The urgency lent this field project by the rapid nature of development on Bahrain was continually reinforced even through the duration of one year’s fieldwork. Hence a concurrent strand to the archaeological excavation and survey was provided by an inventory of extant Islamic shrines, usually linked with the Shi’ah communities, which was also completed. These shrines (see Bahrain 2) are often ephemeral structures, but which in many instances were themselves linked with archaeological sites, i.e. in being built upon mounds, or enshrining ruined or near ruined buildings. This recurrent reworking of, and relationships with the past was also evident in a mosque and its associated pool, Abu Zaydan, which were cleared, and which indicated the incorporation and inversion of pre-Islamic, possibly Dilmun structural elements including a large sacrificial altar in Bilad al-Qadim (see Bahrain 3). Moreover, Abu Zaydan offered a glimpse, potentially, into past concepts of materiality in relation to water, often ignored, or treated as conceptually analogous with today (see Publications). 

Currently I am working with Rachel MacLean on an archaeological guidebook to Bahrain, The Archaeological Guide to Bahrain (Archaeopress) for publication in 2010.

 

L-R. [1]. Map of part of Gujarat. [2]. Carnelian sample collection in progress.

2) Source Analysis and the African Western Indian Carnelian Bead Trade. Gujarat, India (2000).

Following a reconnaissance survey in 1997, I completed a field season collecting carnelian samples in Ratanpor district of Gujarat in June 2000. This was undertaken in co-operation with Dr Kuldeep Bhan of MS University, Baroda, and was funded by the British Academy. The research project of which this fieldwork formed a part had grown from the results of my excavations in Gao and Timbuktu in Mali (see below). Specifically because within the bead assemblages recovered from these two cities were carnelian beads which I had putatively identified as 'visually' Gujarati imports (see India Carnelian Beads). These in turn were potentially indicative of trade, albeit indirect, between Western India and West Africa, perhaps from as early as the eighth century AD. This visual identification was based upon the presence of characteristic bead types such as long bicones, as well as the deep red colour of the beads, both features of Gujarati carnelian beads, in particular those produced in the town of Cambay/Khambhat. However, rather than basing these parallels on visual analysis alone, chemical analysis was required in order to indicate a match between the Gujarati and West African carnelian beads.

To achieve this carnelian samples were collected from various locations in Gujarat including shallow mine workings, and merchants and contractors stockpiles, as well as the manufacturing sites (see India Carnelian Beads). Carnelian bead manufacture continues to this day, but on a much reduced scale to that formerly found. Thus an ancillary element of this project was the making of a film on the processes involved in bead manufacture including roughing out, fine shaping, drilling, polishing, grinding, and baking to alter the colour (the characteristic deep red). This film is available from me for teaching purposes free of charge (Contact). The carnelian samples were ultimately analysed using UV-LA-ICP-MS in conjunction with Dr Dave Polya of the University of Manchester and Dr Kym Jarvis of Kingston University, and were compared with the archaeological beads from Gao and Timbuktu, and various other sites in West Africa. The compositional data indicated some degree of overlap between the Gujarati and West African samples but also indicated that further research was required especially with regard to standardisation (see Publications), and this has now been completed by Dr Sharon Fraser as a PhD project. A third supplementary ethnographic element to this field project involved visiting the Siddi or Habashi community in Ratanpor village. The Siddis being, potentially, a living testimony to African-Indian connections for they are a people of African origin (see India Carnelian Beads).

 

L-R. [1]. Map of West Africa. [2]. Excavations in progress in Timbuktu, Sept. 1998.

3) Trade, Urbanism, and Religion on the Saharan Frontier. Timbuktu (1996-1998).

Timbuktu is a place, or at least a name, synonymous in many people's imagination with the remote and exotic. In more prosaic reality it is a city located on the River Niger in Mali where proximity to the river permits occupation very close to the edge of the Sahara desert, thus facilitating access to trans-Saharan trade routes. Hence considering its fame it is surprising that Timbuktu had never been the focus of systematic archaeological investigation until the programme of survey and test excavation I completed there between 1996 and 1998. This fieldwork had a variety of aims that slotted within the framework of the overall research project on trade, urbanism, and religion on the Saharan frontier. These included assessing the origins of the city and its role in local, inter-regional, and long distance trade. A further associated aim was to attempt to delimit the spatial extent of Timbuktu over time via survey, and also to evaluate the importance of the city in comparison to its now largely forgotten neighbour, Gao (see below). The research was completed in co-operation with the Malian Institut des Sciences Humaines, their representative Mr Nafogo Coulibaly, and Mr Ali Ould Sidi of the Bureau de Patrimoine Culturel in Timbuktu. The fieldwork was funded by the British Academy and the McDonald Institute for Archaeological Research, University of Cambridge.

Initially a survey programme was completed. But unfortunately the survey was not as simple as originally envisaged owing to modern occupation, flooding, and shifting sands obscuring the archaeological deposits (see Mali Timbuktu). This meant that delimiting the primary areas of occupation based upon survey data alone was impossible for what were apparent were islands or snapshots of archaeology rising through the sands rather than a comprehensive archaeological overview (see Publications). This is perhaps unsurprising for a single sand storm has been recorded as depositing up to a metre of sand. The factor of sand accumulation was also found to be of great significance in the five test excavations dug at various locations in Timbuktu.

These excavations were completed in areas which either the survey results, historical records, or oral testimony suggested might be of potential significance. Key in this respect were excavations completed adjacent to and in the vicinity of the extant Sankore mosque which was possibly founded in the fourteenth century AD (see Mali Timbuktu).  However even here, the removal of five metres of deposits only reached back to levels associated with the eighteenth century. Nonetheless, a useful range of material was recovered; including clay tobacco pipes, mollusc shells, locally produced pottery, and beads and bracelet fragments (see Mali Timbuktu). Ultimately, what can be concluded from this exploratory foray into the archaeology of Timbuktu is that further research is needed, but pending this it would seem that Timbuktu was never as important as its neighbours such as Gao, except in western popular imagination.

Finally, as a postscript, it should also be noted that during the course of the fieldwork we recorded the existence of two British war graves which had been overlooked by the Commonwealth War Graves Commission. These were the graves of the Merchant Navy seamen, Chief Engineer William Souter (died 28 May 1942, age 60) and Able Seaman John Turnbull Graham (died 2 May 1942, age 23). Both men had been prisoners of the Vichy French authorities and were kept in appalling conditions in Timbuktu. Our (re)discovery of their graves allowed the reinstatement of the stones in an upright position, their cleaning, and the rebuilding and painting of the wall surrounding the graves (see Mali Timbuktu).

 

L-R. [1]. Map indicating the position of the Dahlak Islands. [2]. Survey of an Aksumite? temple.

4) Reconnaissance Survey of the Dahlak Islands, Eritrea (1996).

In June 1996 I completed a short reconnaissance survey of the site of Dahlak Kebir, on the Dahlak Islands off the coast of Eritrea in the Red Sea (see Publications). This research was undertaken in part, to be honest, for the excitement of being able to undertake archaeological research in virtual terra incognita and within a recently founded nation state, Eritrea, then newly emergent from the civil war with Ethiopia. The reverse of this being that the absence of a formal antiquities legislation meant that work had to be restricted in scope, though that which was completed was undertaken courtesy of the National Museum in Asmara. The fieldwork was funded by St John's College, Cambridge, and the British Institute in Eastern Africa, and I was accompanied by Dr Charles Spence and Mr Yassin Adem.

The survey recorded a range of remains seemingly indicative of occupation at Dahlak Kebir from the Aksumite through to the Ottoman periods, hence perhaps from the first century BC through to the latter sixteenth century AD when the islands were occupied by the Ottoman Turks. In making this suggestion the existence of prehistoric occupation is not denied, which is almost certain to have existed, but this was not a focus of the field research described here. Possible Aksumite remains included the remains of a well-built structure situated on top of a small mound, and of especial interest was the use of spiral carved marble columns in this building with one found still in situ and another nearby (see Eritrea - Dahlak Islands). This structure perhaps represents the remains of a small church, not inconceivable considering the proximity of Ethiopia, and might date from the second to fifth centuries AD. Otherwise, the bulk of the visible remains appeared to date from later Islamic periods. These included extensive cemeteries with associated black basalt Kufic inscribed tombstones and collapsing above-ground tomb structures, as well as numerous cisterns, rock cut and lined with plaster, used for storing water (see Eritrea - Dahlak Islands). The latter are possibly testimony to the trade in slaves in which the inhabitants of the Dahlak Islands were involved. An extensive stone-built settlement area with possible associated harbour was also recorded. A wide range of archaeological materials were also present scattered on the surface of the settlement and adjacent areas. None of these were removed, but observations made in-situ indicated trade items such as various types of imported Chinese ceramics, Islamic glazed ceramics, glass and carnelian beads, coral fragments, and glass vessel and bracelet fragments (see Eritrea - Dahlak Islands and Publications).

 

L-R. [1]. Map indicating the Trans-Saharan trade routes. [2]. Visit by dignitaries to the site, Oct. 1996.

5) Trade, Urbanism, and Religion on the Saharan Frontier. Gao (1993-1996).

My formative field research project was completed (1993) as part of my PhD research, and subsequently also formed a component of a post-doctoral project (1996) which ultimately led to my book The Archaeology of Islam (see Publications). The fieldwork was funded by the British Academy, the UAC of Nigeria Fund, and the McDonald Institute for Archaeological Research, University of Cambridge. The research was completed in co-operation with the Malian Institut des Sciences Humaines, and their representatives Mr Nafogo Coulibaly, Mr Sekou Berte, and the late Dr Tereba Togola, as well as Mr Elmoctar Toure of the Musee in Gao.

Gao was the capital of the Songhai empire, the last and greatest of the 'medieval' West African empires, and flourished between the fifteenth and late sixteenth centuries AD. Historical sources, for example the accounts written by Arab geographers and travellers, had been exhaustively combed by historians, but the archaeology of the Gao region had largely been neglected. Hence, in part, the rationale behind instituting this research project was to redress this disciplinary imbalance. Moreover, reconstructions of societies of the Sahelian fringe (i.e. south of the Sahara) were still often portrayed through the tint of North African 'Arabized' and 'Islamised' spectacles, even if this was interpretively innaccurate.

Thus the two seasons of excavations (and an initial preliminary reconnaissance survey) were targetted at letting the archaeology 'speak' rather than reconstructing it based upon an already predetermined historical 'blueprint'. To achieve this excavations were completed in the area of Gao known via oral tradition to have been linked with its 'medieval' heyday, Gao Ancien. Further excavations were focussed upon other sites likewise locally deemed significant in this respect, including Koima on the opposite bank of the River Niger, the Gadei quarter of Gao, and Gao-Saney with its cemetery and twelfth century imported Spanish marble Islamic grave stones and associated tell site. The excavations uncovered structural remains including part of a possible mosque, a rich merchants' house or palace, and a dry-stone defensive wall with an associated gatehouse in Gao Ancien all dating from the twelfth to thirteenth centuries AD (see Mali Gao 1). In Gadei the architectural traditions were found to differ, as part of a collapsed mud roundhouse was recorded, an architectural form not found in Gao Ancien. At Gao-Saney large-scale looting to obtain objects for the antiquities market meant less survived (see Mali Gao 1), but this was more than compensated for by the extensive assemblages of all categories of archaeological materials recovered (see Mali Gao 2 and Publications).

The latter included items of trade; glazed ceramics from North Africa and Spain, glass from Egypt and Palestine, alabaster possibly from Yemen, as well as numerous beads including, again, putative Gujarati carnelians (see above). Great quantities of locally produced ceramics were also recovered which far outweighed imported pottery in quantity. Whilst the faunal remains recorded indicated that the inhabitants of Gao consumed both domesticated animals such as cattle, sheep, and goats, as well as fish such as Nile perch, and hunted animals, for example reedbuck, which inhabited the river bank environments, but which are absent in the region today due to over-hunting. Botanical evidence indicated that rice and pearl millet were present, and hence in all probability consumed, and cotton was also being grown as seemingly represented by cotton seeds recovered (see Mali Gao 2 and Publications). Yet the most spectacular faunal material was not related to diet but to trade. This consisted of a cache of some sixty hippopotamus tusks which had been placed within a pit and laid on wooden beams below the rich merchants' house or palace in Gao Ancien. This cache was radiocarbon dated to the late ninth - early tenth centuries AD and provides evidence for a trade, probably trans-Saharan, in hippopotamus ivory (see Mali Gao 3 and Publications). Epigraphic evidence also indicated that the process of Islamisation was slow and did not apply to all the population. Furthermore, this was not achieved by conquest and colonisation by Arabs from North Africa as was once thought. Instead, local Songhai names, rather than Arab ones, were found on Muslim tombstones indicating indigenous conversion to Islam, including from what must have been relatively early Islamisation contexts of the twelfth to thirteenth centuries AD.

In summary, the results of the excavations in Gao indicated the indigenous nature of Gao and by implication the Songhai empire. They also indicated the complexity inherent in conversion to Islam and its long drawn out and syncretic nature. Finally, of especial interest was the seeming absence of deposits in the excavated units actually associated with the period of the Songhai empire itself, i.e. the fifteenth to late sixteenth centuries AD. Supplementary to the excavations ethnographic data on a rare survival of a type of sewn boat was collected and an example brought back to the UK (see Mali Gao 3 and Publications).