History I NEED HELP ON MY HOMEWORK

McIntoshandMcIntoshInlandNigerDelta.pdf

Home >History homework help >History I NEED HELP ON MY HOMEWORK

Journal of African History, aa (1981), pp. 1-22 Printed in Great Britain

THE INLAND NIGER DELTA BEFORE THE EMPIRE OF MALI: EVIDENCE FROM JENNE-JENO*

BY RODERICK J. McINTOSH AND SUSAN KEECH McINTOSH

F R O M the sixteenth century until the late nineteenth century, when French colonial mercantile interests altered centuries-old trade patterns in the Inland Niger Delta, the city of Jenne dominated the central delta as a major commercial centre. Linked to Timbuktu by 500 kilometres of navigable riverway, Jenne was a major entrepot in the long-distance trade networks along which passed major commodities like gold, salt, copper, slaves, and kola, and a host of minor commodities.

Historical sources permit a great deal of certainty about the preceding two statements. For periods earlier than the fifteenth century, however, Jenne retreats into historical obscurity. The failure of any Arab or European source prior to the mid-fifteenth century to mention the town by name has convinced various historians of the Western Sudan that Jenne did not exist substantially prior to this time. Such a conclusion is contradicted by the oral traditions of Jenne, which state that the town was founded in the second century after the Hegira (i.e. in the eighth century A.D.). Until now, however, the negative evidence of the early written sources has tended to prevail over the testimony of the oral traditions. Archaeological investigations undertaken in 1977 at the ancestral site of Jenne, known as Jenne-jeno or Djoboro, have demonstrated that the settlement was established by the third century B.C. and had developed into an urban centre of considerable proportions by the eighth or ninth century A.D. In this article we present a summary of the excavation results and briefly consider the implications of the archaeological data for our understanding of the development of the Western Sudan during the first millennium A.D.

We begin with a discussion of historical sources for the Western Sudan in order to assess the quality of available information on the origins of the city, the reasons for its growth, and its function within the extensive trade networks of the second millennium A.D. The available sources include: the scholarly compendia of mediaeval Arab geographers and historians based in North Africa, Spain, and the Near East, European accounts, local histories

* Djenne was the preferred spelling during the colonial era. More recently Malians have dropped silent letters and accents from many placenames. The authors are grateful to Dr Alpha Oumar Konare, now Minister of Culture in Mali, for granting official permission for the research. The assistance of M. Hama Bocoum and M. Ali Tamboura was also central to the success of the field season at Jenne. Special thanks are extended to Professors Brian Fagan, Merrick Posnansky, Albert Spaulding and Dr John Alexander for their suggestions and guidance on the research.

Funding was generously provided by the National Science Foundation, the University of California, the American Association of University Women, and the Crowther-Benyon Fund of Cambridge University.

This article was written while one of the authors (R. J. M.) held a postdoctoral fellowship provided by the American Council of Learned Societies. An earlier version of this article was presented at the African History Seminar, Institute of Commonwealth Studies, University of London, on 14 March 1979.

0021-8537/81/2828-2050 $02.00 © 1981 Cambridge University Press terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0021853700018983 Downloaded from https://www.cambridge.org/core. Access paid by the UCSB Libraries, on 10 Oct 2021 at 05:32:03, subject to the Cambridge Core

https://www.cambridge.org/core/terms

https://doi.org/10.1017/S0021853700018983

https://www.cambridge.org/core

2 RODERICK AND SUSAN McINTOSH

(taWikhs) written in Arabic by native African scholars, and indigenous oral tradition.

Chronology of written documentation

The earliest details on the Western Sudan are provided by Arab geographers and historians, none of whom mentions Jenne by name. From the works produced between the eighth and the fifteenth centuries, a fairly detailed picture emerges of principal political, ethnic and religious divisions in the Sahara, and of important trade routes from North Africa to the southern trans-Saharan trade termini of Awdaghust, Walata, Ghana, and Tadmekka. The transfer of commercial control from Arab or Berber hands to black hands in the southern Sahara is reflected in the Arab sources by a decline in the amount and precision of detail given for trade routes and commercial centres south of the Sahara.1 The southern frontier of Arab knowledge of the Western Sudan seems to be the Senegal and Niger Rivers,2 which the Arabs believed to be but one river, the Nile, ultimately flowing past Cairo and into the Mediterranean. South of the river, the Arabs knew only that a pagan tribe of blacks, the Lamlan, lived. Beyond the Lamlan lay uninhabited territory.3

It is difficult to know how much of the vague reportage of sub-Saharan Western Sudan by Arab authors stems from genuine ignorance or confusion (possibly created in part by deliberate obfuscation on the part of Sudanese traders seeking to conceal the trade routes to the gold fields) and how much is a function of selective reporting. It is possible that the towns of black Africa held little fascination for cosmopolitan Arab scholars, many of whom wrote for royal patrons who possessed black slaves. Particularly since many black societies were pagan at this time, Arab authors may have been reluctant to give them detailed treatment.4 Whatever the reason, the pattern of vague and confused reportage for sub-Saharan Africa is sufficiently consistent in the Arab sources that we may conclude that their failure to mention Jenne cannot be assumed necessarily to mean that the town was not yet in existence.

Ibn Battuta requires special mention here because, unlike other Arab authors of this period, he actually travelled through the Western Sudan

1 Ibn Hawqal's Saharan itinerary, for example, is quite consistent with respect to travel times between Sijilmassa, Awdaghust and Ghana, and Kawkaw, Marandet and Zawila. But travel times between the (presumed) Sahel/north savanna centres of Kugha, Sama, and Kuzan are greatly exaggerated, unless a great deal of doubling back and circular travel were involved (Ibn Hawqal, in J. M. Cuoq, Recueil des Sources Arabes concernant VAjrique Occidentale du VIII' au XVI' Siecle (Paris, 1975), 73). Also, reports of towns near or on the 'Nile' (i.e. the Niger or Senegal Rivers) are almost invariably vague, and range from confusing to contradictory, as in the case of Sama, Barisa/Yarasna, Ghiyaru, and Kugha, variously mentioned by Ibn Hawqal, al-Bakri and al-Idrisi, among others.

2 But only as far south as perhaps 15° and 140 latitude. Even when describing pagan Lamlam territory, which is about as far south as Arab knowledge extended, al-Idrisi provides details of a very dry savanna or Sahel habitat, where the inhabitants raise camels and goats and eat camel meat dried in the sun (al-Idrisi, in Cuoq, Recueil, 132).

3 al-Idrisi, in Cuoq, Recueil, 131; al-Harrani, ibid. 248. 4 In the introduction to Ibn Hawqal's Surat al-Ard we find an example of this attitude:

' I have not described the countries of the blacks in the west... because the characteristics of organized states.. .are utterly lacking among them', translated in J. S. Trimingham, A History of Islam in West Africa (Oxford, 1962), 1-2.

terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0021853700018983 Downloaded from https://www.cambridge.org/core. Access paid by the UCSB Libraries, on 10 Oct 2021 at 05:32:03, subject to the Cambridge Core

https://www.cambridge.org/core/terms

https://doi.org/10.1017/S0021853700018983

https://www.cambridge.org/core

JENNE-JENO AND THE INLAND NIGER DELTA 3

0352-3) to the capital of the kingdom of Mali and thus provides numerous details of the sub-Saharan towns through which he passed. Although he visited Walata, Timbuktu and Gao in his travels, Ibn Battuta makes no mention of Jenne. This might appear to be a particularly strong piece of negative evidence for either the non-existence or relative unimportance of Jenne in the mid fourteenth century. There are several reasons, however, why negative evidence from Ibn Battuta should be regarded critically. Excellent as many of Ibn Battuta's descriptions are, there are large gaps in his narrative wherein a journey of many days' duration received only a paragraph's notice.5

Either his daily journal or his memory was very selective, although part of the problem surely was that his notes from twenty-four years of world travel were dictated to a professional writer in Fes who produced a synthesis in only three months' time.6

Ibn Battuta's account of his trip both to and from Mali is particularly sparse in detail; as a result, his itinerary has been the subject of substantial debate. Delafosse7 argued that Ibn Battuta entered the Inland Niger Delta after crossing the Niger southwest of Jenne, from whence he travelled to Segou and on to Mali. If this were true, Battuta's failure to mention Jenne would strongly imply that the town came to play a critical role in riverine transport in the upper Inland Delta only after the mid-fourteenth century. However, both Meillassoux8 and Hunwick9 offer interpretations of the itinerary that keep Ibn Battuta well to the north of the Inland Delta. We are especially partial to Hunwick's reconstruction, which emphasizes the important point that neither on his journey from Walata to Mali nor en route from Mali to Timbuktu did Ibn Battuta mention crossing the Niger River.10 Given that Ibn Battuta travelled overland, probably to the north of the Inland Delta, it is perhaps not particularly significant that he does not mention Jenne, since we would necessarily expect him to do so only if he travelled through the Inland Niger Delta by boat, which he clearly did not. It initially appears puzzling that Ibn Battuta should have travelled overland on his return rather than taking advantage of the efficient transport by boat, especially between Koulikoro and Timbuktu. However, it may well be that his journey, in the company of a merchant, merely reflected the seasonal shift in commercial routes as the water level became too low for successful heavy transport along

6 Ibn Battuta's month-long stay in Gao merited only seven sentences (Ibn Battuta, in Cuoq, Recueil, 316).

* R. Mauny, V. Monteil, A. Djenidi, S. Robert and J. Devisse, Textes et Documents Relatifs a I'Histoire de I'Afrique: Extraits Tires des Voyages d'lbn Battuta, Universite de Dakar, Publications de la Section d'Histoire, No. 9 (Dakar, 1966).

7 M. Delafosse, 'Le Gana et le Mali', Bulletin du Cotnite d'Etudes Historiques et Scientifiques de I'A.O.F. vn (1924), 479-542.

8 C. Meillassoux, 'L'ltineraire d'lbn Battuta de Walata a Mali', J. Afr. Hist, xm (1972), 389-95-

' J. O. Hunwick, 'The mid-fourteenth century capital of Mali', J. Afr. Hist. x\v (1973), 195-206.

10 If Hunwick's interpretation of the southward journey is correct, Ibn Battuta approached the Niger at a point well to the southwest of Jenne, but never crossed the Niger or travelled in the Delta. His return from Mali would have taken him north of the Niger, via Mema, until he reached the Niger again at Timbuktu, from whence he travelled by boat downstream to Gao.

https://www.cambridge.org/core/terms

https://doi.org/10.1017/S0021853700018983

https://www.cambridge.org/core

4 RODERICK AND SUSAN McINTOSH

the middle Niger.11 Claims that Ibn Battuta's overland itinerary proves that the Middle Niger had not yet developed by the mid-fourteenth century as an important transport axis should be reconsidered.

In general, factors of selective reporting and the northern perspective of Arab authors (focusing on overland trade routes to the north of the Niger, particularly trans-Saharan routes) help account for their failure to mention Jenne. We must wait until 1493—6 for a reference to Jenne in an Arab source. At this time, the Moroccan al-Maghili wrote a long epistle to the Askia Mohammed, ruler of the Songhay Empire, concerning the proper conduct of a Muslim ruler and his subjects. In the course of a discussion on the prevention of public nudity, al-Maghili noted that the offence was particularly widespread at Timbuktu and Jenne.12

The first European reference to Jenne occurs in a letter sent from Tuat in 1447 by the Italian merchant, Antonio Malfante. The letter, written in Latin, mentions various civitates bordering the Land of the Blacks, including Chuchiam (Gao), Thambet (Timbuktu), Geni (Jenne), and Meli (Mali).13 As Prussin14 points out, the interpretation of the term civitate is extremely ambiguous. It may be translated as city or state, but Malfante's connotation of the term probably was closest to that of the fifteenth-century Italian city-state, implying the presence of a civitas and an administered cantado, or hinterland. Although Malfante gives no information on the location of Jenne,15 several other details in his narrative suggest that his informant, a wealthy merchant who claimed to have lived in Timbuktu for thirty years, was specifically describing the Inland Niger Delta when he spoke to Malfante of the Land of the Blacks. Malfante recounts:

Every day he [the informant] tells me wonderful things of these peoples [the Blacks]. . .they all go naked, save for a small loin cloth to cover their privates. They have an abundance of flesh, milk and rice, but no corn or barley. Through these lands flows a very large river, which at certain times of the year inundates all these lands. This river passes by the gates pf Thambet.. .There are many boats on it, by which they carry on trade.16

The seasonally inundated land upstream of Timbuktu which produces an abundance of rice can only be the Inland Delta. Thus we learn that trade by pirogue was common on the Niger between the Inland Delta and Timbuktu in the mid-fifteenth century.

Malfante's mission in Tuat was apparently to reconnoitre a commercial 11 The middle Niger is easily navigable only during the flood season (mid-July to

December); as soon.as the floodwaters recede, sandbars lie very close to the water surface. Even pirogue traffic comes to a virtual standstill between Markala and Mopti from March to the end of June (J. Champaud, 'La navigation fluvial dans le Moyen Niger', Cahiers d'Outre-Mer, LV(I96I) , 255-92, esp. pages 259, 281, 287). Ibn Battuta travelled from Mali to Timbuktu in late February/early March (Cuoq, Recueil, 289, footnote 1).

12 al-Maghili, in Cuoq, Recueil, 430. 13 G. R. Crone, The Voyages of Cadamosto (London, 1937), 87; original letter reproduced

in Ch. de La Ronciere, La Decouverte de I'Afrique au Moyen Age, Memoir of the Royal Society of Geographers of Egypt, No. 5 (Cairo, 1924), 1, 151-8.

14 L. Prussin, 'The Architecture of Djenne: African Synthesis and Transformation' (Ph.D. dissertation, Yale University, 1973), 66-7.

16 La Ronciere proposed the equation of Geni with Jenne: La Ronciere, Decouverte, 148, 154-

16 Crone, Voyages, 88.

https://www.cambridge.org/core/terms

https://doi.org/10.1017/S0021853700018983

https://www.cambridge.org/core

JENNE-JENO AND THE INLAND NIGER DELTA 5

route to the interior of West Africa in order to secure gold at its source for Italian interests before it could reach Portuguese merchants on the Guinea coast.17 Malfante's explicitly southern, sub-Saharan interests might help explain why he mentions Jenne, whereas other, Arab writers availing themselves of informants in North Africa failed to do so.

Much greater detail on Jenne is available in the early sixteenth-century Portuguese accounts of V. Fernandes (1506-7) and Pacheco Pereira (1506-8). Both were seamen employed by King Joao II of Portugal to assess the overland and river routes to the heartland of the gold trade in the interior of West Africa. Fernandes specifically mentions the desire of King Joao to reach the cidades of Timbacutu (Timbuktu) and Gyna (Jenne) via the Senegal River, which he clearly believed to be connected to the Niger. Despite the fact that all Fernandes' information was obtained second-hand, his descriptions of the riverine trade network linking Jenne to Timbuktu, and the Wangara-controlled gold trade are detailed and accurate.18 Both Fernandes and Pacheco Pereira describe Jenne as a walled city and emphasize her role in the gold and salt trade; the latter author also mentions the demand in Jenne for tin, copper, and cloth.19 It is interesting to note that during less than a century of Portuguese navigation along the Guinea coast, more was learned about the structure of the gold trade between the mines and the northern Sudanese entrepots than during seven centuries of Arab observation from the north.

Another important source for information on the Western Sudan is Leo Africanus, who travelled extensively in the interior in both 1511 and 1512—14 and provided eyewitness descriptions of both Timbuktu and Gao. However, Mauny20 suggests that Africanus' descriptions of Jenne and the capital of Mali were probably second-hand. Had he actually travelled upstream from Timbuktu to 'the kingdom of Ghinea' and Mali, Africanus could not have claimed, as he does, that the Niger flows westward along that route.21

Africanus' contradictory description of Jenne has caused some confusion. He states that the capital of Genni (also called Gheneoa, Ghinea22) was ' a certaine

17 N. Levtzion, Ancient Ghana and Mali (London, 1973), 133; E. Herbert, 'Timbuktu: a case study of the role of legend in history', in B. K. Swartz and R. Dumett (eds.), West African Culture Dynamics (The Hague, 1980), 439.

18 V. Fernandes, Description de la Cote d'Afrique de Ceuta au Senegal, 1506-7 (translated P. de Cenival and T. Monod), (Paris, 1938), 85, 87.

19 D. P. Pereira, Esmeraldo de Situ Orbis (translated G. H. T. Kimble) (London, 1937), 80-1.

20 R. Mauny, Tableau Geographique de I'Ouest Africain au Moyen Age, Memoire de PInstitut Fondamental d'Afrique Noire, no. 61 (Dakar, 1961), 47, 499.

21 L. Africanus, The History and Description of Africa (translated J. Pory) (London, 1896), 1, 12. Using this reasoning, however, it is difficult to see how Africanus could have overlooked the fact of the Niger's eastward flow at Timbuktu.

82 This is an excellent illustration of the problem of toponymy in tracing the historical sources for Jenne. Medieval writers and cartographers used a bewildering number of spellings - Gyna, Gyni, Genni, Ghinea, Genehoa - which are sufficiently similar that it is often unclear whether reference is being made to a city (Jenne), a kingdom (possibly Jenne or ancient Ghana), or an entire region (Guinea - the Land of the Blacks). Flemish and Portuguese maps of the fifteenth and sixteenth centuries show an area variously designated as Gine, Guinea, Genoia, and Genna. These apparently are all variants of a generic term for Black Africa. During the early seventeenth century the term Genehoa began to appear with the term Guinia on European maps, such as that by Jodicus Hondius in 1606. The former term refers to both a city and a kingdom or region, whereas the latter

https://www.cambridge.org/core/terms

https://doi.org/10.1017/S0021853700018983

https://www.cambridge.org/core

6 RODERICK AND SUSAN McINTOSH

greate village onely', but then goes on immediately to describe the settlement as the capital of a large kingdom. Not only does the capital boast a highly stratified population, including a prince, priests, doctors, lawyers, and merchants, but it also is involved in a voluminous trade along the Middle Niger to Timbuktu.23

In general, the period from the mid-sixteenth to the early nineteenth century is virtually devoid of any European sources offering precise informa- tion on Jenne's location, physical description, or historical development. This gap in the European sources is filled by three major African sources, all specifically concerned with the Western Sudan and the upper Niger Bend: Ta'rikh al-Sudan, Ta'rikh al-Fettash and Tadhkirat an-Nisyan. Of these the Ta'rikh al-Sudan, completed by al-Sa'di in 1655, is by far the richest in detail concerning Jenne. Although his style when describing Jenne is a bit florid ('This city is large, flourishing and prosperous; she is blessed by Heaven and favoured by the Almighty'),24 the role al-Sa'di accords Jenne in the long-distance trade networks agrees well with Fernandes' earlier description:

Jenne is one of the great markets of the Muslim world. There the salt merchants of Teghaza meet merchants carrying gold from the mines of Bitou... Because of this blessed city, caravans flock to Timbuktu from all points of the horizon.25

Al-Sa'di goes on to describe Jenne's productive and well-populated hinter- land, albeit with some degree of exaggeration:

The territory of Jenne is fertile and populous; numerous markets can be found there every day of the week. Be assured that it contains 7,077 villages closely spaced.26

He claims that the territory of Jenne extends from Lake Debo in the north to the Volta Bend in the south, and borders on the Bandiagara highlands to the east.27 It is not clear whether Jenne's territory was defined by political suzerainty, economic domination, or some other means entirely.

In 1828, Rene Caillie became the first explorer to reach Jenne and subsequently provide a detailed account of the city.28 Caillie reported that Jenne was a city approximately 2 | miles in circumference, surrounded by an ill-constructed wall ten feet high and fourteen inches thick, pierced by several small gateways; it housed a population of 8,000-10,000 Mandingo, Fulani, Bambara, and Moors.29 By far the most important details in Caillie's account of Jenne concern the city's role in commerce along the Middle Niger. We learn that in 1828 Jenne was no longer the central point in the gold trade, owing to a westward re-direction of the trade as a result of wars between the Fulani of the Inland Delta and the Bambara of Segou.30 Because of its is apparently a more generic term for the whole of West Africa. By the end of the seventeenth century both the city and the kingdom of Genehoe (Jenne?) were always indicated in addition to the generic term Guinea. See Prussin, Architecture, 76, 276.

23 Africanus, History, 822. 24 al-Sa'di, Ta'rikh es-Sudan (translated O. Houdas) (Paris, 1900), 22. Translated from

the French by S.K.M. 26 Ibid. 22-3. Translated from the French by S.K.M. 26 Ibid. 24. Translated from the French by S.K.M. 27 Ibid. 25. 28 R. Caillie, Travels Through Central Africa to Timbuktu and Across the Great Desert

to Morocco: Performed in the Years 1824-1828 (London, 1830). 29 Ibid. 1, 459-60. 30 Ibid. 1, 465-6; Mungo Park also mentioned that political instability in the region had

disrupted trade - M. Park, Travels in the Interior of Africa (London, 1907). terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0021853700018983 Downloaded from https://www.cambridge.org/core. Access paid by the UCSB Libraries, on 10 Oct 2021 at 05:32:03, subject to the Cambridge Core

https://www.cambridge.org/core/terms

https://doi.org/10.1017/S0021853700018983

https://www.cambridge.org/core

JENNE-JENO AND THE INLAND NIGER DELTA 7

profitable trade in staples with Timbuktu, however, Jenne continued to prosper despite the decline in the gold trade:

Timbuktu, though one of the largest cities I have seen in Africa, possesses no other resources but its trade in salt, its soils being totally unfit for cultivation. The inhabitants procure from Jenne everything requisite for the supply of their wants, such as millet, rice, vegetable butter, honey, cotton, Soudan cloth, preserved provisions, candles, soap, allspice, onions, dried fish, pistachios. Bousbehey and Toudeyni being only supplied with the grains which the merchants of Timbuktu receive from Jenne would of course be reduced to famine if the trade between these latter two cities should be interrupted.31

Caillie went on to describe how agricultural produce, cloth, dried fish and slaves were bulked at Jenne, where they were brought by individual producers or small volume traders.32 At Jenne, staples bound for Timbuktu and points north were loaded on to pirogues, of which there was a large commercial variety measuring 90—100 feet long by 12-14 feet wide, drawing 6—7 feet water depth, and capable of carrying 60-90 tons.33

Through Caillie's account, Jenne appears as a prosperous commercial centre for regional trade based on a highly productive, diversified, and well-populated regional hinterland. The city's dual role as both an entrepot in the long-distance gold and salt trade and a major production and bulking centre for regional commerce in staples is revealed. We now see that al-Sa'di's claim that Jenne is the cause of Timbuktu's commercial success was by no means idle flattery.

There were no further eyewitness accounts of Jenne until French conquest in 1893, although the city figures prominently in traditions of the theocratic state of Seku Amadou in the first half of the nineteenth century.34 It is only with the French conquest that Jenne's correct geographical position on a tributary linking the Niger and Bani Rivers finally appears on a map. Thus was ushered in a period of intense interest in Jenne, highlighted by Dubois' account of his visits to Jenne and Timbuktu35 and Monteil's classic monograph on the city.36 However, French conquest also initiated the present decline in the city's commercial status in favour of Mopti, the port town at the Bani-Niger confluence with year-round access by motorable road.

Summary of historical documentation

It is in many ways extraordinary that the Western world remained so completely ignorant of Jenne for so many centuries. In striking contrast to its sister-city, Timbuktu, which was well-known, the sources discussed above indicate that the city was almost a total unknown prior to the nineteenth century. Part of the reason for Jenne's obscurity is surely, as L. Prussin37 and

31 Caillie, Travels, 11, 57-9. 32 Ibid. 11, 202-3; ' n Vol. 1, 435-8 Caillie suggests that certain villages specialize in the

production of particular commodities. 33 Ibid, ii, 59. 34 A. H. Ba and J. Daget, L'Empire Peul du Macina (1818-1853) (Paris, 1962); W. A.

Brown, 'The Caliphate of Hamdullahi c. 1818-1864: A Study in African History and Tradition' (Ph.D. thesis, University of Wisconsin, 1969).

35 F. Dubois, Timbuctoo the Mysterious (translated D. White) (London, 1897). 36 Ch.Monteil,MonographiedeDjenne(Tu\\e, i9O3);Ch.Monteil, UneCiteSoudanaise:

Djenne (Paris, 1932). 37 Prussin, Architecture, 77.

https://www.cambridge.org/core/terms

https://doi.org/10.1017/S0021853700018983

https://www.cambridge.org/core

8 RODERICK AND SUSAN McINTOSH

E. Herbert38 suggest, that most European and Arab traders and explorers approached the Western Sudan from the north, for which Timbuktu was the natural terminus from the fourteenth century onwards. The most detailed information about Jenne is furnished by those few travellers and merchants who either approached Jenne from the south (e.g. Caillie) or obtained information from the Guinea coast (Fernandes and Pacheco Pereira) from Dyula (Wangara) traders for whom Jenne was the northern terminus of the trans-savanna trade routes.

The available written documentation provides a fairly consistent picture of the city's function within the trade networks of the second millennium A.D. Jenne's prosperity depended on its dual role as a critical transhipment point for long-distance trade, and as a major production and exchange centre for regional commerce in staples and raw materials. Two factors appear to have been critical to Jenne's development: its situation at the most south-westerly point of the navigable Inland Niger Delta, and its extraordinarily productive agricultural hinterland. Although the historical documents give us no direct information on Jenne's existence before the fifteenth century, we have suggested that neither do they prove that the town was not an important trade centre prior to that time. For direct references to the town's origins and growth, we must turn to the oral traditions.

Oral tradition

Unlike oral traditions in many other Mande-speaking parts of West Africa, Jenne's traditions are fragmentary. In the sense of verbally related and retained chronologies of events passed on from one generation to another in the form of minstrel epics (like those recited by Bambara griots), oral traditions are essentially non-existent at Jenne. The city has no griot,3S and it has no unified coherent tradition of its pre-Islamic history. This is understandable in view of the 'multi-ethnic, polygot, heterogeneous charac- ter' of the city.40 Six or more different ethnic groups co-exist in Jenne41

without relinquishing their individual sense of ethnic and historic identity. As no one group is dominant, no one group has become the guardian of the city's history.

An additional factor which undoubtedly compromised the integrity of earlier oral tradition was the growing influence of Islam and literacy in Jenne. Native Muslim scholars, such as al-Sa'di, transcribed some of the local oral traditions into Arabic, although the pagan past was always discounted. As literacy and Islam spread, the verbal transmission of tradition may have diminished in frequency and importance, especially if the oral traditions were subverted by state-approved standardized histories like those created by Seku Amadou in the nineteenth century.42 Jenne's oral traditions have clearly undergone a good deal of chronological telescoping and political editing through the centuries. However, those details which have survived these

38 E. Herbert, 'Timbuktu', 450. 38 D. Robinson, personal communication (1978). 40 Prussin, Architecture, 94. 41 Bambara, Bobo, Nono (Marka), Bozo (Sorogo), Fulani and Songhay. 42 Brown, Caliphate, 4-7; M. Johnson, 'The economic foundations of an islamic

theocracy - the case of Masina', J. Afr. Hist, xvn (1976), 484.

https://www.cambridge.org/core/terms

https://doi.org/10.1017/S0021853700018983

https://www.cambridge.org/core

JENNE-JENO AND THE INLAND NIGER DELTA 9

processes provide an alternative to slavish dependence on historical docu- mentation for information on Jenne's chronology.

Three sources furnish the bulk of what is known of Jenne's oral traditions: Ta'rikh al-Sudan, Monteil's Monographie de Djenne, and Delafosse's Haut Senegal-Niger. All accounts agree that Jenne was founded in the eighth century A.D., that the city was first situated on the site of Zoboro, and subsequently moved to its present location. The Ta'rikh al-Sudan provides the following account:

In the beginning, the town had been built in a place called Zoboro; later it was moved to the spot where it still exists today. The former town was situated near the modern town in a southern direction... The town was founded by pagans in the mid-second century A.H....The inhabitants did not convert to Islam until towards the end of the sixth century A.H.43

The site of Zoboro is still well known to the inhabitants of Jenne; it is a mound situated three kilometres southeast of Jenne and also known as Jenne-jeno ('ancient Jenne' in Songhay), Jenne-sire, Djoboro, and Do-Djoboro.44

There is disagreement, however, over which city is indicated by the foundation date: Jenne or Jenne-jeno (Zoboro) ? Monteil45 believes that the eighth-century date refers to Zoboro, whereas Trimingham46 seems to accept it as the foundation date for Jenne.

Oral traditions collected by Monteil47 and Delafosse48 add flesh to the bare bones of al-Sa'di's report. Zoboro was initially a Bozo settlement.49 According to the traditions, Soninke-speaking Nono (Marka) entered the upper Inland Niger Delta from the northwest and superimposed themselves on the Bozo autochthones.60 The most commonly accepted derivation of the name Jenne is from the Nono 'Djanna' (as Jenne was originally pronounced), meaning 'Little Dia', thus commemorating the town from whence the Nono came.51

The duration of the Nono occupation at Jenne-jeno prior to moving the city to its present site is not specified in the oral traditions collected by Monteil and Delafosse, although both imply that the move had taken place before the conversion to Islam of the 26th chief of Jenne, Koi Kombora, between A.D. 1200 and 1300.62 In fact, all we can state with confidence is that Jenne-jeno must have been abandoned by at least A.D. 1468, at which time Sonni Ali garrisoned his troops there.53

Several historians have found it difficult to accept the eighth-century date claimed for the city by oral tradition. Delafosse54 felt that Jenne must have

43 al-Sa'di, Ta'rikh, 23. Translated from the French by S.K.M. 44 The tendency, demonstrated here, for different ethnic groups to assign place names

in their own language raises the possibility that Jenne-jeno was known by another name earlier in time. 46 Monteil, Cite, 33.

46 Trimingham, History, 31. But also see ibid., 63. 47 Monteil, Cite, 20-33. 48 M. Delafosse, Haut-Senegal-Niger (Paris, 1912), I, 253-70. 48 Monteil, Cite, 32; Delafosse, Haut-Senegal-Niger, 1, 263. 60 Monteil, Cite, 31. 61 The Nono originally founded Dia, on the edge of the Inland Delta, from which Nono

groups budded off to found new towns, including Sansanding, Diakha-sur-Bafing and Jenne. Monteil, Cite, 31-2; Monteil, Monographie, 263, 286.

68 Monteil, Monographie, 285; Delafosse, Haut-Senegal-Niger, 1, 269-70. 63 al-Sa'di, Ta'rikh, 26. 64 Delafosse, Haut-Senegal-Niger, 1, 269-70.

https://www.cambridge.org/core/terms

https://doi.org/10.1017/S0021853700018983

https://www.cambridge.org/core

IO RODERICK AND SUSAN McINTOSH

been founded c. A.D. 1250 as a direct consequence of the Soninke dispersal following the destruction of Ghana in c. 1240 by the Mansa of Mali. The Nono, he argued, would have entered the Inland Delta in great numbers at this time.

Levtzion56 discounts both of al-Sa'di's dates (c. A.D. 800 for Jenne's foundation; c. 1200 for conversion to Islam) as 'much too early'. He argues that Jenne owed her existence to two things: Timbuktu, and the opening of the Akan goldfields (usually believed to have occurred in the fourteenth century although this is by no means proven). Timbuktu was founded only in the early twelfth century and increased slowly in prosperity until it became very prominent in the fifteenth century. Levtzion reasons that the close commercial relations between Timbuktu and Jenne suggest that these two important market towns developed simultaneously ;56 thus, Levtzion supports Delafosse's date of A.D. 1250 for the founding of Jenne.

Triaud prefers to see Jenne as an artifact of Islam, founded near the end of the thirteenth century by Muslims in co-operation with indigenes. He dismisses the oral traditions as pure fabrication on the part of the inhabitants of Jenne who, 'suffering from the late creation of their town, needed to fabricate a history that gave them more ancient origins'.57

Despite decades of commentary on the chronology of Jenne's origins, Monteil's statement that 'we know nothing of the first centuries of Jenne'58

was still appropriate to the state of knowledge in 1977. The oral tradition versus historical reconstruction debate was clearly stalemated: archaeological research offered the most promise for resolution of this issue. This was one of several questions we sought to answer by excavating at Jenne-jeno in 1977.

The 1977 excavations at Jenne-jeno

The site of Jenne-jeno is a roughly teardrop-shaped mound which rises up to 8 metres above the floodplain of the Inland Niger Delta approximately 3 kilometres southwest of the present-day city of Jenne ( i3°53'2o"N, 40 32' 25" W). The mound rests in an abandoned channel of the Bani River (Fig. i), which now flows four kilometres to the southeast of Jenne-jeno; it is surrounded by floodwaters from September through December annually. The total surface area of the site is 330,810 square metres (33 hectares or 82 acres) and features visible on the surface include mudbrick house foundations, funerary urns, concentrations of iron and slag, and segments of a city wall (Fig. 2).

Excavations were carried out at Jenne-jeno from early February to late April 1977 as part of a programme of archaeological research designed to provide a preliminary framework for interpreting the prehistory of the western Inland Niger Delta. A major goal of the excavations was to establish a stratigraphic, ceramic, and radiocarbon chronology for the site which could then be applied to the investigation of other sites in the vicinity of Jenne. Other issues of interest included the chronology of technological and

66 Levtzion, Ancient Ghana, 159. 69 Levtzion, Ancient Ghana, 157; also Herbert, 'Timbuktu', 433. " J. L. Triaud, Islam et Societes Soudanais au Moyen Age, Recherches Voltaiques no.

16 (Paris, 1973), 128. Translation from French by S.K.M. 68 Monteil, Cite, 37.

https://www.cambridge.org/core/terms

https://doi.org/10.1017/S0021853700018983

https://www.cambridge.org/core

JENNE-JENO AND THE INLAND NIGER DELTA I I

^JENNE W * . V I C I N r r Y

Fig. i. Location of Jenne and Jenne-jeno in the Inland Niger Delta of West Africa.

economic change — for example plant domestication and the introduction of iron and, most relevant to the present article, the possibility of early urban development in the Inland Niger Delta. Since much of the information we hoped to extract from the site was of a chronological nature, vertical excavation units measuring three metres on a site were used, and rigorous stratigraphic control was maintained. Two excavation units, M i and M2, were dug 110 metres apart in a central area thought to have been residential because of the numerous mudbrick house foundations visible on the surface. Excavation unit JF1 sampled a cemetery area near the southeast periphery of the mound. We felt that excavation of these units would permit a preliminary assessment of differences in the nature and depth of the deposits at various points on the mound. Clearly, however, the possibility of sample bias is so great in these excavations that extreme care must be exercised in extrapolating from the excavated data to the unexplored areas of the site.

Summary of excavation results

In the two central excavation units (M 1 and M2), well-stratified, continuous cultural deposits extended to a depth of 5 metres, at which point sterile floodplain alluvium was encountered, thus confirming our initial impression that Jenne-jeno is a tell. In both units, iron and slag were present in the earliest levels, indicating that the original settlers at Jenne-jeno knew the use and manufacture of iron. A domestic function throughout the occupation is suggested for M 1 by the presence of large amounts of domestic debris (animal bones, broken pottery, etc.), charcoal and ash concentrations that may have

https://www.cambridge.org/core/terms

https://doi.org/10.1017/S0021853700018983

https://www.cambridge.org/core

12 RODERICK AND SUSAN McINTOSH

Contour interval 0-5 m

- " " - ' 0-25 m Contour

cm 1 Vertical datum

Not to

scale

Depression

'.".'. City wall

• Excavation unit

» Cement marker

. Funerary urn

o Round house; cylindrical brick

4 Square house; cylindrical brick

I n Square house; square brick

i V House, uncertain shape; cylindrical brick

0 SO 100 metres 200

1 . . . . I . I . I . I

Fig. 2. Topographic map of Jenne-jeno, Mali.

been hearths, and mud brick and tauf (coursed mud) structural remains. The residential character of M 2 is less certain, however. Several mudbrick house foundations were uncovered within the upper 15 metres of M2, but associated deposits failed to reveal the abundant domestic refuse and hearth deposits so characteristic of M 1. The recovery of several unusual classes of artifacts, including a terracotta statuette59 and pottery with raised relief

59 R. J. Mclntosh and S. K. Mclntosh, 'Terracotta statuettes from Mali', African Arts, XII (i979), Si-3.91.

https://www.cambridge.org/core/terms

https://doi.org/10.1017/S0021853700018983

https://www.cambridge.org/core

JENNE-JENO AND THE INLAND NIGER DELTA 13

representations of human and animal figures, from deposits inside the M 2 mudbrick structures may indicate a ritual rather than domestic function during part of the occupation.

On the southeastern periphery of the mound, JFi provides information on another functionally distinct area: the cemetery. Burial practices for both male and female adults comprised inhumation in large funerary urns covered with either a pot-lid or an inverted pot, or regular inhumation outside urns. Remains of eleven individuals, seven of whom had been interred in urns and four regular inhumations, were excavated from J F i . There were no accompanying grave-goods, other than an occasional broken pot and a sliver of a copper ring. Cultural deposits in JF1 extended to a depth of 15 metres.

A gross chronology for the excavated material from Jenne-jeno was established by means of stratigraphic analysis, pottery seriation, and radio- carbon dating. Each aspect of the enquiry was pursued independently, and the conclusions of each were then compared. One of us (R.J.M.) carried out stratigraphic analysis of the two central excavation units (M 1 and M 2). These units were located only n o metres apart and exhibited several marked stratigraphic similarities. Three corresponding sets of levels (bench levels) were identified with M 1 and M 2, based on the virtually identical texture and colour of deposits. It is thought that each pair of bench levels was laid down contemporaneously in M 1 and M 2 by the same depositional process or event (Fig. 3)-

A study of ceramics from Jenne-jeno60 revealed that the popularity of certain pottery attributes, including carinated form, certain types of twine decoration, painted decoration, and sand tempering, exhibited consistent, unimodal change through time, such that a 'battleship curve' or some segment of it was formed when the relative frequencies of these attributes were plotted through time (represented by stratigraphic ordering). Of course, the reliability of the pattern of change thus obtained depends heavily upon the amount of stratigraphic control maintained during excavation. For this reason M 1 served as the standard for ceramic seriation because the degree of correspondence between excavated levels and natural strata was high. Attributes which proved to be time sensitive in M 1 also exhibited battleship curves for the pottery from M 2 and JF 1. The conclusions about contem- poraneous bench levels in M 1 and M 2 were confirmed by the results of pottery seriation: the pottery from each pair of levels identified as contemporaneous on the basis of stratigraphic analysis proved to have virtually identical relative frequencies of time-sensitive attributes.

The independent conclusions drawn by the two authors about the relative chronology of the three major excavation units were confirmed by an internally consistent series of nine radiocarbon dates which also served to anchor our relative chronology in calendar years.61 Eight of the nine Jenne-jeno radiocarbon dates come from M 1 and M 2; they are displayed in Fig. 3, which indicates their stratigraphic position in relation to the M i and M2 bench

80 The pottery from Jenne-jeno and methods used in its analysis are fully described in S. K. Mclntosh and R. J. Mclntosh, Prehistoric Investigations in the Region of Jenne, Mali. Cambridge Monographs in African Archaeology, no. 2 (Oxford, 1980), part i.

61 In discussing radiocarbon dates, we follow the convention of using ' b.c.' and ' a.d.' when speaking of simple, uncalibrated dates, and ' B . C ' and 'A.D.' when dates are expressed in calendar years after calibration.

https://www.cambridge.org/core/terms

https://doi.org/10.1017/S0021853700018983

https://www.cambridge.org/core

M l

M 2

M od

er n

L ev

el 1

2 —

a. d.

5 20

± 1

Ph as

e IV

c. A

.D .

90 0-

14 00

Ph as

e H

I c.

A .D

. 4

0 0

-9 0

—

ad .

11 20

± 1

—

a. d.

1 03

0 ±

15 0

Le ve

l 12

Ph as

e II

c. A

.D .

5 0

-4 0

Ph as

e I

25 0

B .C

-A .D

.5 0

Le ve

l 18

Fi g-

3 -

Ph as

es o

f oc

cu pa

tio n

at J

en ne

-j en

o, i

llu st

ra tin

g st

ra tig

ra ph

ic b

en ch

l ev

el s

(d ep

os ite

d co

nt em

po ra

ne ou

sl y

un de

r si

m ila

r co

nd iti

on s)

a nd

r ad

io ca

rb on

d at

es (

de pt

hs a

t w

hi ch

s am

pl es

w er

e ta

ke n

de no

te d

by t

ri an

gl es

o D z o en C z H o en X

https://www.cambridge.org/core/terms

https://doi.org/10.1017/S0021853700018983

https://www.cambridge.org/core

JENNE-JENO AND THE INLAND NIGER DELTA 15

levels discussed above. The dates suggest that habitation of the site began around 250 B.C. and continued through to the twelfth century A.D. at least. Gradual abandonment of the site was probably in progress soon thereafter, with complete abandonment accomplished by A.D. 1468, at which time the site was garrisoned by the Songhai conqueror Sonni Ali during his siege of Jenne.62 Thus the date RL-794 (modern) from level 4 of M 1 cannot be accepted, even in calibrated form (> A.D. 1530) because doing so would imply that the top 60 centimetres of M 1 deposits accumulated in fairly recent times. We believe A.D. 1400 to be a reasonable terminus ante quern for the abandonment of Jenne-jeno, although it remains possible, given the nebulous state of our current evidence concerning the end of occupation, that the abandonment process was completed as early as A.D. 1200.

Armed with confirmatory information from all three aspects of the investigation concerning temporal relationships of the three major excavation units, a framework for further analysis was created by arbitrarily dividing the deposits into four occupation phases. The definition of each phase was based on the identification of discrete series of stratigraphic events as well as on temporally significant changes in pottery and other cultural characteristics. The chronological outline in calendar years for the four phases is largely derived from the radiocarbon dates, although it must be noted that the standard deviation for most of the Jenne-jeno dates is greater than 100 years, which means that the phase chronology is very approximate indeed. This chronology may be altered by future excavations at Jenne-jeno, as may the defining characteristics of each occupation phase presently outlined; the four occupation phases were conceived merely to facilitate the ordering, analysis and description of the data recovered during the 1977 excavation season.

The early occupation of Jenne-jeno

The earliest phase of occupation at Jenne-jeno lasted from approximately 250 B.C. to A.D. 50. Phase I is currently defined by the absence of permanent mud architecture and predominance of finely made, sand-tempered pottery of Saharan affinity. The first occupants at Jenne-jeno made and used iron, and had a subsistence base that was heavily reliant on aquatic resources, including fish, tortoise, crocodile and waterfowl, and on various bovids.63 Although a definite identification could not be made owing to the high degree of post-depositional fragmentation, it is probable that the bones of a small bovine present in the faunal remains belong to domestic cattle (Bos taurus).™ No evidence of domesticated plants was recovered from Phase I deposits in 1977, but the possibility of sampling error should not be overlooked.

The lower boundary of Phase II is provided by a level of mud wall collapse in one of the central excavation units. Associated charcoal yielded a radiocarbon date of a.d. 40 + 50. From the same level we recovered a substantial sample of well-preserved chaff of Oryza glaberrima. This is the earliest occurrence

62 al-Sa'di, Ta'rikh, 26. 83 Including waterbuck, Redunca redunca. A discussion of the subsistence information

from Jenne-jeno may be found in S. K. Mclntosh and R. J. Mclntosh,' Initial perspectives on prehistoric subsistence in the Inland Niger Delta (Mali)', World Archaeology xi (1979), 227-43.

64 R. G. Klein, personal communication (1978).

https://www.cambridge.org/core/terms

https://doi.org/10.1017/S0021853700018983

https://www.cambridge.org/core

l6 RODERICK AND SUSAN McINTOSH

of African rice yet known. Other economic elements show a great deal of continuity with Phase I; faunal remains continue to be dominated by fish and bovids.

According to the pottery chronology developed and successfully applied to excavation units M i and M2, JF i deposits began to accumulate during Phase II. Over one metre of deposits appears to have accumulated some time between A.D. 50 and 400, based on the relative frequencies of attributes of pottery taken from these levels. The nature of the lowest JF1 deposits suggests that they consist of material eroded from higher areas of the Phase II settlement. The fact that Phase II deposits were present in all three major excavation units (Mi , M2 and JFi ) , separated by a maximum distance of 400 metres, indicates that the Phase II settlement measured at least this distance along one axis. This apparent increase in site size by Phase II may possibly be correlated with the appearance, at the beginning of the phase, of permanent mud architecture and domesticated rice. On the other hand, their seemingly simultaneous appearance may merely be an accident of our extremely small excavation sample. Only further excavation can resolve this question. Also to be clarified by future excavation are the rate of physical expansion of Jenne-jeno in Phase I, and the areal extent of the mound at the beginning of Phase II. On the basis of present evidence, however, it seems reasonable to suggest that Jenne-jeno was growing in size throughout Phase I and was a settlement of considerable dimensions, possibly exceeding 10 hectares, by Phase II.

The urban expansion and subsequent decline of Jenne-jeno

There are indications that the occupation of Jenne-jeno became increasingly intensive in Phase III (A.D. 400—900). An increasing number of wall stumps is visible in section, and the stratigraphy of Phase III deposits becomes noticeably more complex. In M 1, wall stumps are associated with fairly deep deposits, suggesting that occupation of a house could span an appreciable time span, perhaps several centuries. The appearance of crowded cemeteries may be further evidence of population growth and intensified site occupation. All four inhumation burials and six of the nine urn burials in JF 1 appear to have been interred during Phase III, based on the ceramic material contained in the cemetery deposits. The significance of a single radiocarbon date of a.d. 1130+180 (RL-950) on bone taken from one of the urns cannot be assessed at the present time. Bone tends to yield radiocarbon dates that are too young, and this date is younger than expected from the pottery chronology. The date does, however, fall within the expected Phase III time limits when two standard deviations are considered.

Jenne-jeno probably reached her greatest areal extent in late Phase Ill/early Phase IV. Thereafter, the site slowly began to decline. At her apogee, perhaps from A.D. 750 to 1150, Jenne-jeno covered 33 hectares (the present city of 10,000 inhabitants covers 44 hectares). To this can probably be added the 9-hectare expanse of Hambarketolo, a mound immediately adjacent to Jenne-jeno and actually connected to it by an earthen dike (Fig. 1). Pottery from the surface of Hambarketolo indicates that it too was abandoned during Phase IV. Deposits are shallower at Hambarketolo than

https://www.cambridge.org/core/terms

https://doi.org/10.1017/S0021853700018983

https://www.cambridge.org/core

JENNE-JENO AND THE INLAND NIGER DELTA 17

at Jenne-jeno, achieving a maximum depth of approximately 3 metres. This may mean that Hambarketolo was founded later than Jenne-jeno and functioned as part of the city at the time of its fluorescence.

The causes of the decline and ultimate abandonment of Jenne-jeno and Hambarketolo are unknown. As meagre as the data for the chronology of this process are, it nevertheless appears from the archaeological evidence that the beginnings of site abandonment coincide closely with the thirteenth-century date for Jenne's islamization reported by al-Sa'di. It is possible that a new, Islamic Jenne was founded on a new site unpolluted by pagan practices shortly after the king of Jenne(-jeno) or perhaps the community's elite converted to Islam. If, as there is every reason to believe from the historical records, commercial control became vested in the hands of native, Muslim merchants, the new Jenne would soon replace pagan Jenne-jeno as the hub of economic activity. Farmers, fishermen and craftsmen, who undoubtedly comprised the majority of Jenne-jeno's population, would have been attracted to Jenne because of her growing market for their produce. Our hypothetical reconstruction of the decline of Jenne-jeno sees economic incentives at the new, Muslim market centre of Jenne as the primary cause of Jenne-jeno's abandonment. This process was complete by at least the mid-fifteenth century.

The development of Jenne-jeno's rural hinterland

A large site like Jenne-jeno does not evolve in isolation. Although archaeo- logists tried for years to devise a list of traits by which urbanism could be positively identified at an archaeological site,65 no set of criteria was ever recognized as totally satisfactory or universally applicable. A more productive approach has been borrowed from geographers, who have found it useful to investigate the relationship between the urban centre and its surrounding hinterland. They have recognized an increase in the number and variety of services available as one progresses up the regional settlement hierarchy from the agricultural hamlet to the true town. Trigger, for example, defines a city as 'a unit of settlement which performs specialized functions in relation to a broader hinterland \68 The most effective way for the archaeologist to study the emergence and principles of urbanization in a region, therefore, is not only to excavate the culmination of the process (i.e. the largest site in the region) but also to attempt to identify a hierarchy of settlements in which the direction of complexity is revealed by increasing heterogeneity of remains (artifacts and features) and often by size.87 The results of a field survey covering a i,ioo-square-kilometre area of Jenne-jeno's hinterland provide useful insights into the development of this region.

Archaeological sites within a 2 5-kilometre radius to the north and west of

86 Most notably V. G. Childe, 'The Urban Revolution', The Town Planning Review xxi (1950). 3-17-

" B. Trigger, 'Determinants of urban growth in pre-industrial societies', in P. J. Ucko, R. Tringham, and G. W. Dimbleby (eds.), Man, Settlement, and Urbanism (London, 1972). 577-

" R. J. Mclntosh, 'The Development of Urbanism in West Africa: The Example of Jenne, Mali' (Ph.D. thesis, University of Cambridge, 1979), 74-7.

https://www.cambridge.org/core/terms

https://doi.org/10.1017/S0021853700018983

https://www.cambridge.org/core

l8 RODERICK AND SUSAN McINTOSH

Jenne-jeno were located and selected for investigation on an explicit sampling basis employing probability theory;68 the approximate date of abandonment for each site investigated was determined by comparing surface pottery collections with the Jenne-jeno ceramic chronology. One of the startling conclusions of the survey was that, of 42 sites investigated in the course of the sample survey, almost three-quarters had been abandoned by the end of Phase IV (A.D. 1400), and none was abandoned before Phase III.69 Almost all sites investigated are tells with deposits exceeding 2 metres in height above the floodplain. Such an accumulation probably represents several centuries of occupation prior to abandonment. Surface remains on several of the largest sites in the immediate vicinity of Jenne-jeno indicate a maximum extent of occupation during Phase III and a reduction in area of more recent deposits before abandonment at a later time. These results, preliminary as they are, imply that site density in Jenne-jeno's hinterland reached its greatest expression in late Phase Ill/early Phase IV, and thereafter began to decline. At its maximum, site density in the survey region may have been as great as ten times the density of occupied settlements in the same region today.

Most of these sites were quite small: 26 per cent are smaller than the smallest presently occupied village and the mean size of 0-95 hectare for sites compares with a mean size of 198 hectares for modern villages. Five sites exceed 10 hectares in size and one nearby site, Kaniana (Fig. 1), even exceeds the area of Jenne-jeno. In general, site complexity, as measured by the number of different artifact and feature classes noted in careful examination of each site's surface, varies directly with site size. Ceramics and features resemble those at Jenne-jeno to a remarkable extent. There is strong indirect evidence, then, that sites in Jenne-jeno's rural hinterland were integrated in a settlement hierarchy with Jenne-jeno as its apex. It must be stressed, however, that in the absence of refined chronological control for sites in a region, direct and satisfactory evidence for the nature and integrative principles of a settlement hierarchy is elusive. Sites in Jenne-jeno's hinterland form a distribution by size which would be recognized by geographers as conforming to a 'rank-size rule';70 however, it is also possible that this distribution is in reality a palimpsest of various non-hierarchical settlement clusters.

The survey results accord well with the excavation results: both indicate that Jenne-jeno and her hinterland were developing rapidly during Phase III (c. A.D. 400-900). Site density reached a maximum in the hinterland in late Phase Ill/early Phase IV, at approximately the same time that Jenne-jeno

68 The sampling methodology is explained fully in S. K. Mclntosh and R. J. Mclntosh, Prehistoric Investigations, part ii. Tells are highly visible on aerial photographs of the survey region; 404 sites were identified by this method alone.

69 The results of archaeological investigations by R. Bedaux et al. at two sites in the Inland Delta provide some support for these conclusions. Both sites examined by Bedaux appear to have been first occupied in the eleventh century A.D. Radiocarbon dates at the site of Doupwil, approximately 90 kilometres northeast of Jenne, indicate that it was abandoned in the fifteenth century A.D.; Galia, 12 kilometres east of Jenne, was probably abandoned within two centuries after Doupwil. R. M. A. Bedaux, T. S. Constandse- Westermann, L. Hacquebord, A. G. Lange and J. D. van der Waals, 'Recherches archeologiques dans le Delta Interieur du Niger', Palaeohistoria xx (1978), 91-220.

70 G. A. Johnson, 'Aspects of regional analysis in archaeology', Annual Review of Anthropology VI (1977), 479-508.

https://www.cambridge.org/core/terms

https://doi.org/10.1017/S0021853700018983

https://www.cambridge.org/core

JENNE-JENO AND THE INLAND NIGER DELTA 19

achieved its maximum areal extent. Jenne-jeno apparently began to decline, and was ultimately abandoned, as part of a general population reorganization in the western Inland Niger Delta, manifested by abandonment of large numbers of sites on the floodplain. The reasons for this population shift are unknown, but the event appears to have preceded the political dislocations caused by migrations of Bambara and Fulani and by the demise of stable government in the region.

Conclusions and interpretations

The 1977 excavations have demonstrated that Jenne-jeno existed as early as 250 B.C., and was rapidly developing during the early centuries of this era into a town measuring a half-kilometre or so along one axis. By A.D. 900 or 1000, it appears likely that Jenne-jeno and neighbouring Hambarketolo formed an urban unit covering more than 40 hectares. It is likely, in our opinion, that the development of local and regional trade networks in the Inland Delta and adjacent areas was a factor in the early expansion of Jenne-jeno. Situated in an alluvial plain lacking in stone and iron ore, at the interface of two major vegetation zones (dry savanna and sahel), the site is a natural candidate for the early growth of inter-regional exchange. Indeed, Phase I deposits strongly indicate early involvement in trade. Not only are the stone beads from the earliest deposits foreign to the Inland Delta, but there is also evidence that the iron must have been obtained from outside the Delta, presumably by trade.71

To the best of our knowledge, there are no sources of iron ore in the Inland Niger Delta suitable for smelting. The presence of slag at Jenne-jeno and many other sites in the survey region suggests that either iron ore or bloomery iron72 with slag still adhering was imported into the central delta. The latter possibility seems the more reasonable, since transport costs would be appreciably reduced by removing the weight of useless silicates at the source area. Smelting in the Jenne area would seriously reduce the tree cover of the Inland Delta, which was probably never great. There is historical documen- tation that in the nineteenth and twentieth centuries Jenne imported bloomery iron or iron preliminarily forged into bars from the Benedougou region near San.73 If, as we suspect, the Benedougou is the nearest source of abundant iron ore to Jenne, the iron trade between the two areas is likely to be of considerable antiquity, perhaps dating back to the foundation of Jenne-jeno.

71 Although nodules, which we originally took to be limonite or haematite, are abundant on the alluvium, samples collected from the floodplain and from Jenne-jeno proved to be not ferric oxide nodules, but concretions of quartz and clay stained by ferric oxides. P. Jacobberger, the geologist who examined these nodules petrographically, concluded that they were completely unsuitable as a source of iron ore (P. Jacobberger, personal communication, 1979). In the course of regional survey, we noted no potentially ferruginous outcrops or lateritic crusts within the Delta proper.

72 The bloom is the spongy mass of metallic iron which results from smelting. In the smelting process, silicates (the principal impurities present in iron ores) form a glass-like slag. The worst of this slag can be detached from the bloom after smelting, but to achieve workable iron the bloom must be forged to help consolidate the metal and to squeeze out any slag remaining in it. H. Hodges, Artifacts (London, 1976), 83.

73 M o n t e i l , Monographie, 1 9 3 . C a i l l i e n o t e d t h a t J e n n e ' s i r o n w a s s m e l t e d in B e r e ( in the Benedougou), where there was a great deal of ferruginous stone on the surface of the soil. Bloomery iron from Bere was exchanged in Jenne for salt. Caillie, Travels, 1, 421.

https://www.cambridge.org/core/terms

https://doi.org/10.1017/S0021853700018983

https://www.cambridge.org/core

2O RODERICK AND SUSAN McINTOSH

Evidence that the exchange networks within which Jenne-jeno functioned had expanded by the fifth century A.D. is provided by copper ornaments in the earliest Phase III deposits. The three closest known sources of copper ore are in the Sahara: Akjoujt, Nioro, Air.74 This evidence strongly implies that inter-regional trade for Saharan copper existed by the mid-first millen- nium A.D.75 Other Saharan commodities carried in this early trade can only be guessed at: salt is a major possibility, given the historically documented demand in the Sudan for Saharan salt and copper during the second millennium. Unfortunately, salt trade is difficult to document archaeologically. If Saharan copper and salt were reaching Jenne-jeno in any quantity by the fifth century A.D., then it is reasonable to suggest that the navigable middle Niger was functioning as an important north-south transport axis by that date. The ease of riverine transport, at least seasonally, and the annual migration by pirogue of fisherfolk downstream to the Niger Bend76 would have made this a natural early development.

Commodities sought by Saharan traders in exchange for copper and salt may have been largely agricultural, including rice from the Jenne region and fruit or other savanna products from the Benedougou." Gold may also have been one of the savanna commodities moving along the exchange networks at this early time, although we have no direct evidence for this.78 This hypothetical reconstruction of the development of first millennium inter- regional exchange along the Niger emphasizes Jenne-jeno's role as a producer of surplus staples, such as dried fish, fish oil, and rice, which were potentially desirable to communities in the adjacent dry savanna and Saharan zones. Such a theory explains Jenne(-jeno)'s location more satisfactorily than do interpretations of the town's development which stress its excellent situation for riverine trade with the Niger Bend or its defensible location. Indeed, Jenne is perhaps not in the best location for trade down the Niger. The most direct waterway from Jenne to the Niger, the Souman-Bani, is too shallow for travel during four months of the year.79 Rather, Jenne(-jeno)'s location is optimal with respect to three different requirements: proximity to a richly productive agricultural hinterland; access to the Benedogou savanna highlands to the southeast; and riverine transport, throughout most of the year, to the Saharan contact zones at the Niger Bend. We suggest that Jenne-jeno's location at

74 Mauny, Tableau, 307. 75 This conclusion is supported by the radiocarbon dates of a.d. 550+ 100 and a.d.

640+100 for the apparently pre-Islamic copper industry at Marandet in the Ai'r. M. Posnansky and R. J. Mclntosh, 'New radiocarbon dates for northern and western Africa', jf. Afr. Hist, XVII (1976), 183.

76 The Bozo of the Jenne region penetrate beyond Lake Debo seasonally, following the migration of shoals of oil-rich Alestes or Nile perch. L. Sundstrom, Ecology and Symbiosis: Niger Waterfolk, Studia Ethnographica Upsaliensia no. 35 (Uppsala, 1972), 41, 134.

77 A list of agricultural and pastoral products of the floodplain or savanna nearby which were bulked at Jenne at the time of French penetration is provided in Monteil, Monographie, 41-62, 215-25.

78 The accessibility of Jenne to the Lobi goldfields should not be overlooked. Meniaud notes that 'Le Lobi est peut-etre la region du Haut-Senegal-Niger ou l'or est le plus abondant, le plus facile a exploiter, et, en tout cas, le plus beau'. J. Meniaud, Haut- Senegal-Niger (Paris, 1912), 11, 193.

79 For example, the pirogue in which Caillie travelled from Jenne to Mopti in March 1828 had to be off-loaded several times in order to clear sand congestion. Caillie, Travels, 11, 2. See Monteil, Monographie, 14-15, 26-7.

https://www.cambridge.org/core/terms

https://doi.org/10.1017/S0021853700018983

https://www.cambridge.org/core

JENNE-JENO AND THE INLAND NIGER DELTA 21

the southwestern extreme of the navigable and agriculturally productive Inland Niger Delta promoted its growth as a trade centre where Saharan commodities like copper and salt could be traded for dried fish, fish oil and rice produced in the Inland Delta, and where savanna products, including iron from the Benedougou, could be obtained with a minimum of overland travel in exchange for salt, copper, rice, fish and other staples.

The historically documented existence of trade centres by the eleventh century in the agriculturally impoverished Niger Bend area offers the first confirmatory evidence, however circumstantial, that the Inland Delta was firmly established as the agricultural support system for the Niger Bend. It is probable that sizeable settlements on the Niger Bend could not produce enough food locally to maintain their population, as in the case of Timbuktu. Al-Bakri mentions the towns of Safanku and Bughrat along the Niger route to Tirekka, located at the southward turning of the river:80 Tirekka's market attracted the merchants of Ghana and Tadmekka,81 the latter of which is described as 'a great town... better built than Ghana or Kawkaw. The inhabitants... import millet and all other cereals from the Land of the Blacks'.82

More direct evidence for Jenne-jeno's participation in inter-regional trade during the first millennium A.D. will have to come from archaeological research. The above explanation for the town's growth is merely an hypothesis which attempts to account in a reasonable way for the archaeological fact that Jenne-jeno existed throughout the first millennium A.D., that it was receiving copper, probably of Saharan origin, by the fifth century, and that it had grown into an urban centre of considerable proportions by A.D. 900 or 1000.

The conspicuous silence of the historical sources about the early town remains a puzzle. Might there be alternative names for the site present, but unrecognized, in the sources ? If the community were late to accept Islam would the Arab sources fail to mention it ? The historical sources were for so long dependent upon second-hand information passed along the long- distance commercial networks. If Jenne-jeno were until relatively late a principal participant in an independent, although ancient and voluminous intra-Inland Delta trade, the outside world may have remained ignorant. Or perhaps there was a successful policy by the merchants of Jenne, at least in the second millennium, to withhold information about the southern sources of gold and the wealth of the city itself. Trade in staples and luxury goods down the Niger was controlled in recent centuries by the merchants of Jenne; there is a mound outside Jenne called Wangaradaga, where the Dyula from the south, forbidden to enter the city, waited while subordinates of the chief of Jenne negotiated with the merchants of the city for the best price for their goods. The merchants of Jenne were in a position to control the flow in both directions of information, which, as speculators have known throughout time, is a critical commodity to trade.

80 al-Bakri, in Cuoq, Recueil, 106. 81 Cuoq, Recueil, 105; according to the Ta'rikh al-Sudan, Tirekka was populated by

Sanhaja merchants and it declined as T i m b u k t u prospered: Montei l , Cite, 33. 88 Cuoq, Recueil, 107. Trans la ted from the French by S . K . M . Elsewhere we have

argued that Jenne-jeno's agricultural surplus may have helped support the early urban centre of Kawkaw (Gao) - S. K. M c l n t o s h and R. J. Mc ln tosh , Prehistoric Investigations,

https://www.cambridge.org/core/terms

https://doi.org/10.1017/S0021853700018983

https://www.cambridge.org/core

22 RODERICK AND SUSAN McINTOSH

SUMMARY

The dates and circumstances of early references to Jenne have led historians to conclude that the city originated relatively late in time. It is widely believed that the city developed simultaneously with Timbuktu in the mid-thirteenth century as an artifact of trans-Saharan trade. Persistent oral traditions of the foundation of Jenne in the eighth century are generally discounted.

Recent archaeological excavations at the ancestral site of Jenne-jeno have established that iron-using and manufacturing peoples were occupying the site in the third century B.C. The settlement proceeded to grow rapidly during the first millennium A.D., reaching its apogee between A.D. 750 and 1100, at which time the settlement exceeded 33 hectares (82 acres) in size. The archaeological data are supported by the results of site survey within a 1,100-square-kilometre region of Jenne's traditional hinterland. During the late first millennium A.D., several nearby settlements comparable in size to Jenne-jeno existed, and the density of rural settlements may have been as great as ten times the density of villages in the hinterland today.

Evidence from excavation and survey indicates that Jenne participated in inter-regional exchange relations far earlier than previously admitted. The stone and iron in the initial levels at Jenne-jeno were imported from outside the Inland Delta; levels dated to c. A.D. 400 yield copper, presumably from distant Saharan sources. The importance of the abundant staple products of Jenne's rural hinterland, including rice, fish and fish oil, is examined in a reassessment of the extent of inter-regional commerce and the emergence of urbanism during the first millennium A.D. Jenne-jeno may have been a principal participant in the founding of commercial centres on the Saharan contact zone of the Bend of the Niger, rather than a product of the luxury trade serviced by those centres.

https://www.cambridge.org/core/terms

https://doi.org/10.1017/S0021853700018983

https://www.cambridge.org/core