Mumba Rockshelter in northern Tanzania presents one of the richest and most complete archaeological sequences in East Africa for the Middle Stone Age through the Iron Age. Past excavations of the shelter revealed an extremely rich lithic and faunal assemblage, but were problematic, either because of poor excavation and recording methods (in the 1930s), or because the materials were never fully studied (in the 1979/1981 excavations). In both cases, excavators had concluded that the shelter contained a deposit without visible separation between archaeological levels. Re-excavation of Mumba, using modern techniques for recording spatial data, show that the previous geological and archaeological sub-divisions of the shelter deposits need much revision. The results of the excavation have implications for the interpretation of the “transitional” Mumba Industry in the Pleistocene levels and for the co-occurrence of ceramic traditions in the Holocene levels.
Ongoing archaeological research at North Lake Eyasi has produced a wealth of information, including a new hominid fossil and several archaeological sites dating to the end of the Middle Pleistocene. One of the sites (WB9) has been excavated and has produced evidence of multiple processes in its formation, including evidence of functional associations of stone tools and faunal remains which are scarce for this time period. The stone tool industry is based on a core and flake industry, which is not very diagnostic and attributed to MSA. Earlier heavy-duty tools classified as Sangoan may derive from the underlying Eyasi Beds. The stratigraphic provenience of previous fossil hominids is unknown. Surface collections from the Eyasi lake, thus, comprise two different sets of stone tools and fossils, which can only be clearly differentiated in the field. This advises against the use of previously curated collections as a homogeneous sample. Earlier definitions of the Njarasa industry should be revised. This work presents results on the paleoecology of the area and of its paleontological and archaeological information, with special reference to the excavation of WB9, the most complete site discovered in the area so far. This contributes to the limited information available about site functionality and hominid subsistential behaviour in East Africa during the end of the Middle Pleistocene. A technological study from WB9 also shows the variability of stone tool traditions at this time.
Recent re-excavation of Mumba Rockshelter unearthed an unbiased lithic sample from Bed V. Technological analysis has permitted a reinterpretation of the so-called Mumba Industry, a transitional industry between Middle and Later Stone Ages originally defined by Mehlman (1989). Our data confirm Mehlman’s observation that the “evolutionary” markers in Mumba Bed V are basically typological. However, our study differs from his in that we classify all of Bed V as LSA based on the combined analyses of typology and technology in our excavated assemblage. From a technological perspective, no changes have been observed throughout the sequence, and continuity is the main technological characteristic of the series. The only transitional marker from Lower through Upper Bed V is the appearance of the geometric crescent in the latter, taking into account that microliths exist throughout the sequence. This evidence casts some doubts on previous interpretations and underscores the need to recover a larger sample using modern excavation techniques. It also stresses the need to define the MSA/LSA transition in better terms, combining techno-typological criteria.