Cultivating the rainforest
The nature, inception, and rate of spread of agriculture in Central Africa, especially in areas now covered by tropical rainforest, is based on quite patchy data sets. Nonetheless, there is sufficient evidence to allow a reasoned reconstruction of the key developments and drivers of change.
Two separate phases can be identified: an initial Stone-to-Metal phase marked by the presence of ceramics associated with polished stone axes and hoes between c. 1500 bce and 100 ce; and an overlapping early Iron Age, signalled archaeologically by the first evidence for iron-working and new styles and forms of ceramics, between c. 400 bce and 1000 ce. Research has shown that the presence of ground stone artefacts such as axes and hoes is not always a good indicator of the presence of food producers (in many cases they may be a far better indicator of terminal LSA hunter-gatherers), and that some pre-farming communities also made pottery (as for example at Shum Laka rock shelter, northwestern Cameroon). Nevertheless, the marked increase in sites across the Congo basin after c. 600 bce and the introduction of new technologies likely signal demographic increase and southward migration of incipient slash-and- burn agriculture. From the distribution of ‘Stone-to-Metal Age' (SMA) sites (i.e. those associated with initial adoption of iron metallurgy in conjunction with continued widespread use of stone tools), the main migration routes seem to have been from south-central Cameroon eastwards along the northern boundaries of the rainforest, and southwards along the Atlantic seaboards of what are now Equatorial Guinea, Gabon, and Congo, with some eastward movement along the Ogooue River as far as modern-day Lope (central Gabon), between c. 800 and 300 bce. The homogenous nature of the pottery over this entire strip implies that this was a single cultural group, as also suggested by the restricted settlement preference during this phase for hilltops and other dominant positions along river courses.[1072]This southward movement may well have been facilitated by forest fragmentation as a consequence of the weakening of the Atlantic monsoon from around 1500 bce.
A few hundred years later, more extensive forest regression appears to have triggered the southward movement of a new group of people with knowledge of iron metallurgy as well as farming, who were speakers of proto-Western Bantu languages. The earliest of these sites, at Oliga in southern Cameroon, is dated to the mid first millennium bce. Subsequent expansion by these metal-using communities across areas already occupied by the makers of Malongo ceramics and further inland across areas that now lie within the forest zone was fairly rapid between c. 600 bce and 100 ce. There followed a decline in settlement and possibly even a ‘population crash' leading to the abandonment of large parts of the interior, perhaps as a result of environmental impacts triggered by widespread forest-burning associated with charcoal production and slash-and-burn agriculture.[1073]It is likely that yams were an important component, and perhaps also millet and even banana, as possibly documented at the site of Nkang, situated c. 70 km northwest of Yaounde, Cameroon, dated to between 800 and 350 BCE.[1074] Recent work in southern Cameroon indicates that pearl millet was being cultivated at Abang Minko and Bwambe-Sommet by around 400-200 BCE in an area that currently lies within the forest zone, along with the exploitation of a range of forest-based plant resources, including oil palm and the edible nut Canarium schweinfurthii. The cultivation of Bambara groundnut is also attested at the nearby site of Akonete.[1075] Both pearl millet and Bambara groundnut are savanna crops that originated further north and are adapted to drier conditions than those that currently prevail in southern Cameroon. This would imply that annual average rainfall was lower than today when these sites were occupied. In turn, this might explain why, even if banana/plantain were being cultivated a few centuries earlier further north at Nkang, they were not recovered on the more southerly sites.[1076]
As at Nkang, much of the available evidence concerning subsistence strategies in Central Africa comes from pits.
These are especially common on sites dating to between 900 bce and 600 ce, and some are associated with LSA/Stone-to-Metal Age traditions, while others clearly have an EIA affiliation. Most pits are typically between 1.5 and 3 m deep and contain mixed ceramic, stone, and on occasion metal assemblages (hence an EIA affiliation), along with the remains of oil palm and Canarium and other edible nuts. Their distribution extends across the forest zone from the Atlantic coasts of Cameroon, Congo, and Gabon to as far north as Nanga Eboko, central Cameroon. Faunal remains at these sites are rare, although ovicaprids are known from Nkang and Toube 1 (Gabon). The evidence from several rock shelters on the other hand, including Shum Laka and Abeke (Cameroon), points to a reliance on wild species,[1077] while the exploitation of marine shellfish is documented at some of the coastal sites in Gabon and the Republic of Congo. In the latter, several Ceramic LSA sites, such as Tchissinga West on the Loango coast, are dated to between c. 1300 and 100 BCE, with deep pits filled with carbonized oil palm nuts and occasional flat grinding stones and distinctive globular-necked jars. EIA sites, distinguished by the presence of a very different style of ceramics and the presence of iron artefacts and slag, first appear in this area around 100 bce;[1078] however, other than evidence for the exploitation of oil palm, no direct material evidence for subsistence practices has been recovered from either category of site on this part of the coast. Extensive surveying along the main river channels in the Zaire basin has also documented a range of similar sites, although with their own distinctive ceramic styles, that suggest an initial establishment of farming communities in these areas between c. 800 and 100 bce.[1079]