LIBRARY AND MUSEUM REPORTS 108. SOME NOTES ON THE SETTLEMENT-MOUNDS OF CENTRAL GONJA by R. Duncan Mathewson* Much of the archaeological research which has been done under the Volta Basin Research Project has been in Central Gonja within the vicinity of the Volta Confluence. This consisted of excavations on six mound sites and the surveying and mapping of a number of other prominent mound groups. These mounds commonly appear in clusters as at Butie, Old Buipe and the excavated site of New Buipe (York 1967 etc.); or in groups which are strung out along the White Volta as at Kisoto and Kpechilegbe (Mathewson 1967), Many of these mounds are about twenty feet high and are roughly oval in shape with a long diameter of about two hundred feet. In many instances these large mounds are associated with smaller mounds which are only three or four feet high, These mounds may be interspersed among the larger mounds as at Kpechilegbe, or encircling one individual mound as at Juni. There appears to be a very close association between mounds and baobabs, as they are commonly found on the edge of sites - sometimes measuring over ten feet in diameter. The excavations on the Juni and Kisoto mounds have clearly indicated that they represent the collapsed remains of multi-storeyed buildings and probably consist of three main successive periods of construction, Although all that remains of these structures is the lower floor and its corresponding wall stumps, the depth and nature of the deposit which covers these occupation floors clearly indicates that these structures must have had an upper storey, and that this deposit represents the collapsed remains of its swish walls. This deposit is usually about four to five feet thick, on top of which lies the bottom floor of the next building phase, To some extent the collapsed swish walls have been contained by the bottom courses of the thick exterior walls which helps to account for the progressive build-up of the mound, and explains to some degree the high angle of rest that these mound deposits sometimes have. * M r. R. Duncan Mathewson is a Research Fellow in the Department of Archaeology. 1 , A brief description of this work is found in Archaeology in the Volta Basin - 1963-66, V.B.R.P. and Department of Archaeology, Legon, 1967. LIBRARY AND MUSEUM REPORTS 109. These multi-storeyed structures had a ground floor which comprised of small rectangular adjoining rooms, the smallest of which probably measured about 91 by 12'. The exterior walls were over a foot thick in the bottom courses, which in some cases extended beneath the ground floor level. ^ These bottom floors had a hard, pounded surface which up-warped slightly at the wall juncture. Some of the walls were plastered with a fine silty-clay mixture which had been applied over the rough swish surface. The occurrence of clay cylinders which are identical to the drain pipes on the flat roofed buildings in Timbuktu and throughout the Sudan, indicate that these structures must have had a flat roof. The exterior walls were never meant to take the full weight of the upper storeys, but served primarily as 'curtain walls' which contained an internal timber structure that probably took the main strain of the upper levels. There is some evidence that timber up-rights existed along the inside of the walls. The alignment of these supporting timbers is not at all certain, but they must have supported cross beams upon which smaller timbers were laid; thereby, forming a wooden platform, which was the main structural feature of the upper floor. There is good evidence to indicate that swish was applied on top of the platform and in between its wooden joints which must have helped to bind it together, as well as producing a smooth floor surface. The dating of these mounds has not yet been completely resolved. This will ultimately depend upon the detailed pottery analysis of the individual mounds and their relationship to the tobacco-pipe chronology, The absence of pipes from the Juni and Kisoto mounds indicate that their successive building periods were occupied prior to the mid-17th century and the local introduction of the pipe (Ozanne 1962). On the other hand, the two later phases of the Krunkrun- mboi mound (Flight 1967), as well as the later periods of the New Buipe mounds, contain pipes which indicate that these sites were occupied respectively until the late 17th and 18th century. The pottery from these mounds primarily consists of a number of painted wares which are decorated by simple geometric designs. One of the most common forms of decoration are line patterns which are sometimes accompanied by varying motifs of hatchings and crosses. More elaborate designsdo exist, especially at 2. It may well have arrived in the north at a slightly earlier date - possibly within the first quarter of the 17th century. LIBRARY AND MUSEUM REPORTS 110. New Buipe, as well as painted wares which have a zoned impressed and incised decoration. How this great variation in painted designs might be chronologically related is still uncertain? however, some similarity does exist which should enable some correlation to be made with the pottery of the upper periods (IV-VIII) in the New Buipe sequence to that of the painted wares from the other mound sites. The Kisoto Bowl is a very distinctive pottery form that occurred in great quantities in the Kisoto mound, and as it has been subsequently found at four other excavated mound sites, it has become an invaluable cross-dating 'fossil1. It had been noted earlier as occurring as surface pottery with painted wares which were associated with mounds in the confluence area (Davies 1964), but further survey work has demonstrated that its distribution is centred primarily in the area on both sides of the White Volta just north of the confluence. Excavations have shown that it is primarily a pre-pipe pottery form and although it has a fair degree of uniformity, small variations exist in rim form, decoration, and shape which has provided a useful chronological index for the immediate pre-pipe levels (Flight and Mathewson- in preparation). The distribution of these mound sites conforms very closely to the known area of Gbanya settlement and to some extent reflects their occupation of iCentral Gonja during the mid-16th and 17th century (See Wilks 1966). Although a closer dating of these mounds is still rather uncertain,^ there is a strong suggestion that the occupation periods at Juni and Kisoto might be contemporary with Periods V and VI at New Buipe.4 • Although the Gbanya had become the dominant element in Central Ghana by 1600 AD, it is difficult to deterine what are the specific intrusive Gbanya traits and what should be attributed simply to pre-existing autochtonous cultural elements. How far this painted pottery should be considered as part of the Gbanya complex, thereby reflecting an earlier underlying Mande tradition, is very uncertain (Mathewson 1968a). The pottery sequence at New Buipe has revealed painted pottery which long ante-dates the arrival of the Gbanya (York 1968). It may well 3. See Flight 1968 for a tentative chronology. 4. Two C-14 dates have been received from the Juni mound which gave mean dates of 1460 AD. for the earliest period and 1630 AD. for the latest. These dates will be assessed at a later time. /U-J LIBRARY AND MUSEUM REPORTS 111. be that there are two distinct painted pottery traditions superimposed at New Buipe - an earlier type of zoned impressed and incised painted wares and a geometric and design painting style which occurs in the later Gbanya levels. It is very tempting to associate these multi-storeyed buildings completely with the arrival of the Gbanya. The large concentration of mounds within the Mpaha-Butie, Kusawgu-Tuluwe, Kafaba-Larmisa and Buipe areas strongly suggests that they could very well be a primary Gbanya feature. However, multi-storeyed buildings represent a savannah type of architecture which is widespread throughout the Western and Central Sudan, and it is quite possible that they were already existing in the area prior to the Gbanya invasion.* Largely through the efforts of Hausa and Bornu traders," a number of these towns probably had by the end of the 15th century developed into centres which largely controlled \he trade across the Volta Basin (Mathewson 1968b, 1968c). element within the confluence area' which to some extent developed and maintained the flow of trade which attracted Muslim traders, and thereby establishing commercial links with the Mossi and Hausa States. Some of the mounds in this area may in fact represent initial Dagomba occupation,8 which was prompted by the economic importance of the confluence area as the major northern outlet for the highly valued forest commodities of gold and kola. It is highly likely that there was at this time a significant Dagomba important trading •4 17 5. Excavations at Yendi Dabari have indicated the existence of possible multi- storeyed buildings. (See Shinnie and Ozanne 1962). 6. Muslim groups at Mpaha, Kafaba, and Buipe claim to have settled before the arrival of the Gbanya and to have come from Hausa or Bornu (Goody, 1964; Levtzion 1966). 7. Kasawgu (Kawsaw of Tamakloe) was apparently initially a large Dagomba settlement which was the scene of very heavy fighting with the Gbanya. (See also Goody 1964). 1-4 X- 8. Bilegas (a Dogbane word) are water cisterns which have been dug down through the surface laterite into the soft decomposed bedrock, They are associated with mounds in areas which do not have an alternative water supply. Although their distribution is unclear, many occur in the Butie area and it is probable that they were initially a Dagomba feature which developed as a result of contact with Hausa we 11-diggers. LIBRARY AND MUSEUM REPORTS 112. Although the Gbanya may not have initially introduced these buildings It is into this area, they certainly became closely associated with them. quite clear that many of these mounds represent occupation at'different periods. It has been noted that there were four mounds still being occupied at Butie as late as 1914 (Cardinal 1931).' A mound cluster at Butie indicates very recent abandonment, as on top of one of the larger mounds, a rectangular floor plan is still recognizable along with exposed timber supports, and a partially collapsed swish wall (Mathewson 1967). These mounds must be the ones that were still being occupied just over fifty years ago. However, this must be a unique instance, and it is probable that the descriptions of this very late occupation indicate that it represents in part an adaptation and utilization of pre-existing mounds which had long since been abandoned (See Tominlison 1954). There is no doubt that these buildings would have served as a highly effective defensive position. During the Gbanya invasion of Dagomba, in about the first decade of the 18th century, "Kumpati" built flat topped houses ('houses of Navrongo type with no grass roofs') which the Gbanya defended against the oncoming Dagomba (Tait B). built at the various places where he encamped, some of which might still be seen in "Duni", a village of Singalu" (Tamakloe 1931). These fortifications were supposedly These multi-storeyed buildings represent an architectural tradition of the savannah zone which ecologically belongs to the low rainfall areas north of the forests. The apparent existence of this type of building at Bono-Manso (Meyerowitz 1951) signifies however, the penetration of major northern cultural influences into the forest fringe, the extent of which has yet to be determined. 9. The main Butie mound field consists probably of over thirty mounds. These are very closely set together and are of varying shapes which may indicate different types of buildings. Their date is very uncertain, but they may have been occupied as late as the 19th century when Butie was a stronghold against Ashanti penetration. 10. Tamakloe mentions that the swish had been mixed with honey instead of water. A similar tradition is connected with Jakpasere, a building \n the Kpembe- Salaga area, which is reputed to have been built by "Jakpa" out of a mixture of swish and shea-butter. The full excavation of this site is presently being undertaken by the writer. •* * 9\ Jfcr-S M *-- Xi m * 3 LIBRARY AND MUSEUM REPORTS 113. References Cardinal, A . W ., 1931, The Gold Coast Davies, O ., 1964, 'Gonja Painted Pottery*. Trans. Hist. Soc. Ghang 7, pp. 4 - 1 1. Flight, C ., 1968, 'The Settlement-mounds of Central Gonja: Provisional Chronological and Historical Context1. Forthcoming UNESCO' Conference, Ouagadougou, Upper Volta. Goody, J ., 1964, 'The Mande and the Akan Hinterland1. The Historian in Tropical Africa. Ed. - Vansina, Mauny, and Thomas, pp. 193-212. Levtzion, N ., 1966, 'Salaga -A Nineteenth Century Trading Town in Ghana1 Ann. of Israel Orien, Soc. 2, pp. 207-244. Mathewson, R.D., 1968a, 'The Painted Pottery Sequence in The Volta Basin1. The West African Archaeological Newsletter 8, pp. 24-31 , 'Kitare: An Early Entrepot in the Volta Basin1. 1968b, VI - Pan-African Congress of Prehistory, Dakar, Senegal. 1968c, Iron Age1 - Forthcoming. 'A Proposed Chronological Framework for the Ghanaian Meyerowitz, E.L-R., 1951, The Sacred State of the Akan Shinnie, P.L. and Ozanne, P., 1962, 'Excavations at Yendi Dabari1. Tfarts, Hist. Soc. Ghana 6, pp. 87-118. Tait 'Dagomba Drum Histories'. (Tait B). I.A.S. Tamakloe, E.F., 1931, rA Brief History of the Dagbamba People1. Tomlinson, H . H ., 1954, Gonja Peoples'. 'The Customs, Constitution and History of the LIBRARY AND MUSEUM REPORTS 114. Wilks, I ., 1966, 'A Note on the Chronology and Origins of the Gonja Kings'. Ghana Notes and Queries 8, pp. 26-28. York, R . N ., Mathewson, R.D., Calvocoressi, D ., Flight, C, 1967, 'Archaeology in the Volta Basin 1963-66, V.B.R.P. and Department of Archaeology, Legon. York, R . N ., 1968, New Buipe'. 'Pottery Analysis: Further Methods and Results from Vl-Pon-African Congress of Prehistory, Dakar, Senegal, t. •'.