INDIVIDUAL RESEARCH REPORTS 63. SOME REFLECTIONS ON AFRICAN TRADE ROUTES There will be four main factors in the recon- struction of pre-European African history south of the Sahara; the recording of oral history, the study of anthropology, archaeological research and geographical analysis, all four must always be co-ordinated, in some N instances they will blend. There is none of the northern antithesis between tradition and the written source for normally the few . chronicles are essentially segments of oral history re- corded at different periods in documentary form as in the case of the chronicles of Kilwa and of Pate. The fac- tual value of traditional history must always depend on the character of the social setting in which it has been preserved, adapted and at times created. In consequence the valid study of oral history is inseparable from scientific anthropology. But it is also interwoven with archaeological research whether the study of sites or of beads or of pottery forms; African tradition has often provided archaeology with its clues as in Uganda, Archeo- logy is providing the study of traditional history with a new technique for objective checking. But geographical analysis combined with a per- petual awareness of economic factor must give the set- ting for any reconstruction of the African past. Per- haps the most immediate need is to apply the evidence of oral tradition,' archaeology and geography to the re- discovery of pre-European trade routes. There are four main groupings of such routes to be investigated. There are the lines of coastal traffic along the shores of the Indian Ocean and of the Atlantic, and perhaps along the deserts of the Sahara which formed an #1 •r i 64. INDIVIDUAL RESEARCH REPORTS inland sea. There are the tracks by which exports from far into the interior reached the coast and were unti- mately forwarded to foreign markets. There is the net- work of internal trade within Africa brought into being by local needs for salt or iron or copper. Finally there is the possibility of transcontinental trade routes. And in each case we are only at the beginning of any scientific study. In this paper I am tentatively making suggestions based primarily on personal experience. \ The trade route of which most is known is that of the coastal traffic along the African shore of the Indian Ocean. It is relatively well documented by Graeco-Roman and Arab geographers. Its traditional his- tory has been recorded fully. An archaeological survey from the coasts of Somalia to those of the Transkei was finally completed in the spring of 1963. Yet even here there are still some elementary questions unanswered. How far south did this coastal traffic extend before the Portuguese rounded the Cape? Until the winter of 1962 I had assumed that it coincided, even if only roughly, with the area of the monsoon and that Sofala was the most southerly mainland calling-port to possess sig- nificance. But that November I worked on finds from the Mapungubwe site in the north Transvaal and they prove contact with Indian Ocean trade. Most of these are very difficult to date - "trade-wind" beads and glass frag- ments - but they include some sherds of Chinese porcelain, Yuan ware celadon, which are clearly thirteenth century. Mapungubwe is an obvious clearance centre for copper from the Messina workings and gold from the north Trans- vaal; and air photography shows that it is the most easterly of a line of sites which stretch into Bechuana- land. It seems likely therefore that there is a still undiscovered entrepot at the mouth of the Limpopo river. * A INDIVIDUAL RESEARCH REPORTS 65. It has long been known that But perhaps the Indian Ocean trade route went still farther south. I think that I found, evidence of this in 1963 during an archaeological survey of the shores of the Transkei. sherds of Ming porcelain have been reported from Pond- oland, now the the Transkei coast. It was supposed that they had come ashore from sixteenth century Portuguese wrecks. It is clear from their surface texture that this is true of about thirty per cent of the sherds that I have examined. But at four sites in Pondoland, notably round at St. John and at Msicaba, and in the museum at East London there are fragments of Ming blue and white of the fifteenth or early six- teenth century which show no signs of being sea-tossed. More conclusively f have found numbers of Gujerati car- nelian beads, characteristically'fifteenth century types and carefully facetted, which show no mark of erosion; there were about eighty or ninety of them at Port St. John alone, and I have been told of others still brought in to up-country stores. It seems at last tenable that in the fifteenth century Pondoland was visited by Is- lamic traders from the north as the most southerly catch* ment area for ivory. If in the fifteenth century there was a sea route from Somalia to the shores of South Africa were there transcontinental routes which would link this with the western coast? There are two possibilities; one leading from Kilwa in what is now Tanganyika to the old kingdom of the Congo and what was to be later San Salvador. The other from the Adan kingdom in what is now Somalia to northern Nigeria perhaps by way of Bornu. Traditions of the latter route are recorded by Burton. The first route is the only one that is yet documented. •-"?• • -# Mf.. • .J • •* 66. INDIVIDUAL RESEARCH REPORTS Ibn Battuta who visited Kilwa island in 1332 described an overland route that led from there to West Africa through the land of the Lam Lam'. No geographer, except Sir Hamilton Gibb, took this account very seriously; but this year it has received a rather surprising cor- roboration from two eighteenth-century French sources. Both were found by Dr. Freeman-Grenvilie and will be published in his forthcoming book The French at Kilwa Island. Morice writes in 1777 the fresh water sea or which I have spoken to you, is I estimate, a month's march from the ocean. This fresh water sea has a rise and fall of eight feet. It takes two days to cross it in a boat rowed by six oarsmen. On the other side lies a huge country which has been crossed by natives of the country, who, after two month's travelling, found the ocean and saw ships there manned by Europeans. There is every reason to believe that this is the coast of Angola. Cossigny added a few years later -> an African caravan leaves Kilwa every year and proceeds to the west coast. There is a great lake which they say is a fresh water sea, in the middle of which there is a large is- land. They cross this lake on pieces of wood and make a halt at the island. The Africans assert that at the end of the journey there is a salt sea. They find there vessels like ours and Europeans to whom they sell their slaves. The agreement in their reports does not permit of any doubt as to the truth of the fact. The fresh water sea is presumably Lake Nyasa. It is tenable that the original route wound its way from Kilwa to the old kingdom of the Congo. There is still the tradition of a trade path from Kilwa to the southern end of Lake Nyasa. There was a nineteenth-century report of a series of stone forts along it, but this has never been investiga- ted by an archaeologist. I once came across a possible clue to another transcontinental trade route. I never had the opportunity INDIVIDUAL RESEARCH REPORTS 67. to follow it up, so it is perhaps worth recording in some detail in case some other archaeologist should have the chance to do so. In the winter of 1950 the government of the British Somaliland Protectorate invited me to make an archaeological survey. During it I worked on a group of twelve deserted towns along the frontier of British Somaliland and Ethiopia and carried out preli- minary excavations at four of them. They are often of considerable extent; three of them, Amud, Abara and Gargesa, contain the remains of between 200 to 300 stone houses; in some cases the Walls still rise to eighteen feet. It seems clear that they represent vanished cities of the late medieval kingdom of Adal. I examined twenty- six silver coins from them; the latest were struck at Cairo by Sultan Qait-Bey (1468-89); the earliest, also from Cairo, were struck by Sultan Barquq (1382-99). All were from mints either at Cairo or at Damascus. For the purpose of this essay there are three points worth noting. First, that these sites can be plotted in a pattern of a rather devious trade path of short marches. Secondly, that I was able to link them with an Indian Ocean termi- nus at Sa'ad-al-Din island, four miles north east of the modern port of Zeila. Thirdly, that during nearly twenty years field work throughout Africa I have never come across such evidence of late medieval wealth. There is the use of a silver coinage, unlike the copper which alone proliferates round Kilwa. There are the occasional gold coins; none have been reported elsewhere in East Africa. There is the quite exceptional quality of the porcelain. Fine celadon was predominant in each of the sites that I examined, found sometimes on the surface, sometimes beneath the .surface to the depth of seven and a half inches, sometimes in the tight-packed middens that often rose to between four and five feet. There were sherds with a white crystaline body and green-white glaze without crackle, or glazed olive-brown; others with a grey granular body with a sea-green glaze or a blue-green glaze crackled all over. I also found a few e it- 68. INDIVIDUAL RESEARCH REPORTS fragments of Ming, red and white, and others of white porcelain with an opaque bluish tint. There was also a good quantity of early Ming blue-and-white; fragments of bowls ornamented with tendril scrolls on a slightly blued ground with black spotting at the end of the strokes, or of bowls with floral designs in strong out- lines of grey or black-blue. The most reasonable hypothesis would seem tOv be that I was working along a short segment of a wealthy trade path of the late fourteenth and fifteenth centuries that had one ultimate terminus in early Ming China and another northward, up the Red Sea, in Cairo. The signs- of almost excessive wealth suggest that this track led far into the interior of Africa and even possibly tapped a gold supply; gold and ivory were two imports needed in late medieval China. Documentary evidence may yet be found to show that one pilgrim route from Mali to Mecca bent south of the Christian bastion of fifteenth-century Ethiopia and followed some such path as this. There is a third possibility of transcontinental traffic which should be mentioned even if, as I suspect, later research discards it. The great medieval port of Aidhab lies far up the Red Sea, to the north of Suakin. It was destroyed in 1426. The height of its prosperity would seem to have been in the late fifteenth century. I have never been able to visit the site but I have worked on objects from it both in the British Museum and in the museum at Khartoum. They were oddly similar to those from the Adal towns. It is not impossible that another pilgrim route from the west came to Aidhab by way of Darfur; traffic with medieval Nubia alone does not seem sufficient to explain its wealth. Finally late medieval Cairo was the apex of two trade paths, one leading- to the kingdoms of the western Sudan, the other to the ports on the east coast. v i 4 INDIVIDUAL RESEARCH REPORTS 69. It may have been Cairene merchants who first popularized in the West trading techniques that they had feecome familiar with in East Africa - the use of cowries for barter, possibly systems of gold weights. The riches of fifteenth-century Cairo were largely based on the African trade. But it seems probable that, while Cairo had traf- ficked with East Africa since the twelfth-century Fatimids, effective Cairene penetration into the West Sudan did not begin till the accession in 1382 of the first of the BurjlMamluks, Sultal Saif al-Din Barquq. Foreign merchants may have been active along all these routes. But this would not have been the normal pattern in pre-European African trade routes. It would now seem more likely that throughout Africa there were wide areas in the interior where trade goods circu- lated by barter, freely or sluggishly. At points near the circumference of such areas tracks came into ex- istence leading, however deviously, to markets along the coast frequented by foreign merchants. But until well into the nineteenth century, both east and west, it was normally traders from the interior who penetrated to the coast, not coastal traders penetrating into the interior. At least in East Africa the final section of such tracks will usually be found to follow rivers; for even when these were not navigable by raft, they provided a secure supply of drinking water. This will explain why so many of the medieval trading centres are by river mouths or opposite to them. In Kenya Malindi was by the old opening of the Sabaki which led down from the Kenya uplands; Kipini was at the mouth of the Tana, and goods could be shipped from there to Lamu and Pate. Mombasa was served by a system of creeks and streams. In Tangan- yika the wealth of Kisimani Mafia in the thirteenth and fourteenth centuries was perhaps partly due to its posi- tion opposite the mouth of the Rufiji. There may always have been a track along the changing course of the Pan- gani and the streams that fed it. During a brief archaeo- 70. INDIVIDUAL RESEARCH REPORTS logical survey in Portuguese East Africa I noted how much the trade was dependent on the river systems of the Sabaki and the Buzi. Sofala was a river port. But there were other routes that crossed rivers but did not follow them. Mr. Roger Summers has found traces of one which led from the gold-producing area round Penhalonga and the Ziwa people on the eastern border of southern Rhodesia to medieval Kilwa; this would seem to have crossed the Zambesi at Tete. But such routes may have been cut as wide as a swathe as the goods drifted from village to village in an un- organized traffic. It used to seem likely that most of the internal trade in pre-European Africa was after this fashion and.that supply centres in southern central Africa were first stimulated in the twelfth century by the distant demand Of Muslim traders on the eastern coast. Now I am not so certain. Dr. Brian Fagan has just discovered a new culture at Ingombe Ilede in North- ern Rhodesia. It is reported to be well evolved, and a quantity of beads prove its contact with the Indian Ocean area. A radio-carbon dating gives 850 A.D., plus or minus. It was presumably a centre of copper diffusion, like Katanga in the Congo. Tn the last months M. Mauny and Dr. Freeman-Gren- ville severally have worked on the references to the Wak-Wak peoples in the Book of the Marvels of India, the Kitab Ajaib al-Hind. They have established that there was an organized Indonesian penetration into East Africa during the tenth century. The demand for copper and for gold along the coast may have been pre-Muslim, created by Chola traders from south-east India and by the Sri Vijaya people from Sumatra and west Malaya. By the ninth century it may have been strong enough to have produced trade paths from Ingombe Ilede and Katanga and the entrepot that may have marked the earliest Zimbabwe. During the last two years the tendency in African archaelogy has : A . * INDIVIDUAL RESEARCH REPORTS 71. been to favour earlier rather than later datings. Perhaps in the next few years it may become increasingly apparent that the diffusion centres radia- ted west as well as east and south. A pressing need is an archaeological survey in Angola. This might begin with the Kola complex to the south-west of Benguela. It seems probable that Angola will be found closely linked with Southern Rhodesia and with Katanga copper ingots common to both. The skeleton of pre-European African history and archaeology will be reconstructed by the discovery of its trade routes. It now seems likely that the skeleton will prove to be pan-African. Gervase Mathew.