Introduction Archaeological Survey
Total Page:16
File Type:pdf, Size:1020Kb
NYAME A KUMA No. 49 June 1998 of imported ceramics and glass bottles (Chami and Msemwa 1997a). This suggests that the Rufiji delta The 1996 Archaeological Reconnais- and the neighbouring area were connected to the sance.North of the Rufiji Delta ancient transoceanic trade early in the fmt millenium AD (Casson 1989). Felix A. Chami and Bertram B. Mapunda The presence of sites with similar cultural ma- Archaeology Unit terials in the islands of Kwale, Koma and Mafia also University of Dar es Salaam suggests that the EIW people were widely spread f!0. Box 35050 over the region. Apparently they would exploit re- Dar es Salaam, Tanzania sources from the hinterland environment for local or long distant trade. The area near the delta and along the river was probably conducive for settle- Introduction ments due to abundant marine and riverine resources, particularly fish, shellfish, hippos and mangroves. This is a preliminary report of an archaeo- Alluvial soils and scattered small lakes around the logical reconnaisance conducted in the Rufiji &is- delta would also provide for agriculture and fish- trict, Coast region in July and August 1996. The ing. The 1996 archaeological study in the Rufiji dis- district of Rufiji comprises the whole of the Rufiji trict was therefore expected to further enhance the delta, some few off-islands near the delta, and the already established knowledge on the ancient cul- immediate hinterland of the delta extending to about ture of the central coast of Tanzania. 100 km into the interior. The district is on the cen- tral coast of East Africa, about 100 km south of Dar-es-Salaam, Tanzania from latitude 7O35' to Archaeological Survey 8O00' S and longitude 38O30' to 39" 20' E (Map 1). The survey work took place between the 23rd The archaeological research was organized by the and the 31st July 1996. Survey was extensive rather Archaeology Unit of the University of Dar es Sa- than intensive. The purpose was to cover as large an laam as fieid training for archaeology students. The area as possible for the purpose of obtaining a good Archaeology Unit has annual field school sessions, amount of data. In many places, however, we were which involve, apart from training, archaeological hindered by dense vegetation. Only open places, e.g. survey and excavations. farms (usually less than three acres), road, and foot This area was selected for research in order paths were properly surveyed. Work concentrated to fullfil a long term campaign of surveying the on the northern half of the Rufiji district. This in- entire central coastal area of Tanzania. The cam- cluded Bungu Ward to the north, Nyamisati and paign was initiated in 1986 by the Archaeology Unit Mchukwi to the East, Kibiti at the centre and Ikwiriri with a reconnaisance around Mkiu, in Kisarawe, and Mkongo to the south. To the west the survey today known as Mkuranga district (Schmidt et al. work reached Mloka and the eastern part of the 1992; Chami 1988, 1992). To date, a large part of Selous game reserve (Map 1). We also conducted a this area has been investigated, thanks to field casual reconnaisance south of the Rufiji river par- schools (Chami and Kessy 1995) and individual ticularly around Utete town and the free animal hunt- researchers, particularly, Felix Chami and Paul ing zone south of Utete. Msemwa (Chami 1994; Chami and Msemwa 1997a, Three different survey techniques were em- 1997b) and Emanuel Kessy. The previous results ployed based on the vegetion coverage of the area point to the Rufiji district as an area worthy of ar- First, dnving slowly along motorable roads, observ- chaeological research. The littoral area of the dis- ing for features and concentration of artifacts such trict near the delta has been found to contain re- as potsherds, slag, and lithics along the road, on the cent, Later Stone Age (LSA) and Early Iron Work- walls of road cuttings or further off the road where ing (EIW) sites. These represent the earliest known visibility allowed. We also paid attention to places settlements on the littoral of the central coast of disturbed by charcoal burners, borrow pits and dump- East Africa The excavation of an EIW site on the ing pits. Once materials were found, the car was old shore, near the delta at Kivinja, yielded remains NYAME AKUMA No. 49 June 1998 Map 1: Archaeological sites near the Rufiji Delta. Potsherds 0c Charcoal Reddish brown (5YR 4/31 B DojF~fd(2.5 YR . Reddish brown (2.5YR 5/41 Reddish brown (5YR 5/31 Red (2.5 YR 4/6 Red (10 R 5/61 NYAME AKUMA No. 49 June 1998 Figure 2: Late Stone Age artifacts from the lowest cultural layer of Nkukutu-Kibiti. stopped for thorough examination and recording. settlements would occur where it is safe, adminis- When necessary a hoelshovel test was applied to tratively feasible or culturally appealing (e.g ., strong examine the subsurface. This technique of driving spiritual connection with ancestors). For this rea- slowly while observing was employed in three types son, it is not unusual to find that some areas have of areas: been settled continuously for decades, centuries or (i) Flat lands such as hill tops, hill bottoms, millennia. So observing the soil disturbance by the and plain land. Since gully erosion that could ex- modem dwellers can reveal the hidden past. pose subsurface materials is minimal in such a land- (ii) Areas with road-cuts and errosions. Road- scape, walking along the road proved to be fruit- cuts and erosion tend to expose materials lying un- less. der surface. Walking, rather than driving, enabled (ii) Areas with low visibility and accessibil- us to observe closely the gullies and road cuttings ity. These included areas with dense forests, long for both archaeological and geological phenomenon. grasses, itching runner plant (Mucuna pruriens or (iii) Areas found to have scatters of artifacts, upupu in Swahili) and marshes /swamps. features or sites. Such places needed close observa- (iii) Protected and dangerous areas such as tion, including subsurface testing and recording. military camps or wild animal reserves. Third, we walked along transects. This tech- Second, we walked along the road and foot nique was used in open farms and grassland areas paths. This technique was employed on three ar- cleared by bush fires. In addition to observing the eas: (i) in and around density settled areas. In such occurance of archaeological materials on the sur- areas we focused on the wall of wattle and daub face, hoelshovel tests were also used in such areas houses because experience from other places has to check subsurface materials. indicated that daub reveal artifacts such as pottery and slag that had been collected in borrow pits Results (Mapunda 1995). The assumption is that sin& the prehistoric times people have tended to settle near Recent remains of local ware, dating probably important resources such as water, food, raw mate- between the 17th and the 19th centuries, were found rials for artifacts or easy trade connections. Some and some collected from the towns of Kibiti, NYAME AKUMA No. 49 June 1998 Nyamakonga and Kinyaya. Associated with the lo- However, one site was found at Temboni vil- cal ware were Chinese sherds of the same period. lage, about 2 km from Kingwila. It is on a ridge The areas between Kibiti and the Mchukwi junc- oriented east-west. The site is deeply cut by the tion, about 6 km southeast of Kibiti as well as the road, exposing a deep subsurface layer of quartz vicinity of the Mchukwi town (Map 1) were the gravels. The top 30-60 cm below the surface con- most productive survey areas. The fmt site, EIW, sists of a concentration of pottery of the PW pottery was found at Mkukutu, just 1.5 km from the centre tradition. The potsherds look like those of Mchukwi of Kibiti. Cultural materials including potsherds, (Figure 5j), suggesting the 11th to the 13th century slag and lithic chips were observed eroding from period occupation. Associated with the pottery were the floor and the walls of a track made by bulldoz- spindle whorls (Figure 6g and 6h) suggesting the ers for lomes to fetch gravel quarried to the south- practice of cloth making. west of the site (Map 2). Part of this site is badly destroyed by this quarrying activity. A close ex- Archaeological survey on the Ikwiriri brown- amination of the potsherds both from survey and ish soils yielded no substantial data, as did the search excavation revealed the site to be of Limbo tradi- on lowlands further east towards the heart of the tion (Figure 3; see Chami 1988,1992). delta as far as Muyuyu village. A large portion of this lowland had too many streams and dense thicket Further south, at Kingwila village, we found abstracting free accessibility. A survey south and another EIW site composed mainly of pottery of south west of Kibiti along the road from Kibiti to the Kwale tradition (Figure 4a and 4b; see Soper Mng'aru village, yielded two sites. The first was 1967). It seems the major part of the site was de- found at Mwangia village, about 2 km from the stroyed by.road construction activity. Had archae- Kibiti-Mkongojunction. A scatter of potsherds, di- ologists surveyed the land before road construction agnostic of the terminal EIW period, was found erod- the two EIW sites would have been discovered and ing from the road Figure 4c-g). The second site was studied before the aforementioned destruction. Un- found at Mng'aru village, overlooking a stream fortunately there is currently no such policy in Tan- which separates the village from Mng'aru prison.