Archaeology of a Rural Landscape (Tana Delta, Kenya)
Yanis MOKRI, “Archaeology of a Rural Landscape: The Case of the Tana Delta in Kenya.” Mambo! Vol. XVI (8), 2019.
Since the 1940s, the history of coastal Kenya relies essentially on monumental trading centres studies of the last millennium and a half. On the contrary, hinterlands of those remarkable cities were not subjects of research before the late 1970s. During the 1980s two Kenyan archaeologists wrote their PhD thesis on coastal hinterlands, highlighting the interactions with the cities. G.O. Abungu proposed several trade network models between the Upper Tana River and the mouth of the same river, based mostly on similar decorated potteries. However, he did not take into account the deltaic area of the Tana River sufficiently, though he suggested its role of hinterland for Swahili cities. Therefore, my PhD aims to elaborate a history of this region based on the archaeological finds for the last 2 000 years. The study aims to shed light on the rural dynamics and the organisation of the past populations who inhabited the Tana Delta during this time scale. As such, I would like to interrogate the models proposed by my predecessors, based on archaeological and historical material. For this purpose, I try to understand the way people occupied and managed their environment, determining the subsistence and handcraft activities involved into a micro and macro scales, namely the site or village and the Delta. In a broader context, the Tana Delta can be analysed in a global scale: African inland and Indian Ocean trade networks.
The research is part of a broader multidisciplinary programme upon the deltas of Eastern Africa called Western Indian Ocean Deltas Exchange Research. It aims to show the importance of deltas for humans and local ecology, questioning the relationship of the people with every component of the deltaic areas. A historical/archaeological project has been thought to fill a gap in the ancient history of deltas, especially for Kenya’s Tana Delta.
This paper presents the results of the first archaeological season led in 2018, whose purpose was to map most of the archaeological sites and gathering oral information on the local history. It allowed me to draw a first distribution of inhabited sites, past and present. Three of them were selected for deep investigation, based on their situation and the material found. These three sites are discussed after an explanation about the methodology and the material sought.
Material and Methods
The communities inhabiting the Delta are differentiated by their language and activities. Three of them looked more important to reconstitute the local history, due to their ancient presence in the area: Pokomo (Bantu speakers), Orma (Cushitic speakers) and Waata (Cushitic speakers). According to the oral traditions, they practiced respectively agriculture, pastoralism and hunting-gathering. Since their arrival in the Tana region between the 15th and the 17th centuries, their relationships have entailed conflict and trade. Each group provided the others with their own set of production. Obviously there is no tight barrier, allowing the Pokomos to rare animals like chicken or the Waatas having small farms for instance. Moreover, today and maybe in the past, they have all practiced fishing.
As a hinterland, the Tana Delta required a well-structured system where each group could be satisfied while supplying the cities market constantly with ivory, horns, and ostrich eggs, among others. What exactly was this system for those communities and the ones before them? There will be no answers here, but a body of evidence which should be explored with more data in the future.
1.1 Human sciences investigation
After a first eight-day survey in Garsen in March 2018, four zones were deemed interesting for deep investigation: Lake Shakababo, Lake Moa, Kau, and the former mouth of the Tana River.
The second mission, spanned over 5 months between July and December 2018, was auspicious. The main objective was to draw an archaeological map of the area gathering every “historical” feature that could indicate human occupation, like abandoned and inhabited villages, ruined buildings, farms, and even trees.
The research protocol followed three interwoven methods: survey, interviews, and excavation. Old settlements were spotted by relying on oral information and meaningful artefacts found on the ground surface, i.e. sherds of pots. It became clear that observation and questions to dwellers about local history and handcraft practices were important. Questions focused on pot making to determine a period when the production stopped, excluding thereby a potential modern production.
Since the 1980s, pots have been replaced by metal utensils, giving a Terminus Ante Quem. Material found was picked into bags the date of the collection registered, GPS recorded and photographed. Further analyses like excavation were undertaken on sites of high interest.
1.2 Environmental study
Different approaches are employed to reconstitute the past of the Tana Delta. Peopling history is also studied through an environmental scope, through which landscape questions how human made use of it and participated to transform it. Landscape can be defined as the result of the interaction between human and nature on the long term.
Vegetation is one of the most informative constitutive elements. Plants are studied as present and past indicators of human activity. They are easily noticeable, encouraging a land cover study to get clues upon the organisation and the use of land. Their origin helps to reconstitute the interaction at a local and a global perspective. The introduction and movement of plants, especially exotic plants, had consequences on the societies, leading to modification on the local economies as shown in many studies around the world. As one can observe today, natively exogenous plants lead position into the farms of the Tana Delta. In overlapping periods of importation and historical/archaeological events it may give ideas on the local history, explaining partly the reasons why some communities settled into this flourishing area for instance.
Beside this social consideration, vegetation may inform on ecological adaptation and transformation. The idea of millennia pristine area without any human activities in the intertropical belt is no longer considered valid today. Since they are able to produce and use tools, mankind exert a force on the environment. Henceforth landscape is considered as a palimpsest of those interactions, allowing researchers to reconstitute the history of a studied area. Plants are chosen not for their yielding function only but also to contain weathering of soils, to boost fertility, for their protective functions or even aesthetic features. Clues of this past alteration should be supplied in combining archaeological material with data from a current paleoecological study of that region conducted by Peter Gitau, a Kenyan PhD student at the National Museums of Kenya (NMK).
Livestock is also studied through a similar scope as it gives ideas on network and migration, establishment and occupation.
Finally, the human/environment link is thought as a dialectic phenomenon whereby the two parts affect one another, exerting pressure on each other.
2.1. Lake Shakababo: a witness of medieval network
Fieldwork started in early August and focused on Lake Shakababo, where ceramic remains are scattered all along the shore. On the southern side, at Sango Farm Mtetemo, pot sherds and freshwater shells appeared in bulk on top of the area. ‘Mtetemo’ is a shamba of 23.5 ha. The type of pot sherds found is similar to the Early Tana Tradition (ETT) dated from the 7th to the 10th centuries AD yielded in large quantity by a test excavation. Indeed, the last days of August were devoted to a test excavation of 1 x 2 meters that uncovered a miscellaneous and a huge amount of material including ceramics, bones (fish, sheep, and undetermined), iron slags, charcoal, shells (freshwater, marine), round beads made of shell. The presence of cowries attests direct or indirect connections with littoral.
Furthermore, pot holes and what appeared as a base of an earthen building were found at 60 cm under the ground, last level reached. Time playing against us, unfortunately, the geological level was not found. A coming mission will open a new test pit near the former to draw the entire stratigraphy and corroborate the last work.
2.2. Undetermined settlements at Lake Moa
Lake Moa was surveyed during the month of October where a lot of broken pots and one ceramic spoon were found along the shore. The remarkable size and decoration appearing different from the ETT of Mtetemo. They look quiet recent comparing to the latter due to their very good conservation. Based on oral stories and good preservation, one can suppose an ancient production (about 19th century) made by a community inhabiting the place at this time, Giriama or Waata. However, an ongoing study on this material is adjusting and revising this hypothesis. Some sherds present a decoration depicted by researchers as distinctive ETT motifs. A test excavation is planned to contextualize those artefacts and others to enlighten the past of Lake Moa.
2.3. A worldwide connected site: Kau
Oda Branch and Matomba Branch, respectively on the southern and northern parts of the Tana River, meet where Kau village stands. Here starts an ecotone transition from fresh water to saline water implying a change in ecosystem. This crossroad village was likely a major place for trade and harbouring for ships coming from the coastal towns. It could have facilitated circulation towards inland. It is not a surprise then if Kau used to be home of the Nabahani sultanate during the first half of the 19th century and, as such, an important trade hub for the region. Indeed, archaeological artefacts were found in this inhabited village of a Muslim Pokomo community. A Muslim cemetery with tombs of saints attests to this role. Sherds of pots coming from all over the world were discovered like Chinese ceramics (blue-and-white, stoneware), European ware (floral decoration lying on a white slip), and what appeared to be Islamic ware (yellow slip). Until now they are the only witness of this type of ceramic discovered in the Tana Delta. Local production was found dating probably from the 19th century. It is supported by upstream discoveries along the Oda Branch at Hanasa for instance, where Indian merchants and the British army were settled at the time. It should be noted that ancient agricultural activity at Kau is witnessed by a millstone used today as a gravestone, among others. Much more evidence would be gathered with upcoming surveys and excavations.
Many other places have been surveyed along the Oda Branch, based on a map of the area dated from 1893 drawn by W. Fitzgerald. The villages have provided good clues on local 19th century history, opening perspectives for further investigations.
The Tana River meets the Indian Ocean at the Ungama Bay via four outlets. The main outlet reached the Indian Ocean by the North and is bordered by the town of Kipini where the ruins of Ungwana stand. The three other outlets are distributed along Ungama Bay. About 6.9 Km from the ocean to inland consist of sand dunes raising several meters in height and cultivated by farmers. Despite oral information telling about vestiges, none was discovered at the place explored (2°38’60” S; 40°15’50” E).
All the material is kept at the NMK in Nairobi, and is partly under study. All the information recorded within this five-month fieldwork integrates a GIS database; an archaeological map of the Tana Delta is in progress. Additionally, with further investigation in the above-mentioned places, Dominic Mathina from Nile Surveys and Geosolutions Ltd. conducted a topographical survey in November 2018.
While several preceding works suggested the historical importance of the Tana Delta, this 2018 fieldwork brought material evidence. The three sites explored of Mtetemo, Moa and Kau made it possible to cover a wide area. They open windows on different spaces and times, allowing a diachronic approch of the region. The archaeological material should give hints on the systems scrutinized at the levels of the village, the community and the Delta. As such, the system is understood as the organisation of social and economic lives of a community.
Through this work, it is expected to reconstitute parts of an ancient rural world poorly known in coastal Kenya. The overall ambition is to provide a better understanding of the rural networks at a local scale with Mtetemo and Moa, and at a global scale with Kau, clearly opened to the Indian Ocean network. These reconstitutions will be possible only with a broader range of data, meaning more excavations in the future.
I would like to thank people and institutions for allowing me to conduct research in the best conditions: Kenyan partners, the NMK and the NACOSTI; and French institutions, IFRA, IRD, IMAF and University Paris 1 – Panthéon Sorbonne. Thanks to Kenedy Otoi, Francktough Daido, Aziz Mbarak and the Tana Delta people.
 Henry Mutoro, 1987, “An Archeological Study of the Mijikenda Kaya Settlements on the Hinterland Kenya Coast”, PhD diss, University of California; George Abungu, 1989, “Communities on the river Tana, Kenya: an archaeological study of relations between the delta and the river basin, 700 – 1890 A.D.”, PhD diss, Cambridge University.
 The notion of environment here must be understood as wide as possible, where people develop in space and environmental ecology.
 This project is registered in NACOSTI as: “DeltArcheoGIS – Archaeological mapping of the Tana Delta”.
 Waata are associated to the Cushitic speakers Oromo. See A.Kassam and A.B.Bashuna, “Marginalisation of the Waata Oromo Hunter-gatherers of Kenya: Insider and Outsider Perspectives”, Africa: Journal of the International African Institute, 74 (2), 2004, pp.194-216. https://doi.org/10.3366/afr.2004.74.2.194
 The outlet today was not the same during the past centuries. The Belazoni canal built during the first half of the 19th century changed the course of the Tana River.
 This concept used in archaeology defines the chronological limit before which an event occurred with certainty. It is the final limit of this event, here the pot making.
 For a recent one, see Christian Thibon et al., 2019, “Dissemination of the American Plants in the Great Lakes Region: General and Sub-regional Approaches, the Western Kenya”, Les Cahiers d’Afrique de l’Est, n°52. https://journals.openedition.org/eastafrica/452
 W. Balée and C. Erickson, Time and Complexity in Historical Ecology: Studies in the Neotropical Lowlands, 2006; G. Saulieu, M. Elouga, B. Sonké, Pour une écologie historique en Afrique centrale, 2016; S. Rostain, Amazonie : jardin sauvage ou une forêt domestiquée. Essai d’écologie historique, 2016.
 “Garden, field” in Kiswahili. No crops grown there but many Azadirachta Indica (mkilifi) seem artificially planted, as well as mango trees.
 Only 3 ha were mapped and just 0.12 ha were surveyed with material gathered on the ground.
 Also known as Tana Ware or Triangle Incised Ware (TIW). See M. Horton and F. Chami, “Swahili Origins”, in The Swahili World, A. LaViolette and S. Wynne-Jones (eds), 2017; S. Wynne-Jones and J. Fleischer, “Ceramics and the early Swahili: Deconstructing the Early Tana tradition”, African Archaeological Review, 2011. https://doi.org/10.1007/s10437-011-9104-6
 Giriama is a community belonging to the famous Mijikenda, « nine towns », a Bantu speaking community composed by nine “tribes” leaving on the East African coast. Waata are known as the most ancient people of the region, according to the Pokomo, traditionally hunter-gatherers.
 C. Coret, 2016, La refondation d’une cité swahili à Witu. Écriture de l’histoire et légitimation du pouvoir au nord de la côte est-africaine (1812-1895), PhD Diss., Université Paris 1 – Panthéon Sorbonne. http://www.theses.fr/s90807
 This assertion seems to corroborate the words of G. Abungu saying that no exotic artefacts were found inland in an archaeological context. However, the cowries at Mtetemo give an opportunity to reassess this statement.
 C. Coret mentioned the presence of a jail at Kau.
 W. Fitzgerald, Travels in East Africa, 1894.
 This medieval stone-town is already well studied. Two other medieval towns are close to Ungwana: Mwana and Shaka. See G. Abungu, op.cit.; J. Kirkman, 1966, Ungwana on the Tana, The Hague, Mouton.
 Olivier Hamerlynck informed me about sherds of pots he found few years ago.
 A total of 5 516 sherds of pot from all sites are recorded. More than 3 200 come from Mtetemo.