteen mausoleums and five hundred coral tomb stones. This important Maldivian site to the southern atoll was the perfect place to explore the uniqueness of the ancient Maldivian mosques. This site that is not on the tentative heritage list: Hulhumeedhoo, despite the fact that this site has the largest cemetery in the Maldives. It is considered as the oldest and largest cemetery in the Maldives with 500 coral stone tombs and 15 mausoleums. The funerary compound includes four mosques. Six were built originally but only four remain: Koagannu mosque (1397 AD.), Boadha mosque (1403 AD.), Athara mosque (1417 AD.) and Fandiyaru mosque (1586 AD.) (Fig. 3).

mini Pradines Figure 02

Figure 2: Photo of the compound of Koagannu Cemetery’s in Hulhumeedhoo island.

mini Pradines Figure 03

Figure 3: Plan of the compound of Koagannu Cemetery’s in Hulhumeedhoo island.


After an initial visual identification of the site, it has been decided that three test pits (1, 2 & 3) would be realized along two of the four mosques (Mosque n° 1 Fandiyaaru Miskiy, test pit 2 and 3; and Mosque n° 3 Bodha Miskiy, test pit 1) due to their locations. Likewise, the nearby space available has been taken into account to ensure that everything runs smoothly around the visible groups of burials; equally with local traditions when they were assumed to be pertinent and susceptible to help onto the archaeological investigation. Each Test pit has been recorded as a unique component where its numeral name has been used to write down the first stratigraphic unit by employing the hundreds digit as follows: Test pit 1 (US100); Test pit 2; (US 201) and Test pit 3 (US 300) (Fig. 4).

mini Pradines Figure 04

Figure 4: Photo of the Test pit 1 showing its relation with the Bodha Mosque

In our current state of knowledge, five phases of occupation has been distinguished before the 14th-15th century until nowadays with a possible sub-division inside the third phase (Figs. 5, 6, 7 and 8).

mini Pradines Figure 05

Figure 5: Photo of the Fandiyaaru Mosque

mini Pradines Figure 06

Figure 6: Plan and north section of Test pit 2 showing its relation with Fandiyaaru Mosque

mini Pradines Figure 07

Figure 7: Photo of Test pit 3 and its relation with Fandiyaaru Mosque

mini Pradines Figure 08

Figure 8: South section of Test pit 3 and its relation with the Fandiyaaru Mosque's basement

Phase 1

So far, it appears that the site at least was strongly occupied a few centuries before the construction of Fandiyaaru Miskiy. The 30 cm thick organic layer 210=308 shows important human activities through the huge amount of very tiny charcoal inclusions, the occurrence of four seashell species within it as well as the possible coral wall structure 307. Although we cannot be sure whether this phase must be related to a pre or early Islamic period, one should rather stand to a generic denomination when speaking of Phase 1, i.e. the pre-Mosque period. However, there is no doubt that, whoever those people were and came from, they arrived on a virgin land exempt of any earlier civilization(s) for they left behind an occupancy layer directly upon the sandy substrate (211). Probably those latter ones must be dated back to before the 15th-16th century, which is the approximate date of the next phase. Further researches should furnish much more archaeological information and improve the date of that first phase.

Phase 2

The hardest phase to be recognized is perhaps the second one because of the radical soil change that was noticed in both the Test pit 2 and 3: we leave a very organic layer (210=308) to a much less or even non-organic one (209=306), almost sterile of material. The question is to know whether we here assist to an abandonment of the site, a human activity change or just a levelled ground event. Yet, it is unlikely that it is about the Fandiyaaru Miskiy since its basement cut context 303 that is overlying the layer referred at issue above (209=306). The presence of one Indian carinated rim pottery of orange colour suggests attaching this Phase 2 around the 15th/16th century. Moreover, even though the Test pit 1 did not provide such a similar context, it is tempting to associate the deposit 106 to this same chronological phase as it contained a piece of pottery that might be linked to an identical typological dating. This second phase is seemingly ending with the deposit of context 205 in front of the Fandiyaaru Miskiy’s main entrance and context 303 in its north-western part. And, eventually, the three sub-circular pits (206, 207 & 208) closed definitely that Phase 2.

Phase 3a

This is during the Phase 3a that one can observe the construction of the Fandiyaaru Miskiy within the 16th century regarding on the previous observations. The archaeological survey allowed the understanding of the architectural chronology: the manufacture seems easy but well thought so that both the dismantlement and the rebuilding of the monument could be easily undertaken elsewhere. Besides the fact that the rectangular basement 305 (roughly 4,04 x 3,68 x 0,20 m) was only comprised of one single course of sand stone boulders, the whole structure was, somehow, supported by another “substructure”. It is actually a rectangular coral podium of 19 cm high which presents four moulded steps: the lowest and largest one forms a 3,83 x 3,34 x 0,08 m rectangle whereas the upmost and smallest one by 3,21 x  2,70 x 0,03 m. Above it, the first closed wall (3,21 x 2,70 x 0,09 m), whose only 29 cm high is visible, had been added before the shaped plinth of 14 cm high that snaps into place over it and acts as an interface with the further architectural elements and the current white floor tiles (5 cm thick). The façade walls are comprised of five courses joined together like a puzzle without any binding agent at all and go up to a height of 1,78 m. The upmost part of the structure is made up of wooden intertwined elements and so is the framework that supports a modern iron roof. Moreover, the presence of a well and a pathway are associated to the construction of the mosque but it is tough to tell if the footpath has been used to connect the mosque up to the door entrance.

Phase 3b

A spreading out of the prayer room outside the mosque shall be considered for the Phase 3b since two parallel walls (204), 10,08 m of distance from one another and 1,28 m away from the eastern side of the mosque, were found cutting context 205 that approximate belongs to the 16th century. This plausible wideness of the sacred place may suggest an important attraction  and, hence, a development of the site in such a way that people felt the need to increase the prayer area and host as much worshipers as possible. Shall we also recognize the construction of the unique minaret to this phase according to the high level of its basement (5 cm under top soil) or to the Phase 3a. In the latter, we may thus state the moving of the architectural element in a later period mostly if one refers to the highest tomb that provides a similar observation and we do know that it has been moved at least once in the past, maybe twice.

Phase 4

The digging of the grave 107 behind the qibla of the Bodha Miskiy seems to be the only trace of the human presence at the site and the marker of the fourth phase. The burial may date to the 17th/18th century according to the materials found upon its surface layer 105, in particular the three fragments of a porcelain plate. Its shape and decoration remind the British porcelain that used to be produced for a commercial exportation towards China during that time; especially when a fake Chinese sign happens to be painted on the internal surface. A similar complete object was observed in the archaeological museum’s collection of Malé.

Although the excavation did not provide further indication about this phase, there is no doubt that other graves must be dated to the same period such as the highest coral tomb and probably a few more. Therefore, a first inventory of the ancient visible burials has been made by counting the tombstones. Three groups have emerged according to three space organization patterns regardless the gender and age distinctions: these comprised inside a full compound or an individual pen, these partially delimited by a compound and these following a wall compound, namely being in the same alignment of it. Among a minimum of 650 individuals noticed, it has resulted that 223 belong to the first category, 282 to the second and, finally, 145 to the third one. It’s probable that the stelae placed in the same alignment than a wall compound might not be in situ. This is because numerous wall and podium parts have been reused either to enclose partially or entirely the tombs. The arrangement of the stelae could have been used to refurbishing the cemetery but the exact purpose still remains unknown except if that was for gaining more space in order to host more people.

Phase 5

This last phase is well characterized by the construction of a modern porch expansion 202 that encompasses three sides of the Fandiyaaru Miskiy. Even though it is not proved archaeologically, it is probable that these recent buildings present all around the four mosques might have been built at about the same chronological period due to their architectural design. Finally, even if the mosques do not seem gathering as much believers as they used to, there is no doubt that a certain attraction remains at the site nowadays, carried by local traditions surrounding it or by looking at it. Indeed, apart from the customs starring the first settlers, the site centre gives the impression to retain people with a particular care. This concern is well marked by the existence of the current level of circulation (101=201=301) which is levelled with small leached broken whitish-grey coral stems.


Excavation at Fenfushi Old Friday Mosque (Ari Atoll)

The second site was the Friday mosque of Fenfushi on Alifu Dhaalu Atoll was renovated between 1692-1701 CE during the reign of Sultan Mohamed of Dhevvadhu on the site of an earlier mosque built by Kallhukamanaa and still in use today. The mosque is surrounded by pre Islamic (Buddhist) structure with a huge bathing pool. The small island of Fenfushi has very old oral traditions and an old cemetery (Figs. 9-10). The craft lacquer and turn wood traditions originates from this island (Figs. 11-12). Our mission in Fenfushi had also another objective to implement the map created previously and to prepare some surveys to see the relation between Buddhist and pre-Islamic settlements and Islamic settlements. The excavations run from 16-20th November 2017 within the Old Friday Mosque’s enclosure, known as Aasaary Miskiy, which is located in the north-western part of Fenfushi Island, Ari Atoll. Only two Test pits (1 & 2) were possible due to the limited time and excavated with the same recording system as previously.

 mini Pradines Figure 09

Figure 9: Aasaary Mosque compound on Fenfushi island      

mini Pradines Figure 10

Figure 10: So-called Budhist bathing tank on Fenfushi island

mini Pradines Figure 11

Figure 11: Aasaary Mosque prayer room and lacquer works

mini Pradines Figure 12

Figure 12: Aasaary Mosque and the wooden lacquered mihrab

The excavations carried out at the Aasaary Cemetery enabled to determine three important phases that occurred within the site but the absolute chronological periods could not be established entirely due to the lack of datable material. According to one of the eldest people of the island the ancient bathing tank has been given to the community by Sultan Mohamed Dhevvadhu (1692-1701) as a gift resulting of the dryness that had occurred in the region during his time. This hence suggests its non-relationship with the mosque. The rediscovery of a very similar modern structure, built in the 1980s, with the purpose of bringing fresh water to the islanders may confirm these comments (Figs. 13-14). Consequently, the ancient bathing tank could not be related to the Buddhism era of the Maldives and should be considered as the first water tank ever built on the island. The groundwater and rainfall thus permitted to fill both water containers even though some wells surrounded the mosque.

mini Pradines Figure 13

Figure 13: Plan of the modern and ancient water tanks near Aasaary Mosque

mini Pradines Figure 14

Figure 14: Photos of the modern and ancient water tanks near Aasaary Mosque

Furthermore, it was only known by tradition that the current mosque had also been built during the reign of Sultan Mohamed Dhevvadhu above an earlier mosque which was said to have been constructed from coral stones but without any carvings by one of the former royal couples of the region, Kalhukamanafaanu and her husband. The excavation of the Test pit 2 has demonstrated the assertion of this older mosque through the rediscovery of its foundation trench filled by two courses of beach slabs and boulders which formed its basement. Whereas the date of its construction could not be given because of the lack of datable material, it was shown that the mosque had been destroyed ca. 1538-1635, meaning about a century and a half before Mohamed Dhevvadhu’s reign. This tends to hypothesize that the Sultan did not demolish the first mosque but rather wanted to re-build the sacred place for the community. It seemed that he was very turned towards to the well-being of his community mostly when combining his willing for providing fresh water thanks to the construction of the water tank. Nevertheless, it could be considered that the Sultan may have wished to reconnect his family with the first royal couple whose tombs are visible on the southern side of the Old Friday Mosque. Those enigmatic rulers were Al-Wazeerul Haaju Hasssan his beloved wife who was believed to be the king’s daughter of Malé.

However, one can wonder why it took that long before seeing a new mosque. Again, according to tradition, the above cited rulers might also have built Kuda Miskiy –“Small Mosque”– which is located to the southwest of the Aasaary Cemetery in another boundary where, following an official document –Faiykolhu–, Kalhukamanaa’s son, Mohamed Mathukkalaa, was buried in the cemetery around the Small Mosque. If so, it shall be considered that the sacred area there has been using instead, contrary to the current situation nowadays.

To conclude, the excavation allowed determining not only the true function of the so-called “bathing tank” and the similar modern structure 102/103 but also demonstrating scientifically the relative chronology of the standing mosque’s construction and the existence of an earlier building underneath. The investigation showed that both structures have been separated by a destruction layer which was possible to date more or less precisely. Thus, the site had three main phases: the building of the first mosque, its destruction ca. 1538-1635, the rebuilding of the mosque around Mohamed’s reign (Figs. 15-16), the construction of the modern water tank ca. 1980. In the end, the protective structures surrounding the ancient water tank and the mosque might be added as the final and fourth phase dated to the 21st century having occurred at the site.

mini Pradines Figure 15

Figure 15: Plan of Test pit 2 and its relation with Aasaary Mosque

mini Pradines Figure 16

Figure 16: West and East sections of Test pit 2 and its relation with Aasaary Mosque's basement


Conclusion: continuity or influences of Buddhist architecture on local Islamic architecture 

We prepared the first plan of the Meedhoo (Fandiyaaru) compound. Our plan includes the four mosques and the whole cemetery. We did three test pits around two major mosques within the compounds. Our major discovery was to contradict the stated fact that the site was built on a previous Buddhist site. We didn’t find any evidence of this. Despite the indisputable Buddhist occupation on the island, our excavations revealed that the Islamic sites (mosques and cemeteries) were not built on previous pre-Islamic buildings. Our excavations also revealed that the site is not as old as it was described in the oral tradition. The site is not from the 12th century but likely from the 16th-17th centuries. Our plan and the full results of our archaeological investigations will be sent to the Maldivian Heritage Department and UNESCO in order to add this site to the list already prepared and to include the Southern Atoll to the World Heritage List. We also found another small mosque to the south of the island. This mosque is identical to the mosque in the large compound and should be also protected.

In Fenfushi, we implemented the map previously created. This implementation was a major advance because we added another old mosque and a further modern mosque, including another cemetery. This plan will be extremely important in the future as it considerably enlarges the zone that needs to be protected. During our excavation of the old mosque, we discovered an old platform of a previous mosque built in sand stone, and not in coral. We couldn’t date this mosque but we know now that it was built before the 16th century as we did a radiocarbon dating on charcoals (Beta-analytic) in the destruction layer previous to the construction of the second mosque. This date correspond well to the local tradition and the epigraphy of the cemetery and the second mosque. Our test pits confirm also that there were no pre-Islamic settlement or structure under the old mosque of Fenfushi. This mosque is similar to the old mosque found in the bush to the north of the main site (on the new map). We excavated a recent and modern water tank (cistern) and this excavation helped us to reinterpret the so-called “old bathing tank”. We found absolutely no evidence to support the interpretation as a pre-Islamic Buddhist bathing tank. Our investigations suggest that the bathing tank was in fact a water tank built by the sultan for the local community.

Our two excavations to the north and south of the Maldivian atolls bring very important information.  We were expecting to find Buddhist temples structures under these old Maldivian mosques but we found nothing. Some test pits were done in the great mosque in Malé by my colleague Mauroof Jamel but they found nothing. So at least for us, the two sites that we investigated the mosques were not built on Buddhists temples. It’s the story of a non-result but it helps to have a better understanding of the relation between tradition, history and archaeology and to propose a new chronology of the Maldivian mosques and their relation with previous religions and architectures. 

Professor Stephane Pradines
Institute for the Study of Muslim Civilisations
Aga Khan University



Aga Khan university (AKU); Max van Berchem Fondation; UNESCO New Delhi; Department of Heritage of the Maldives ; Military Engineers, Maldives National Defence Force (MNDF) and the World Monument Fund (WMF)



Bell, H.C.P., "The Maldives Islands; Monograph on the History, Archaeology, and Epigraphy", lère éd. Colombo, Ceylon Government Press, 1940; rééd. Malé, NCLHR, 1985; rééd. Malé, Novelty Printers & Publishers Pvt. Ltd., 2002. 
Carswell, John, “Mosques and Tombs in the Maldives Islands”, in Art and archaeology research papers 9, 1976, pp. 26-30. 
Forbes, Andrew, "The Mosque in the Maldive Islands: A Preliminary Historical Survey". In: Archipel, volume 26, 1983. pp. 43-74. 
Forbes, Andrew D.W., "The Pre-Islamic Archaeology of the Maldives Islands". In: Bulletin de l'Ecole française d'Extrême-Orient. Tome 76, 1987. pp. 281-288. 
Jameel, Mauroof, "Coral Stone Mosques of Maldives. The Vanishing Legacy of the Indian Ocean", Gulf Pacific Press, Los Angeles, 2016, 272 p. 
Kalus Ludvik, Guillot Claude. "Inscriptions islamiques en arabe de l'archipel des Maldives". In: Archipel, volume 70, 2005. pp. 15-52. 
Reynolds, C., “The mosques in the Maldives islands: further notes”, in Archipel 28, 1984, pp. 61-64.