Leguminosae constitute an important proportion of the charcoal sampIes recovered at archaeological sites in the West African savannas. Identification of these fragments to a level below family or subfamily was problematic, because a comparative survey was missing. Therefore, the wood anatomy of 31 species (23 genera) of Mimosoideae and Caesalpinioideae growing in the Sudanian savannas of West Africa was examined. The species were grouped into 18 types according to wood anatomical structure. The types represent single species or genera (fourteen types), two genera (three types) or three genera (one type) . The following features are regarded as suitable for a reliable delimitation and identification . Heterocellular rays and storied structure allow for a first differentiation. Enlarged, non-bordered vessel-ray pitting , nonvestured vessel-pits, silica, axial canals, septate fibres and crystals in non-eharnbered ray cells are additional features characterizing few or single types. Types without these features are delimited less easily. Parenchyma distribution and ray width are, due to variability, not as reliable , but remain necessary features for identification. Types characterized by these features only may not always be recognized correctly. Quantitative features of the vessels are not regarded as helpful for the differentiation within the set of examined species. A table (Table 1) summarizes the results for easy reference.
Identification of archaeological or soil charcoal in a species-rich biome, such as the Central African rainforest, is challenging because of the large number of woody taxa with similar and overlapping wood anatomical features. Valid environmental or archaeological interpretations can only derive from reliable and transparent identifications that allow comparison of and referencing between different charcoal assemblages. The identification of 30 archaeological charcoal types from the site Dibamba in southern Cameroon serves as a starting point for a discussion on classification and naming. These 30 types are fully documented and illustrated in the Supplementary Online Material (SOM). The discussion underlines the basics of “good practice” of charcoal identification in a speciesrich tropical environment. The value of differential diagnosis is stressed, as is the importance of leaving identification levels on higher taxonomic level if necessary. We argue that the level of identification must be reflected in the name of the charcoal type. Names of charcoal types are written in small capitals to clearly distinguish them from botanical taxa with which they are not necessarily identical. The Dibamba charcoal assemblage offers the first and so far unique possibility to directly comprehend human impact on the structure and composition of West Central African rainforest over the last 3000 years. The paleoenvironmental significance of the results presented here will be subject of a forthcoming publication.
Settlement activities of the Nok Culture considerably decreased around 400 BCE and ended around the beginning of the Common Era. For a better understanding of the decline of the Nok Culture, we studied the charcoal assemblage of the post-Nok site Janruwa C, dating to the first centuries CE. Janruwa C differs from Middle Nok sites in ceramic inventory and a wider set of crops. 20 charcoal types were identified. Most taxa are characteristic of humid habitats such as riverine forests, while those savanna woodland charcoal types that had been dominant in Middle Nok samples are only weakly represented. The differences between the Middle Nok and post-Nok assemblages do not indicate vegetation change, but rather different human exploitation behaviors. It seems that the Nok people avoided forest environments while in the first centuries CE, other, possibly new populations settled closer to the forest and were more familiar with its resources. The new exploiting strategies might be explained as adaptation to changing environmental conditions. Our results, together with data from other palaeo-archives in the wider region, point to climatic change as a potential factor for the decline of the Nok Culture. We argue that erosion on the hill slopes, maybe due to stronger seasonality, was responsible for land degradation after 400 BCE and that the Nok people were not flexible enough to cope with this challenge through innovations.
Since 2003, a joint research project of the universities of Frankfurt and Tübingen (Germany) has explored the changing interrelationship of environment and culture in the forest-savanna regions of West and Central Africa. This paper provides the first archaeological and archaeobotanical results of three field seasons in the rainforest of southern Cameroun. Excavations were carried out at Bwambé Hill in the vicinity of Kribi at the Atlantic coast as well as at Akonétye, Minyin and Abang Minko’o, all located in the hinterland near Ambam. At all sites a number of pit structures, which contained mostly ceramics, were excavated. In addition, at Akonétye two graves with rich ceramic and iron offerings were unearthed. They seem to be the oldest graves with iron objects yet known in Central Africa.
A large body of archaeobotanical material was retrieved from the structures excavated (charcoal fragments, charred fruits and seeds, phytolith and starch samples). Of high importance is the presence of pearl millet (Pennisetum glaucum) at Bwambé Hill and Abang Minko’o in archaeological contexts dated to about 2200 bp. Charcoal and pollen data indicate that the ancient settlements were situated in a closed rainforest which was, however, massively disturbed and partly substituted by pioneer plant formations.