8
The al-Khamis mosque on Bahrain: A note on the first and second phases DAVID WHITEHOUSE Corning, New York The ruined al-Khamis Mosque is situated near the north end of Bahrain, 10 kilometres from Manama. The building was partly excavated in 1986 by the Mission Arche ´o- logique Franc ¸aise, under the direction of Monik Kervran, who distinguished three phases of construction (1). This note is a comment on the remains and interpretation of the first and second phases. It does not address the question of the dates of these phases, for which very little evidence exists. For the sake of convenience, the qibla wall is described as facing west (in reality, the qibla is a few degrees south of west); the sides of the building are described as facing north and south, and the fac ¸ade as facing east. Phase 1 was represented in the excava- tion by part of the qibla wall and a small part of a limewashed, earth floor (Fig. 1) (2). The wall, which was 0.60 m thick, was constructed with rubble and lime mortar. The mihrab was contained in the thickness of the wall. It was a semi-circular niche 0.70 m wide and 0.35 m deep. The outside of the wall has no salient corresponding with the mihrab, but Ms. Kervran inter- preted an L-shaped construction, which occupied this position, as a buttress to compensate for the thinness of the masonry behind the niche. If the mihrab was on the axis of the mosque, the internal width of the building was at least 8.75 m. Outside the qibla wall are several graves that evidently form part of a graveyard of undetermined extent. The L-shaped feature, however, was not the only construction abutting the outside of the qibla wall. About 1.50 m to the north of it was a second, almost identical con- struction. The plan in Figure 1 appears to show the footing of the first feature extend- ing along the western arm of the L and beginning to return at the end of it; this suggests that the feature is complete. The plan of the second feature is incomplete in Figure 1 but it appears more fully in Figure 2. The size and shape of the two features, and the fact that the plan of one is almost a mirror image of the plan of the other, suggest the possibility that, rather than buttresses, they were parts of a separate building. Indeed, if this was the case, and if the building was symmetrical and square, it could have been a tomb with sides approxi- mately 5.15 m long, comprising four L-shaped piers supporting a dome. Such tombs, of varying sizes and elaboration, have a wide distribution in the Gulf region, as well as farther afield (3). Figures 1 and 2 do not reveal conclusively the chronological relationship between the mosque and the putative tomb, although Arab. arch. epig. 2003: 14: 95–102 Printed in Denmark. All rights reserved 95

The al-Khamis mosque on Bahrain: A note on the first and second phases

Embed Size (px)

Citation preview

The al-Khamis mosque on Bahrain:A note on the first and second phases

DAVID WHITEHOUSECorning, New York

The ruined al-Khamis Mosque is situatednear the north end of Bahrain, 10 kilometresfrom Manama. The building was partlyexcavated in 1986 by the Mission Archeo-logique Francaise, under the direction ofMonik Kervran, who distinguished threephases of construction (1). This note is acomment on the remains and interpretationof the first and second phases. It does notaddress the question of the dates of thesephases, for which very little evidence exists.For the sake of convenience, the qibla wall isdescribed as facing west (in reality, the qiblais a few degrees south of west); the sides ofthe building are described as facing northand south, and the facade as facing east.

Phase 1 was represented in the excava-tion by part of the qibla wall and a smallpart of a limewashed, earth floor (Fig. 1) (2).The wall, which was 0.60 m thick, wasconstructed with rubble and lime mortar.The mihrab was contained in the thicknessof the wall. It was a semi-circular niche0.70 m wide and 0.35 m deep. The outsideof the wall has no salient correspondingwith the mihrab, but Ms. Kervran inter-preted an L-shaped construction, whichoccupied this position, as a buttress tocompensate for the thinness of the masonrybehind the niche. If the mihrab was on theaxis of the mosque, the internal width of

the building was at least 8.75 m. Outsidethe qibla wall are several graves thatevidently form part of a graveyard ofundetermined extent.

The L-shaped feature, however, was notthe only construction abutting the outsideof the qibla wall. About 1.50 m to the northof it was a second, almost identical con-struction. The plan in Figure 1 appears toshow the footing of the first feature extend-ing along the western arm of the L andbeginning to return at the end of it; thissuggests that the feature is complete. Theplan of the second feature is incomplete inFigure 1 but it appears more fully in Figure2. The size and shape of the two features,and the fact that the plan of one is almost amirror image of the plan of the other,suggest the possibility that, rather thanbuttresses, they were parts of a separatebuilding. Indeed, if this was the case, and ifthe building was symmetrical and square, itcould have been a tomb with sides approxi-mately 5.15 m long, comprising fourL-shaped piers supporting a dome. Suchtombs, of varying sizes and elaboration,have a wide distribution in the Gulf region,as well as farther afield (3).

Figures 1 and 2 do not reveal conclusivelythe chronological relationship between themosque and the putative tomb, although

Arab. arch. epig. 2003: 14: 95–102

Printed in Denmark. All rights reserved

95

Ms. Kervran’s interpretation of one of theL-shaped features as a buttress indicatesher belief that they are either contemporaryor that the mosque is earlier. If this iscorrect, the tomb (if that is what it was)stood in the graveyard that developedoutside the qibla wall. If the tomb turnsout to be earlier, one might conjecture that itwas the shrine of a local saint and that themosque was built in its immediate vicinity.

The excavated remains of phase 2 aremore extensive (4). Although nothing sur-vived to a height greater than 0.60 m aboveground level, the plan of the building isclear, at least in outline. The new mosquewas 8.0 m wide and 6.30 m deep internally.The builders preserved the position of theoriginal qibla wall (5), but they employedreused ashlar blocks to construct newlateral walls and a new facade. The newwalls were one block (0.35 m) thick. Eachlateral wall contained the remains of asingle entrance, near the east end. Twostone blocks, which are equidistant fromone another and from the east and westwalls, presumably were the bases of woo-den columns supporting a line of beams,parallel to the qibla and west walls, onwhich poles (probably mangrove poles)would have been laid in an east-westdirection to form the roof. To the west ofthe mosque was a courtyard, which was thesame width as the mosque and was 9.40 mdeep. The remains of the walls did notsurvive to a sufficient height to show the

positions of an opening or openings eitherin the east wall of the mosque or in thewalls of the courtyard, and it is alsopossible that additional openings existedin the north and south walls of the mosque(cf. Fig. 3).

Ms. Kervran drew attention to a singlearchitectural parallel for the mosque of phase2. Unhappily, in doing this, she confused twomosques in the Sultanate of Oman: the

Fig. 1.

The al-Khamis Mosque: remains of phase 1 (after

Kervran & Kalus, La mosquee: Fig. 11).

Fig. 2.

The al-Khamis Mosque: plan of phase 2 (after Kervran

& Kalus, La mosquee: Fig. 13).

D. WHITEHOUSE

96

Fig. 3.

Siraf: plan of the mosque at Site C in periods 2a and 2b (after Whitehouse, Siraf III: Fig. 19).

THE AL-KHAMIS MOSQUE ON BAHRAIN

97

Mosque of the Qadi at Sohar, which shelikened to phase 2 of the al-Khamis Mosque,and a small mosque at Nakhl on the BatinahCoast, which she likened to phase 3. In fact,the small mosque at Nakhl (now destroyed)was a roughly square building with, at thecentre, a single teak column to support theroof. Each side of the mosque had twoentrances and the facade had three, whichprovided access from a rectangular court-yard with a small ablution area. Its familylikeness to the al-Khamis Mosque is evident.The Mosque of the Qadi (also destroyed) hadan internal courtyard, similar to the court-yard in phase 3 (6).

Phase 2 of the al-Khamis Mosque and themosque at Nakhl are just two examples of agroup of mosques which have a widedistribution and range in date from theninth century to the present day. Themedieval members of this group that arenearest to Bahrain are at Siraf on the coastof Iran. At Siraf, the excavation of five smallmedieval mosques, four of which had morethan one phase of construction, and theplanning of surface remains of five othermedieval mosques, provided no fewer thaneighteen plans of mosques of the periodbetween the ninth and the twelfth centuries.With one exception, these mosques form ahomogeneous group, with the followingcharacteristics (Figs. 3 and 4) (7):

1. Like all buildings at Siraf, they wereconstructed with rubble and mortar. Weassumed that most, but perhaps not all,had flat roofs made of poles, matting andmud.

2. They were small, ranging in size from5.4 · 5.5 m to 9.7 · 10.2 m internally.

3. They were square or more often rectan-gular, usually but not always with eitherone or two transverse arcades to supportthe roof. The mihrabs were contained inrectangular salients. The wall oppositethe mihrab had from one to three

doorways and the lateral walls hadvarious combinations of doorways, nich-es and windows.

4. In four cases, they had lateral extensions.5. They were usually accessible from a

courtyard, which might be opposite theqibla wall and have the same width asthe mosque itself.

6. Two of the mosques had staircase min-arets.

7. Three were associated with cemeteries.

Phase 2 of the al-Khamis Mosque sharedseveral characteristics with the small mos-ques at Siraf, although, as far as we know, itdid not have a staircase minaret and themihrab (either a survivor from the originalmosque or an imitation of the originalmihrab) was not contained in a salient.Nevertheless, it is impossible to escape theconclusion that the al-Khamis Mosque andthe small mosques at Siraf belong to thesame tradition, which was present in theGulf no later than the ninth century.

When, in 1980, I published the mosques atSiraf, I was unaware that at least five mosquesof this type existed at Samarra’, the capital ofthe Abbasid caliphate from 221/836 to 279/892. Two of the caliphs’ palaces at Samarra’contained freestanding mosques: the Jausakal-Khaqani, which possessed four, and theBalkuwara, which had one (8). The mosquesin the Jausak al-Khaqani were built presum-ably, like the rest of the palace, between221/836 and 227/842. They were larger thanthe al-Khamis mosque and the mosques atSiraf (the largest had internal dimensions ofabout 9.0 · 21.0 m, with a courtyard about11.0 m deep) but all had, opposite the qiblawall, a courtyard that was the same widthas the mosque itself. The mosque in theBalkuwara palace is presumably contem-porary with the rest of the complex, whichwas built in 234/849-245/859. It is of thesame type as the mosques at the Jausakal-Khaqani, although apparently it did not

D. WHITEHOUSE

98

Fig. 4.

Siraf: plan of the mosque at Site M2 (after Whitehouse, Siraf III: Fig. 23).

THE AL-KHAMIS MOSQUE ON BAHRAIN

99

Fig. 5.

Mosques at Siraf and Samarra’ (after Whitehouse, The Smaller Mosques: Fig. 5).

D. WHITEHOUSE

100

have a courtyard. The finds at Samarra’ showthat mosques with plans like the plan of phase2 of the al-Khamis Mosque were present inIraq by the middle of the ninth century.

This traditional form has survived to thepresent day. The excavations at Sirafrevealed an example of approximatelythe fifteenth century, which was associatedwith a small cemetery. Modern mosquesin the villages east and west of Siraf havethe same plan and in many cases have anidentical or similar courtyard (9). Further-more, mosques with the same plan, some-times associated with similar courtyardsare found in many parts of the Sultanateof Oman; indeed, they are the most com-mon type of small Ibadi mosque in theinterior of the country (10). In addition tothe mosque at Nakhl mentioned above,they include, in the interior, the mosque atIbra Ayla, four mosques at Nizwa (al-Muwadd, al-Shawadhna, al-Jinah and al-Sharja), the al-Maytha’ Mosque in Bahlaoasis and the al-c Awayna Mosque inal-Hayl; and on the Batinah Coast, atBandar Jissah, al-Ribat at Salalah, Rakhyut,Bukha, Masjid al-Hamra’ at Jabru, SurTharmaid and Masjid al-Naqa at Sohar.The mosques at Salalah and Jabru areassociated with cemeteries (11).

Although we have no reason to connectthe al-Khamis Mosque or the mosques atSiraf and Samarra’ with Ibadites, it isinteresting to note that the al-KhamisMosque, three of the medieval mosques atSiraf, the fifteenth-century mosque at Sirafand two of the mosques in Oman areassociated with cemeteries.

References1. Kervran M & Kalus L. La mosquee al-Khamis a

Bahrain: son histoire et ses inscriptions. Arche-ologie islamisque 1: 1990: 7–73. The report is in

two parts. In ‘I. Le monument’ (pp. 7–51) Monik

Kervran described the standing remains and the

results of the excavation, while in ‘II. Les

inscriptions’ (pp. 53–73) Ludvik Kalus published

thirteen inscriptions from the mosque, the earli-

est of which records the construction of phase 3.

It may date from 518/1124-5 and in any case it

mentions the future cUyunid emir Abu Sinan

Muhammad, who came to power about 526/

1131-2. The inscriptions, therefore, have little

bearing on the subject of this note, other thanto show that phase 2 is earlier than the twelfth

century.

2. Kervran & Kalus, La mosquee: 17–18.

3. On the Iranian side of the Gulf, a tomb of this type

is often called an imamzadeh (‘‘house of an imam’’).

The form is said to be derived from that of

Zoroastrian fire temples.

4. Kervran & Kalus, La mosquee: 18–20.

5. It is not clear how much of the qibla wall was

preserved by the builders of phase 2. Ms.

Kervran states (p.17) that ‘Le plan [du mihrab]n’avait pas varie, depuis l’origine. Ses dimen-

sions avaient seulement ete quelque peu reduites

par l’epais badigeon dont on avait recouvert la

paroi, a plusieurs reprises: sa largeur etait ainsi

passee de 0,70 a 0,58 m.’ This statement implies

that the original niche survived, but in a gradu-

ally narrower form as coat after coat of white-

wash was applied to it. The elevation in Fig. 12,

however, seems to show that the reduction in the

width of the niche in phases 2 and 3 had nothing

to do with additional layers of whitewash.Indeed, the most obvious interpretation of Fig.

12 is that the builders of phase 2 demolished the

qibla wall down to the level of the new floor and

rebuilt it with a slightly smaller niche, and that

the builders of phase 3 did precisely the same

thing.

6. Kervran & Kalus, La mosquee: 42–43. Despite the

confusion in the text, the plans in Figs. 32 and 33

are correctly captioned. The plan of the mosque

at Sohar was strikingly similar to the plan of the

old Great Mosque at Nakhl: see Costa PM.Historic Mosques and Shrines of Oman. Oxford:

BAR Int Ser, 938: 2001: 105–113, esp. Fig. 154.

According to the published plans, the mosque at

Sohar was about 28.0 m long and 24.0 m wide

externally, while the mosque at Nakhl measures

about 26.0 · 20.0 m.

7. Whitehouse D. Siraf III. The Congregational Mosqueand Other Mosques from the Ninth to the TwelfthCenturies. London: British Institute of Persian

Studies, 1980: 30–57.

8. Creswell KAC. Early Muslim Architecture, ii.Oxford: Clarendon Press, 1940: Figs. 194 and 214;

THE AL-KHAMIS MOSQUE ON BAHRAIN

101

Whitehouse D. The Smaller Mosques at Siraf: A

Footnote. Iran 22: 1984: 166–168.

9. Whitehouse D. Excavations at Siraf: Second

Interim Report. Iran 7:1969: 39–62; Staircase

Minarets on the Persian Gulf. Iran 10: 1972:

155–158.

10. Costa, Historic Mosques: 35.

11. Costa, Historic Mosques: Figs. 47, 53, 75, 92, 97, 113,

173, 179, 187, 193, 216, 247 and 256.

Address:

David Whitehouse

The Corning Museum of Glass

One Museum Way

Corning NY 14830-2253

USA

[email protected]

D. WHITEHOUSE

102