The Palaeovegetation of Janruwa (Nigeria) and its Implications for the Decline of the Nok Culture

Alexa Höhn & Katharina Neumann

Abstract Résumé

Settlement activities of the Nok Culture considerably de- Les indices de peuplement de la culture Nok diminuent con- creased around 400 BCE and ended around the beginning of sidérablement aux alentours de 400 BC puis disparaissent the Common Era. For a better understanding of the decline au début de notre ère. Pour mieux comprendre son déclin, of the Nok Culture, we studied the charcoal assemblage of l’assemblage anthracologique du site Janruwa C, datant des the post-Nok site Janruwa C, dating to the first centuries CE. 1ers siècles CE, a été étudié. Le site diffère des sites Nok au Janruwa C differs from Middle Nok sites in ceramic inventory regard de l’inventaire des céramiques reconnu ainsi que par and a wider set of crops. 20 charcoal types were identified. une plus grande diversité des plantes cultivées. 20 types de Most taxa are characteristic of humid habitats such as river- charbons ont été identifiés. La plupart des taxons sont caracté- ine forests, while those savanna woodland charcoal types that ristiques d’habitats plus humides (ripisylve) alors que ceux de had been dominant in Middle Nok samples are only weakly la savane boisée, qui dominent les échantillons du Nok Moyen, represented. The differences between the Middle Nok and sont faiblement représentés à Janruwa C. La différence entre post-Nok assemblages do not indicate vegetation change, but les assemblages du Nok Moyen et du post-Nok ne présente rather different human exploitation behaviors. It seems that pas un changement de végétation mais plutôt une modification the Nok people avoided forest environments while in the first des modes d’exploitation. En effet, alors que les populations centuries CE, other, possibly new populations settled closer Nok ont évité les environnements forestiers, il semble que, to the forest and were more familiar with its resources. The durant les 1ers siècles CE, de nouvelles populations se soient new exploiting strategies might be explained as adaptation to installées à proximité et étaient plus familières avec les res- changing environmental conditions. Our results, together with sources forestières. L’émergence de ces nouvelles stratégies data from other palaeo-archives in the wider region, point d’exploitations peut être expliquée comme une adaptation to climatic change as a potential factor for the decline of the aux changements des conditions environnementales. Nos ré- Nok Culture. We argue that erosion on the hill slopes, maybe sultats, associés aux autres données paléoenvironnementales due to stronger seasonality, was responsible for land degrada- ouest-africaines, suggèrent un changement climatique comme tion after 400 BCE and that the Nok people were not flexible facteur potentiel à l’origine du déclin de la culture Nok. Nous enough to cope with this challenge through innovations. proposons ainsi que l’érosion des versants, peut-être en raison d’une saisonnalité croissante, soit responsable de la dégradation des terres après 400 BC et que les populations Nok n’étaient pas suffisamment adaptées pour faire face à ce défi par des innovations.

Keywords: West Africa, archaeobotany, archaeological charcoal, vegetation history, human impact, climatic change

Alexa Höhn 8 [email protected] / Katharina Neumann 8 [email protected] * Institute for Archaeological Sciences, African Archaeology & Archaeobotany, Goethe University, Norbert-Wollheim-Platz 1, 60629 Frankfurt a. M., Germany

DOI 10.3213/2191-5784-10296 © Africa Magna Verlag, Frankfurt a. M. Published online 18 Nov 2016

Journal of African Archaeology Vol. 14 (3) Special Issue, 2016, pp. 331–353 331

This article is available under a Creative Commons Attribution 4.0 International License (CC BY 4.0) Downloaded from Brill.com10/10/2021 06:57:41PM via free access A. Höhn & K. Neumann

Introduction sites nearby, Janruwa A and B, and with the results of previous investigations (Kahlheber et al. 2009). The The Nok Culture existed from the mid-second millen- most remarkable feature of Janruwa C is the presence nium BCE to around the turn of the Common Era in of fonio (Digitaria exilis Stapf) and of oil palm (Elaeis Central Nigeria, with settlement activities peaking dur- guineensis Jacq.), which had not been used during Mid- ing the Middle Nok phase between 900 and 300 BCE. dle Nok times (Kahlheber 2010). We hypothesize that Towards the end of this phase, the number of settlements the clear differences between the charcoal inventories of considerably decreased; according to radiocarbon dates, Janruwa C as opposed to Janruwa A and B and others only very few Late Nok sites have been found in the do not point to vegetation change, but instead to a shift research region (Franke & Breunig 2014; Franke of settlement sites from the hill slope to the valley floor, 2016). With the beginning of the Common Era, archaeo- next to the river. In the context of palaeoenvironmental logical sites with Nok attributes disappear completely. data from the wider region, we conclude that climatic New ceramic inventories, totally different in decoration, change might have been involved in the decline of the style and clay composition, and the absence of terracotta Nok Culture. figurines Beck( 2015; Franke 2015), in combination with a higher diversity of crops (Kahlheber 2010), point to a cultural change, possibly related to shifting The site and its environmental settings environmental conditions. Environmental settings Culture and environment seem to have been stable during Middle Nok times. There is no marked variation The core region of the Nok research project is located in the archaeological and archaeobotanical assemblages in the northern High Plains, the hilly foreland of the Jos of Middle Nok sites. For the production of the terracotta Plateau, northeast of the Nigerian capital of Abuja (Fig. figurines only very few clay sources were used Beck( 1). The relief is undulating with striking granite outcrops 2014, 2015). Stringent rules seem to have applied to and elevations between 300 and 900 meters. At Kafan- decoration and form of vessels in comparison to Early chan, a town about 60 km to the east of Janruwa C, mean and Late Nok pottery (Franke 2014). Throughout annual rainfall is about 1800 mm, mean monthly temper- Middle Nok times, subsistence was based on the same ature is 25°C and relative humidity averages about 63% crops and wild fruits: Pearl millet (Pennisetum glau- (Abaje et al. 2010). The rainy season lasts from April to cum (L.) R.Br.) was the most important crop, present October. If not cited otherwise the information on the in almost every site (Kahlheber et al. 2009) and cul- vegetation of the research area is derived from Keay tivated together with cowpea (Vigna unguiculata (L.) (1953), according to whom the region lies in the northern Walp.). The rarity of cowpea in the sites is probably part of the Southern Guinea vegetation zone of Nigeria. due to preservation problems. There is no evidence of Floristically it belongs to the Guinea-Congolia/Sudanian root or tuber crops. Endocarp fragments of Canarium regional transition zone (White 1983), with a mosaic schweinfurthii Engl. and Vitex sp., both with oleagi- of secondary (wooded) grassland, semi-evergreen rain nous seeds and edible mesocarp, have been found in forest and Isoberlinia woodland, the latter being typical many sites, but probably represent only a small part of of the Northern Guinea zone of Nigeria . Woodlands are the wild species collected for food (Kahlheber et al. open stands of trees at least 8 m tall with a canopy cover 2009). From the charcoal assemblages of Middle Nok of 40 % or more and a field layer usually dominated by sites, we assume that cultivation took place in a shifting high grasses; they are not true forests (White 1983). cultivation system with long fallow periods, and there The woodland is annually swept by fire and the degree is no evidence for erosion or land degradation. Prob- of canopy closure may vary, resulting in more and less ably the settlements shifted as well; this could explain open patches of savanna woodland. Gallery forests grow the surprisingly high number of archaeological sites along rivers, often specifically riverine species are found dispersed in the landscape (Rupp 2014). here tolerating periodical inundations. But also species of the lowland rainforest enter savanna regions along the What was the reason for the disappearance of the rivers. Gallery forests reach up to 600 m in width along Nok Culture, after having existed under stable condi- a major river (Porembski 2001), but are often narrower. tions and in a quite conservative manner for about In the vicinity of the Janruwa sites the watercourses are 1500 years? In this paper we explore the hypothetical small, and human pressure has reduced the gallery forests relationship between human impact, climate and veg- to small strips of at most several tens of meters width etation change and their implication for the decline of today. Species present at the Janruwa riverbank include the Nok Culture during the 4th century BCE, through Pandanus sp., Elaeis guineensis, Mitragyna sp., Ficus the study of charcoal samples from the post-Nok site sur Forssk. and Pterocarpus santalinoides DC. Light Janruwa C, dated to the 2nd to 4th century CE. We forests and dry forests are the two zonal forest types that compare the inventory with those of two Middle Nok are present in the Southern Guinea zone. These forests

332 Journal of African Archaeology Vol. 14 (3) Special Issue, 2016

Downloaded from Brill.com10/10/2021 06:57:41PM via free access The Palaeovegetation of Janruwa

Figure 1. Central Nigeria. Location of Janruwa C and neighboring sites (Map: E. Eyub). benefit from humid soil conditions but do not tolerate ties of the sites pearl millet (Pennisetum glaucum) and flooding. The dry forests or forest outliers are known in fonio (Digitaria exilis, D. iburua Stapf.) are widely Nigeria by the term “kurmi” and harbor species of the grown as well, even on poor soils and terraced slopes, drier parts of the lowland rain forest, e.g. Celtis zenkeri whereas the cultivation of yams is confined to better, Engl., Triplochiton scleroxylon K.Schum., Aubrevillea more profound soils. Manioc, rice and sugar cane are kerstingii (Harms) Pellegr. (Jones 1963). They are a mainly found in favorable locations such as valley very characteristic feature of the Southern Guinea zone bottoms and depressions, as in the vicinity of Janruwa and grow in moist but well-drained valleys and on the C, where rice is cultivated. Farming of maize and sor- raised banks of rivers. The light forests correspond to ghum is acknowledged by informants to be important the “forêt claire” of French publications. These forest in local agriculture. The spectrum of crop plants is patches contain a mixture of fire-tolerant and fire-tender diversified by Bambara groundnut Vigna( subterranea trees with no or only sparse grass growth, for instance (L.) Verdc.), cowpea (Vigna unguiculata), groundnut Sterculia tragacantha Lindl., Phyllanthus discoideus (Arachis hypogaea L.), gourds (Lagenaria siceraria (Baill.) Müll.Arg., Malacantha alnifolia (Baker) Pierre (Molina) Standl., Cucumis melo L.), and ginger (Zin- and Albizia zygia (DC.) J.F.Macbr. They resemble the giber officinale Roscoe) as the most important cash transition woodland between savanna and rain forest. Ac- crop. Land use includes husbandry of cattle, goat and cording to Keay (1953) this forest type might have been sheep, poultry keeping and the gathering of wild plants the “climax” vegetation of the Southern Guinea zone. like wild yams and a variety of tree fruits.

Where cultivation is possible, the current natural vegetation has been profoundly modified to wooded The site farmland, with fields including useful trees likeParkia biglobosa (Jacq.) G.Don (syn. P. oliveri J.F.Macbr.) Janruwa C is a flat site, located less than 40 m from and bush fallows, i.e. woodland in various stages of the perennial watercourse called Janruwa (Fig. 1) and regeneration. Tuber crops including yams, manioc and surrounded by hills. A unit of 8 x 12 m was excavated potato as well as maize dominate today’s agriculture of in 2010 to a depth of approximately one meter. Several this vegetation zone (Avav & Uza 2003). In the vicini- features were discerned during the excavation, with

Journal of African Archaeology Vol. 14 (3) Special Issue, 2016 333

Downloaded from Brill.com10/10/2021 06:57:41PM via free access A. Höhn & K. Neumann

feature 6

feature 8 7392 4738 10618

7341 8794 10617 4173

4738 8794 7392 7341 4173 10617 10618

Figure 2. Janruwa C. Location of archaeobotanical samples in the excavation, planum (above) and section (below). Large and bold number: charcoal analyzed. Red dot: radiocarbon date on archaeobotanical sample. (Map: T. Preuss, modified)

features 6 and 8 being pit-like structures (Fig. 2). The connected to small fragments possibly of terracotta, shallow feature 6 contained almost complete pots, suggests intrusional origin, possibly washed onto the grinding stone fragments, a stone axe and stone balls. site at a later time due to erosional processes. The pot- Feature 8 was larger and deeper. Besides pottery and tery inventory is dominated by carved wooden roulette stones, charred plant remains were present and most of decorations, thus classifying Janruwa C as a post-Nok the analyzable charcoal samples come from here. The site (Franke 2015). other features were mostly distinguished by stone set- tings (N. Rupp, pers. comm.). Four of five radiocarbon In order to compare the results of the charcoal results are calibrated to the first centuries CE and thus analysis from Janruwa C to material from Middle Nok date into the post-Nok period; only some grains of sites of the same regional setting, one sample each pearl millet (Pennisetum glaucum) from feature 6 gave from the sites of Janruwa A and B was analyzed. As a radiocarbon age in the early first millennium BCE, the results fitted in with those of other Middle Nok and thus in the Middle Nok phase (Tab. 1). Another sites (Kahlheber et al. 2009) no further samples were date, taken on fonio (Digitaria exilis) from a different analyzed for the time being. Janruwa A, excavated sample of the same feature 6 (Fig. 2), falls into the first in 2006, is located about 400 m to the southwest and centuries CE again. The single date into the Middle upslope of Janruwa C (Fig. 1). It is disturbed by illicit Nok phase, being from above the younger date and excavations of Nok terracotta and accordingly fragments

334 Journal of African Archaeology Vol. 14 (3) Special Issue, 2016

Downloaded from Brill.com10/10/2021 06:57:41PM via free access The Palaeovegetation of Janruwa

Site No. Site name C14 age Calibrated age* Dated material Lab No.

2006/04 Janruwa A 2542±32 bp 801–546 BC Pennisetum KIA 30267 2009/02 Janruwa B 2495±43 bp 792–432 BC Pennisetum MAMS 10744 2010/04 Janruwa C 2709±27 bp 906–810 BC Pennisetum MAMS 11160 2010/04 Janruwa C 1800±40 bp AD 94–338 Fonio Beta-278001 2010/04 Janruwa C 1799±24 bp AD 133–322 Charcoal MAMS 11162 2010/04 Janruwa C 1740±30 bp AD 236–386 Fonio Beta-297287 2010/04 Janruwa C 1719±24 bp AD 252–389 Charcoal MAMS 11161

Table 1. Radiocarbon dates of Janruwa A-C, mentioned in the text. Complete list of dates in Franke (2016). (* OxCal 4.2, IntCal13, 2-sigma).

of terracotta sculptures were found in the six test pits In Janruwa C most of the samples large enough for and scattered on the site (N. Rupp, pers. comm.). The expedient analysis came from Feature 8 (find no. 4738, Janruwa A ceramic inventory is clearly classified into 7341, 7392, 10617, 10618), only find no. 4173 and 8794 the Middle Nok phase (Franke 2015). The charcoal were taken from the matrix between the features (Fig. analyzed comes from unit 5 and was taken at 60 cm 2). The other archaeobotanical samples were small depth. Pearl millet from this sample was dated to the and did not contain many fragments large enough for first half of the first millennium BCETab. ( 1). charcoal analysis, even those taken from Feature 3, an assumed hearth. About 50 fragments per sample were Janruwa B was excavated in 2009. The site is situ- analyzed with a reflected-light microscope (Leica DM ated 700 m southwest and further upslope of Janruwa 4000M) after manually fracturing them along the three C (Fig. 1). A unit of 10 x 12 m was excavated reaching planes of wood (transverse, longitudinal tangential and sterile ground at about 60–70 cm. Pearl millet from longitudinal radial). The accumulation curves (Fig. 3) an archaeobotanical sample taken from a charcoal show that, even though the slope of the curve is flattened concentration was dated to the Middle Nok phase at about 30 fragments, single new types still turn up (Tab. 1); the analyzed charcoal belongs to this sample after 40 fragments. However, the number of fragments as well. Based on the pottery analysis, the site shows is regarded as sufficient, because the interpretation is two different occupation episodes, one in the Middle only based on the dominant charcoal types. Nok phase, the other in post-Nok times, but not related to the sample analysed (Franke 2015). Fragments with similar wood anatomical characters were grouped into types and the types identified. The The analysis of seeds and fruits shows clear differ- type was assigned a name using the InsideWood database ences: In Janruwa C fonio (Digitaria exilis) was present (InsideWood 2004-onwards), the reference collection of in all analyzed macrobotanical samples; fragments of microscopic wood slides of the Frankfurt Archaeobot- oil palm (Elaeis guineensis) endocarp were detected anical Laboratory and the Frankfurt DELTA anatomical in about half of them. Finds of these two species are database of African woods (Dallwitz 1980; Neumann only present in sites dating to post-Nok times. Addi- et al. 2001). In some cases identification was possible tionally Janruwa C furnished pearl millet and cowpea. only after SEM pictures had been taken to clarify the In Janruwa A and B only pearl millet was recorded (S. character of wood anatomical structures. The differences Kahlheber, pers. comm.). between charred and fresh wood (e.g. Prior & Gasson 1993) were taken into consideration during the identifica- tion process. The names of the charcoal types are given Material and methods in small capitals to discriminate the anatomical charcoal types from the botanical taxa, with whom they are not Seven charcoal samples from Janruwa C and one sample necessarily identical (Joosten & De Klerk 2002). each from Janruwa A and B were analyzed. The charcoal fragments derive from the large fraction (2 mm² mesh In order to draw conclusions towards the former size) of the flotated sediment samples. In Janruwa C vegetation at the sites, it is necessary to consider the between five and fourteen liters of sediment per sam- ecology of the species evidenced through the charcoal ple were processed; the sample of Janruwa A derived types. For each charcoal type all species with a similar from 20 liters, the one of Janruwa B from 15 liters of or identical wood anatomy and present in the vegetation sediment. of the Southern Guinea zone of Nigeria are considered.

Journal of African Archaeology Vol. 14 (3) Special Issue, 2016 335

Downloaded from Brill.com10/10/2021 06:57:41PM via free access A. Höhn & K. Neumann

Figure 3. Type accumulation curves of the analyzed samples.

The anatomy of the species comprised in one type is Xylopia spp., Annonaceae not necessarily identical, but the variability of wood Vessels solitary and in groups of up to 4; axial paren- anatomical features in certain species renders definite chyma in narrow 2-celled bands, scalariform; rays identifications of these species difficult and necessitates 4–7-seriate, mostly homocellular of procumbent cells, comprising them in charcoal types. occasionally few square cells; vessel-ray pitting similar to intervessel-pitting but elongated apertures on vessel All dates in the paper are given in the BCE/CE walls. The presence of homocellular rays eliminates system, whenever necessary uncalibrated dates were other Annonaceae than Xylopia. Xylopia parviflora calibrated with IntCal13 and OxCal 4.2 (Reimer et al. Spruce is widespread in tropical Africa; it prefers humid 2013; Bronk-Ramsey 2009). sites and is present in gallery forests in the Sudanian and Guinean zones. X. aethiopica (Dunal) A.Rich. is a lowland rain forest species that occurs in gallery forests Results in the savanna zones.

Short descriptions and discussion of identified charcoal types Chrysobalanaceae I (Fig. 4) Vessels solitary and some with a diameter larger than 200 All charcoal types present in the assemblages (Tab. 2) are µm; axial parenchyma apotracheal diffuse-in-aggregates briefly described based on IAWA Nomenclature (IAWA or in narrow lines, with ca 8 cells per strand; rays uniseri- Committee 1989) and discussed in alphabetical order of ate, some biseriate, homocellular of procumbent cells, the families they are assigned to. Most of the types com- containing abundant silica. Description and micrographs prise several species with similar wood anatomy (Tab. 3), of Parinari curatellifolia type in Höhn (2005). Maran- but species only present in rain forest are omitted, as rain thes polyandra (Benth.) Prance is present in Sudano- forest vegetation at Janruwa during the last 2500 years is Guinean savannas, P. curatellifolia Planch. ex Benth. highly improbable. The habitats of the species assigned also in the Sahelo-Sudanian zone. Neocarya macrophylla to each charcoal type are listed in the type descriptions. (Sabine) Prance ex F.White, Maranthes kerstingii (Engl.) If not stated otherwise, the ecological information is Prance ex F.White and Parinari congensis Didr. grow in retrieved from Keay (1989) and Arbonnier (2002). gallery forests in savanna regions.

336 Journal of African Archaeology Vol. 14 (3) Special Issue, 2016

Downloaded from Brill.com10/10/2021 06:57:41PM via free access The Palaeovegetation of Janruwa

Figure 4. Chrysobalanaceae I. a: transverse section, b: tangential longitudinal section, silica within ray cells, c & d: radial longitudinal sections, silica within ray cells.

Anogeissus leiocarpa, Combretaceae molle R.Br. ex G.Don, C. nigricans Lepr. ex Guill. & Perr.) Vessels mostly in groups of 2–3; axial parenchyma are typical for Sudanian and Guinean savannas and dry paratracheal scanty or vasicentric; rays uni- to biseri- forests on different soil types. ate, heterocellular, rows of square cells within mostly procumbent cells; prismatic crystals in radial alignment Terminalia spp., Combretaceae in the square ray cells. Detailed descriptions and micro- Vessels solitary and in groups; axial parenchyma para- graphs of the type in Neumann et al. (1998) and Höhn tracheal aliform, 4 and more cells per strand; rays uni- to (2005). Anogeissus leiocarpa (DC.) Guill. & Perr. is biseriate, heterocellular with procumbent and square/ present in most of the savanna areas from the semi-arid upright cells mixed. Detailed description and micro- Sahel to the forest-savanna boundary. It is typical in dry graphs of the type in Neumann et al. (1998) and Höhn forests and Sudano-Sahelian to Sudano-Guinean gallery (2005). Terminalia species are present in Sudanian to forests and tolerates temporary inundations. Guinean savannas, T. glaucescens Planch. ex Benth. and T. laxiflora Engl. belong to gallery forests, while Combretum spp., Combretaceae T. macroptera Guill. & Perr. is part of the vegetation Vessels solitary, in two sizes; axial parenchyma paratrache- in bas-fonds and like T. mollis M.A.Lawson can grow al aliform; rays uniseriate, heterocellular with procumbent on clayey soils with bad drainage. and square/upright cells mixed. On the single fragment of this type phloem and idioblasts in rays are not discernible. Alchornea spp., Euphorbiaceae Description and micrographs of this type inHubau (2013). Vessels solitary and in groups of regularly up to 4; The possible Combretum species (C.collinum Fresen., C. thin-walled fibers; rays uniseriate, homocellular of fragrans F.Hoffm. syn. C. adenogonium Steud. ex A.Rich., upright and square cells, some rays with radial canals, C. glutinosum Perr. ex DC., C. lecardii Engl. & Diels, C. then wider; vessel-ray pitting with reduced borders

Journal of African Archaeology Vol. 14 (3) Special Issue, 2016 337

Downloaded from Brill.com10/10/2021 06:57:41PM via free access A. Höhn & K. Neumann

Figure 5. Dialium spp. a: transverse section, b: radial longitudinal section, c & d: tangential longitudinal sections, silica in axial parenchyma cells. and enlarged (gash-like). Detailed description and Dialium spp., Fabaceae, Caesalpinioideae (Fig. 5) micrographs of the type in Eggert et al. (2006) and Vessels solitary and in groups; axial parenchyma in Höhn et al. (2007). Alchornea cordifolia (Schumach. bands, 3 (–4) cells wide; rays 2–3-seriate, homocellular & Thonn.) Müll.Arg. is present in Sudano-Guinean to of procumbent cells; rays and axial parenchyma storied; Guinean savannas, in gallery forests, on river banks silica in axial parenchyma. Dialium guineense Willd. is and in marshes. present in forest margins and in Sudano-Guinean and Guinean gallery forests. Caesalpinioideae I, Fabaceae, Caesalpinioideae Vessels solitary and in groups of 2–3, pits vestured; Pterocarpus spp., Fabaceae, Faboideae axial parenchyma paratracheal aliform and conflu- Vessels solitary and in groups; axial parenchyma ent; rays 1-3-seriate, mostly biseriate, homocellular paratracheal confluent and banded, 2 cells per strand; of procumbent cells; crystal chains in chambered rays uniseriate, homocellular of procumbent cells; rays axial parenchyma. Isoberlinia doka Craib & Stapf and and axial parenchyma storied. The type is described by I. tomentosa (Harms) Craib & Stapf are typical for the Neumann et al. (1998) and, accompanied by micro- Northern Guinean savannas of Nigeria (Keay 1953). graphs, as Pterocarpus lucens type by Höhn (2005). Berlinia grandiflora (Vahl) Hutch. & Dalziel occurs Pterocarpus erinaceus Poir. is the most widespread in gallery forests and borders of forests in Sudanian Pterocarpus in savanna areas. P. santalinoides DC. and Guinean zones. Erythrophleum africanum (Benth.) grows besides water courses in the Sudano-Guinean Harms grows in savanna woodland and light forests and Guinean zone, while P. lucens Guill. & Perr. is (“forêt claire”) in the Sudano-Guinean and Guinean typical for the northern Sahelo-Sudanian and Sudanian zone, E. suaveolens (Guill. & Perr.) Brenan is present savannas. in the drier parts of the rain forest and in gallery forests.

338 Journal of African Archaeology Vol. 14 (3) Special Issue, 2016

Downloaded from Brill.com10/10/2021 06:57:41PM via free access The Palaeovegetation of Janruwa

Figure 6. Millettia spp. a: transverse section, b: radial longitudinal section, c & d: tangential longitudinal section: rays, axial parenchyma and vessel elements storied.

Millettia spp., Fabaceae, Faboideae (Fig. 6) cf. Parkia spp., Fabaceae, Mimosoideae Vessels solitary and in small groups; axial parenchyma Vessels solitary and in groups, pits vestured; axial in bands, 5 and more cells wide, 2 cells per strand; rays parenchyma paratracheal aliform and confluent; fibres 3–5-seriate, homocellular of procumbent cells; rays, in radial alignment; rays 4–7-seriate, homocellular of axial parenchyma and vessel elements storied; crystals procumbent cells; long crystal chains in chambered in chambered axial parenchyma cells. A similar type is parenchyma. Description and micrographs of cf. Parkia described by Hubau (2013). Most of the Millettia species biglobosa type in Höhn (2005).The wood anatomy is belong to the forest zone, but Millettia thonningii (Schum. very similar to that of Afzelia and Guibourtia (Cae- & Thonn.) Baker is widespread and abundant in Nigeria salpinioideae), but due to the very wide rays Parkia in savanna and secondary forest, often on river banks. is more probable. Parkia biglobosa (Jacq.) G.Don is a widespread tree in Guinean and Sudanian savannas, Pericopsis spp., Fabaceae, Faboideae often on loamy and deep sandy soils. Vessels solitary and in groups; axial parenchyma paratracheal aliform and confluent; mostly 2 cells per cf. Khaya spp., Meliaceae strand; rays 2–3-seriate, homocellular of procumbent Vessels solitary and in small groups of 2–3; axial paren- cells; axial parenchyma and rays storied; crystals in chyma paratracheal scanty to vasicentric; rays mostly chambered axial parenchyma. 5–6-seriate, heterocellular with 1–2 rows of marginal cells; fibres septate; crystals in square/upright ray cells. Several Pericopsis species have a similar wood Description and micrographs of Khaya senegalensis anatomy (InsideWood 2004-onwards) but only Peri- type in Neumann et al. (1998). Khaya senegalensis copsis laxiflora is common in Sudanian and Guinean (Desv.) A.Juss. grows in savanna woodland of the Su- savannas, often on rocky or lateritic soils and poor danian to Guinean savannas and in transition types on fallows and in margins of gallery forests. the borders of the forest zone; it is confined to moist

Journal of African Archaeology Vol. 14 (3) Special Issue, 2016 339

Downloaded from Brill.com10/10/2021 06:57:41PM via free access A. Höhn & K. Neumann

Figure 7. Flueggea/Hymenocardia. a: transverse section, b: tangential longitudinal section, c & d: radial longitudinal section, vessel-ray pits simple. situations in the Sudanian zone. K. grandifoliola C.DC. Uapaca differs in having silica and lacking septate is present in the northern, drier part of the forest zone fibres. Hymenocardia acida Tul. is a common species and in the forest outliers of the moister savanna zones. of Sudanian and Guinean savannas, on more or less sandy soils. H. heudelotii Planch. ex Müll.Arg. prefers Lophira spp., Ochnaceae humid sites and is typical for Guinean gallery forests. Vessels solitary and in groups of 2–3; axial parenchyma Flueggea virosa (Roxb. ex Willd.) Royle is present in banded, 3–5 cells wide; 7–10 cells per strand; rays Sahelo-Sudanian to Guinean savannas. It is typical for 2–3-seriate, homocellular of procumbent cells. Lophira disturbed soils and fallows. lanceolata Tiegh. ex Keay is widespread and abundant in Sudano-Guinean to Guinean savannas. Phyllanthaceae I Vessels solitary and in groups of 2–3, diameter about Flueggea/Hymenocardia, Phyllanthaceae (Fig. 7) 150 µm; axial parenchyma paratracheal scanty; rays Vessels very small, diameter up to 50 µm, solitary and mostly 7–8-seriate, heterocellular, procumbent and in groups of 3 and more; axial parenchyma paratracheal square/upright cells mixed or 1–2 rows of marginal scanty; rays 3–5-seriate, heterocellular, procumbent cells, and uniseriate rays of square/upright cells; vessel- and square/upright cells mixed, and uniseriate rays ray pits with reduced borders, mostly gash-like; fibres of square/upright cells; fibres septate; vessel-ray pits septate. Hymenocardia cannot be excluded from this with reduced borders, mostly rounded; crystals in type definitely, but more probable are the following square/upright ray cells. These characters are typical taxa: Antidesma venosum E.Mey. ex Tul., Bridelia for Phyllanthaceae and the type strongly resembles spp. and Margaritaria discoidea (Baill.) G.L.Webster. Flueggea virosa/Hymenocardia acida type documented Antidesma venosum is present in gallery forests and in Höhn (2005). Phyllanthaceae I (see below) has marshes of the Sudanian and Guinean savannas. Most larger vessels and gash-like vessel-ray pitting, and of the Bridelia species also prefer humid sites and

340 Journal of African Archaeology Vol. 14 (3) Special Issue, 2016

Downloaded from Brill.com10/10/2021 06:57:41PM via free access The Palaeovegetation of Janruwa

Figure 8. Rubiaceae I. a: transverse section, b: tangential longitudinal section, c & d: radial longitudinal section, many rows of square/upright marginal ray cells. are present in Guinean savannas and in gallery forest The type corresponds to the Rubiaceae wood type I of of the Sudanian zone, some are present in fallows. Jansen et al. (2002). The charcoal type is described Margaritaria discoidea also prefers humid soils and as Feretia type in Neumann et al. (1998), and as Ru- belongs to gallery forests and forest margins in the biaceae I type in Höhn (2005). Several species of the Sudano-Guinean and Guinean zones. savanna zones with different ecological preferences are included in this type; some thrive along watercourses Uapaca spp., Phyllanthaceae or in swamps, others on dry or rocky soils. Vessels solitary and in groups of 2–3; axial parenchyma paratracheal scanty; rays 4–5-seriate, heterocellular; Sapindaceae I (Fig. 9) vessel-ray pits with much reduced borders, gash-like; Vessels solitary and in groups of 3–4; axial parenchyma silica in ray cells. Uapaca togoensis Pax grows in paratracheal scanty; rays uniseriate, homocellular of Sudano-Guinean and Guinean savannas and gallery procumbent cells; fibres septate; crystals in radial align- forests; it is typical for disturbed soils. U. heudelotii ment in ray cells. Blighia, Lecaniodiscus and Allophylus Baill. is present close to waters in forest regions but have a quite similar wood anatomy. Allophylus can be may occur in gallery forest in savanna regions as well. excluded from the type as crystals in ray cells are not described in InsideWood (2004-onwards); Blighia is Rubiaceae I (Fig. 8) coded there for both uniseriate rays and up to 3-seriate Vessels mostly solitary, diameter up to 50 µm, pits rays. Blighia sapida K.D.Koenig is present in drier vestured; axial parenchyma paratracheal scanty; rays forest areas and forest outliers in savanna regions. Le- 3–4-seriate, heterocellular with several rows of upright/ caniodiscus cupanoides Planch. ex Benth. grows in the square cells and low procumbent cells. It is difficult to understory of forests, forest outliers and gallery forests differentiate wood anatomically within the Rubiaceae. in Sudanian and Guinean savannas.

Journal of African Archaeology Vol. 14 (3) Special Issue, 2016 341

Downloaded from Brill.com10/10/2021 06:57:41PM via free access A. Höhn & K. Neumann

Figure 9. Sapindaceae I. a: transverse section, b: tangential longitudinal section, uniseriate rays, c & d: radial longitudinal sections, crystal (sheaths) in ray cells.

Sapotaceae square/upright cells; crystals in square/upright cells. Vessels mostly in groups of 2, tyloses present; axial Description and micrographs of Celtis integrifolia type parenchyma in narrow 2-celled irregular bands; rays in Neumann et al. (1998) and Höhn (2005). Celtis in- 2–4-seriate, heterocellular with several rows of upright/ tegrifolia Lam. is the common Celtis of savanna areas, square cells; some vessel-ray pits with reduced borders; from Sahelo-Sudanian to Guinean zones but in the drier silica in ray cells and parenchyma. Description and mi- savanna regions only along rivers and streams. crographs of Vitellaria paradoxa syn. Butyrospermum paradoxum types in Neumann et al. (1998) and Höhn (2005). Vitellaria paradoxa C.F.Gaertn. is locally abun- Composition of the samples dant in the Sudanian and Guinean savannas especially near towns and villages. It does not tolerate inundation. The dominant charcoal types at Janruwa C are Rubi- Manilkara obovata (Sabine & G.Don) J.H.Hemsl. is aceae I, Lophira spp., Anogeissus leiocarpa and Flu- present in evergreen forest but it is also a frequent eggea/Hymenocardia, though in varying abundance constituent of riparian forest in the savanna regions and and ubiquity (Tab. 2), with Rubiaceae I, Lophira spp. grows in forest outliers. Other taxa, e.g., Malacantha and Anogeissus leiocarpa recorded in every sample. alnifolia (Baker) Pierre or Pachystela pobeguiniana Flueggea/Hymenocardia type is present with the (Dubard) Pierre ex Lecomte occur in gallery forests in highest number of fragments in a single sample as Sudano-Guinean and Guinean savannas. well as in sum of all samples. However, with 4 to 22 fragments the number of fragments is highly variable Celtis spp., Ulmaceae from sample to sample. This is also true for other types, Vessels solitary and in groups of 2–4, tyloses present; e.g., for Lophira spp. with 2 to 15 fragments. Cf. Khaya axial parenchyma paratracheal aliform and conflu- spp., Chrysobalanceae I, cf. Parkia spp. and Caesal- ent; rays 4–5-seriate, heterocellular, with one row of pinioideae I are also present in many samples but with

342 Journal of African Archaeology Vol. 14 (3) Special Issue, 2016

Downloaded from Brill.com10/10/2021 06:57:41PM via free access The Palaeovegetation of Janruwa

Janruwa C Janruwa B Janruwa A Samples 7341 10617 10618 4738 7392 8794 4173 (2) AB 276 AB 35 Types feature 8 outside features sum ubiquity

Rubiaceae I 9 6 11 6 5 2 9 48 7 . . Lophira spp. 7 10 4 2 2 15 6 46 7 . . Anogeissus leiocarpa 2 6 2 9 2 12 2 35 7 . . Flueggea/Hymenocardia 4 5 10 7 18 . 22 66 6 13 . cf. Khaya spp. 7 8 3 2 1 4 . 25 6 2 5 Chrysobalanaceae I 2 2 1 1 3 9 . 18 6 . . cf. Parkia spp. 1 3 1 4 5 . 1 15 6 . 14 Caesalpinioideae I . 1 1 4 1 6 . 13 5 25 11 Alchornea spp. . 1 1 4 2 . . 8 4 . . Millettia spp. 1 1 1 . 2 . . 5 4 . . Dialium spp. 1 . 1 . . 1 1 4 4 . . Pericopsis spp. 5 1 1 . . . . 7 3 . . Uapaca spp. 3 . . 1 3 . . 7 3 3 4 Sapindaceae I . . . 1 . 1 1 3 3 . . Terminalia spp. 1 . . 2 . . 2 5 3 . 2 Pterocarpus spp. . . . 2 1 . . 3 2 . . Xylopia spp. 1 . 2 . . . . 3 2 . . Combretum spp...... 1 1 1 . . Phyllanthaceae I . . . 1 . . . 1 1 . . Sapotaceae . . . . . 1 . 1 1 . 3 Celtis spp...... 1 Indet. 5 6 13 5 5 7 6 7 7 Sum of fragments 49 50 52 51 50 58 51 361 50 47 Number of types 13 11 13 14 12 9 9 20 4 7

Table 2. Results of charcoal analysis.

fewer fragments. The composition of the samples from Ecological evaluation of charcoal types within Feature 8 and outside this feature does not differ markedly; the samples from Feature 8 are just slightly Lophira spp. and cf. Parkia spp. are the only charcoal more diverse (11 to 14 types) than the two other samples types which unambiguously represent savanna wood- with just nine types each. land. L. lanceolata and Parkia biglobosa are very typical of the Sudano-Guinean and Guinean savannas The composition of the Middle Nok samples and do not thrive in forests. The other charcoal types from Janruwa A and B (AB 35 and AB 276) clearly of Janruwa C are ambiguous concerning information differs from that of the Janruwa C samples. Except value on former vegetation, as they either contain one or for Flueggea/Hymenocardia, the dominant types more species with a wide distribution in all vegetation at Janruwa C (Anogeissus leiocarpa, Lophira spp., types, like Anogeissus leiocarpa, or several species Rubiaceae I) are missing. Instead, the Fabaceae with with different habitat preferences, e.g. as within the Caesalpinioideae I, and cf. Parkia spp. in Janruwa A Rubiaceae I. Accordingly, the charcoal types cannot are present in higher numbers. In Janruwa B cf. Parkia be related to just one vegetation type (Tab. 3). This spp. is replaced by Hymenocardia/Flueggea. Celtis problem amplifies with increasing species richness. spp. is exclusively found in Janruwa A. The two samples Thus, in the Southern Guinea Zone of Nigeria, a direct are also less diverse than those of Janruwa C with only connection between charcoal type and presence of a four, respectively seven types (Tab. 2). Even though the certain vegetation cannot be drawn unambiguously, fragments were more eroded than those from Janruwa while in environments with low tree species diversity, C, the lower number of types is not related to preserva- like the Sahel, a correlation of charcoal types and tion as the number of indeterminates is not higher than habitat is much easier. It is also possible that within in the Janruwa C samples. some types the fragments of one sample or one site

Journal of African Archaeology Vol. 14 (3) Special Issue, 2016 343

Downloaded from Brill.com10/10/2021 06:57:41PM via free access Charcoal type Probable species A. Höhn & K. Neumann Gallery Savanna Light/dry Use as fuel forests woodland forests wood/charcoal

Rubiaceae I Crossopteryx febrifuga l l Feretia apodanthera l l Gardenia aqualla l Gardenia erubescens l Gardenia sokotensis l Gardenia ternifolia l Ixora brachypoda l Mitragyna inermis l l Morelia senegalensis l Pavetta cinerifolia l Pavetta corymbosa l l l Pavetta crassipes l Tricalysia okelensis l l l Lophira spp. Lophira lanceolata l l Anogeissus leiocarpa Anogeissus leiocarpa l l l l Flueggea/Hymenocardia Flueggea virosa l l Hymenocardia acida l l Hymenocardia heudelotii l cf. Khaya spp. Khaya grandifoliola l Khaya senegalensis l l l Chrysobalanaceae I Maranthes kerstingii l Maranthes polyandra l l Neocarya macrophylla l l l Parinari congensis l Parinari curatellifolia l l l cf. Parkia Spp. Parkia biglobosa l l Caesalpinioideae I Berlinia grandiflora l Erythrophleum africanum l l l Erythrophleum suaveolens l l l Isoberlinia doka l l l Isoberlinia tomentosa l l l Alchornea spp. Alchornea cordifolia l l Millettia spp. Millettia thonningii l l l Dialium spp. Dialium guineense l l l l Pericopsis spp. Pericopsis laxiflora l l l l Uapaca spp. Uapaca heudelottii l l Uapaca togoensis l l l Sapindaceae I Blighia sapida l l Lecaniodiscus cupanioides l l l Terminalia spp. Terminalia albida l l l Terminalia avicennioides l l l Terminalia glaucescens l l l Terminalia laxiflora l l l Terminalia macroptera l l l Terminalia mollis l l Pterocarpus spp. Pterocarpus erinaceus l l Pterocarpus santalinoides l l Xylopia spp. Xylopia aethiopica l l l Xylopia parviflora l l Combretum spp. Combretum collinum l l l Combretum fragrans l l l Combretum glutinosum l l l Combretum lecardii l Combretum molle l l l Combretum nigricans l l l Phyllanthaceae I Antidesma venosum l l l Bridelia atroviridis l l Bridelia ferruginea l l Bridelia micrantha l l l Bridelia scleroneura l l Margaritaria discoidea l l l l Sapotaceae Malacantha alnifolia l l Manilkara obovata l l l Pachystela pobeguiana l l Vitellaria paradoxa l l Celtis spp. Celtis integrifolia l l l

Table 3. Species comprised in the respective charcoal types; their aptitude as fuelwood/charcoal and the occurrence in three different vegetation types of the Southern Guinea zone is given: gallery forest, savanna woodland and dry/light forest, based on 344 Journal of African Archaeology Vol. 14 (3) Special Issue, 2016 Keay (1989), Burkill (1985–2000) and Arbonnier (2002).

Downloaded from Brill.com10/10/2021 06:57:41PM via free access The Palaeovegetation of Janruwa must not necessarily represent one species. Fragments Many species characteristic for the two forest with similar anatomy may be derived from two or more types are represented by charcoal fragments in the species, especially when the assemblages do not result assemblage of Janruwa C (Tab. 4). Several charcoal from one fire event, but are rather dispersed charcoals types hint towards the presence of an open type of from many hearth fires. Correlating vegetation types to forest constituting the border of gallery forests or charcoal types may even be difficult on high taxonomic forest outliers to the adjacent savanna. Typical for the levels as with Anogeissus leiocarpa for instance. boundaries of gallery forests are light-demanding spe- Just one matching species belongs to this type, but cies such as Anogeissus leiocarpa, Mitragyna inermis unfortunately A. leiocarpa has a wide ecological range and Alchornea cordifolia (Natta & Porembski 2003). and occurs in several different vegetation types in the In the assemblage of Janruwa C Anogeissus leiocarpa Southern Guinea savanna zone. Thus, in order to draw and Rubiaceae I (including M. inermis) are constantly conclusions about the former vegetation we have to present in higher numbers, possibly due to their quality argue with possibilities and rather look at combinations as fuelwood, while Alchornea spp. (with A. cordifo- of charcoal types and included species. lia), which is not mentioned for fuelwood qualities, is only present in fewer samples and lower numbers. An Most of the species represented by the charcoal additional species occurring in the transition from gal- types of the assemblage are known sources of fuel- lery forests to savanna according to Natta & Poremb- wood or charcoal production (Tab. 3). Exceptions are ski (2003) and present in the assemblage is Millettia Millettia thonningii and Alchornea cordifolia. Within thoningii (as Millettia spp.). Additional widespread the Rubiaceae I only three species are sources of fuel species of riparian forest are Manilkara multinervis, or charcoal. It is more probable that these species can Xylopia parviflora, Lecaniodiscus cupanoides, Khaya be found in the assemblage than those not being used grandifoliola and K. senegalensis (Natta et al. 2002). for fuelwood today. Some of them are rather typical of gallery forests of the Northern Guinean zone or in the transition to Sudanian The interpretation is based on the assumption that climates, but as soil, light regime, and flood frequen- species with good fuelwood qualities that were available cies play a role in the gallery forest compositions, in the vicinity of the site are present in the assemblage the charcoal fragments, possibly representing these constantly. The number of fragments is less important. species, might have come from gallery forest trees at Different fragmentation rates of charcoals dependent Janruwa C. Also Terminalia macroptera and T. glauce- on the taxon are an issue (Chrzazvez et al. 2014), but scens are listed as characteristic for gallery forests in as relative variations are not calculated here, the bias Nigeria (Keay 1953). Typical species of the riverfront is expected to be negligible. and the central part of gallery forests in the Guinean zone like Dialium guineense, Parinari congensis and Pterocarpus santaliniodes (Natta & Porembski 2003) Discussion are represented by charcoal types as well. Especially Chrysobalanceae with Parinari congensis is almost The environment of Janruwa C (100 – 400 CE) constantly present in the assemblage. Natta & Poremb- ski (2003) name nine tree species, out of a recorded Gallery forest or dry forest 200 species, that contribute to almost 50 % of the total The charcoal assemblage of Janruwa C points to the ex- abundance and cover more than 50 % of the basal area ploitation of two vegetation types — savanna woodland in gallery forests of Benin. Out of these nine species and some type of forest. The forest could have been either five are documented in the archaeobotanical remains gallery forest or a “young forest”, typically surrounding of Janruwa C. Four are represented by charcoal the dry forests of the Southern Guinea zone, which are types: Pterocarpus santalinoides (Pterocarpus spp.) also called forest outliers or “kurame” (Keay 1953; Kil- Dialium guineense, Uapaca togoensis (Uapaca spp.), lick 1959; Jones 1963). The terrain at Janruwa C allows Berlinia grandiflora (as Detrieae I) and one, Elaeis for both options: the watercourse just below the site is guineensis, is represented by endocarp fragments. But accompanied by a gallery forest, which could have been within the assemblage these charcoal types are not the much wider under less anthropogenic pressure in the past, most dominant and the other four species noted by and the position at a valley head is likewise typical for the Natta & Porembski (2003) are missing, namely Cola presence of forest outliers (Jones 1963). The charcoal as- laurifolia, Syzygium guineense, Cynometra megalo- semblage indicates that forest was present and exploited phylla and Diospyros mespiliformis. by the people at Janruwa C, regardless of whether it was gallery forest or forest outliers or a mixture of both. The Characteristic for the margins of forest outliers existence of this forest was related to the favorable con- are Anogeissus leiocarpa, Malacantha alnifolia, Le- ditions in the valley with deeper soils and more humid caniodiscus cupanioides, Khaya grandifolia, Dialium conditions than on the slopes of the surrounding hills. guineense and Berlinia grandiflora (Jones 1963). But

Journal of African Archaeology Vol. 14 (3) Special Issue, 2016 345

Downloaded from Brill.com10/10/2021 06:57:41PM via free access Charcoal type Possible species Savanna Gallery Gallery Forest Savanna Present at A. Höhn & K. Neumann woodland1 forests1 forests2 outliers3 fallows Janruwa today

Rubiaceae i Crossopteryx febrifuga ? l Feretia apodanthera Gardenia aqualla Gardenia erubescens l Gardenia sokotensis Gardenia ternifolia Ixora brachypoda Mitragyna inermis l m l Morelia senegalensis Pavetta cinerifolia Pavetta corymbosa Pavetta crassipes Tricalysia okelensis Lophira spp. Lophira lanceolata l l Anogeissus leiocarpa Anogeissus leiocarpa m u l Flueggea/hymenocardia Flueggea virosa l Hymenocardia acida l l Hymenocardia heudelotii cf. Khaya spp. Khaya grandifoliola u Khaya senegalensis Chrysobalanaceae i Maranthes kerstingii Maranthes polyandra Neocarya macrophylla Parinari congensis l Parinari curatellifolia cf. Parkia spp. Parkia biglobosa l l Caesalpinioideae i Berlinia grandiflora l l u l Erythrophleum africanum Erythrophleum suaveolens Isoberlinia doka Isoberlinia tomentosa Alchornea spp. Alchornea cordifolia m l Millettia spp. Millettia thonningii m Dialium spp. Dialium guineense l/m u Pericopsis spp. Pericopsis laxiflora l l Uapaca spp. Uapaca heudelottii Uapaca togoensis l l Sapindaceae i Blighia sapida Lecaniodiscus cupanioides l u Terminalia spp. Terminalia albida l species Terminalia avicennioides l Terminalia glaucescens l l Terminalia laxiflora Terminalia macroptera l Terminalia mollis Pterocarpus spp. Pterocarpus erinaceus Pterocarpus santalinoides l l Xylopia spp. Xylopia aethiopica Xylopia parviflora l Combretum spp. Combretum collinum species Combretum fragrans Combretum glutinosum Combretum lecardii Combretum molle Combretum nigricans Phyllanthaceae i Antidesma venosum l/m Bridelia atroviridis Bridelia ferruginea l l Bridelia micrantha l Bridelia scleroneura Margaritaria discoidea Sapotaceae Malacantha alnifolia u Manilkara multinervis l/m Manilkara obovata Pachystela pobeguiana Vitellaria paradoxa l Celtis spp. Celtis integrifolia

Table 4. Species characteristic for the possible vegetation types in the vicinity of Janruwa C in the past and species present today. Other346 than in Table 3 not all possibleJournal habitats of areAfrican indicated Archaeology but rather Vol. 14if a(3) species Special is Issue, characteristic 2016 for a certain vegetation type (m: border of gallery forest with savanna; u: young forest around forest outlier; 1 Keay 1953; 2 Natta & Porembski 2003; 3 Jones 1963). Downloaded from Brill.com10/10/2021 06:57:41PM via free access The Palaeovegetation of Janruwa again, some typical species, which could strengthen However, some typical savanna woodland species the arguments for the presence of a forest outlier, are like Vitex doniana, Detarium microcarpum and Daniel- missing in the assemblage, for instance Chlorophora lia oliveri (Keay 1953) are missing in the assemblage. excelsa or Triplochiton scleroxylon, both typical of the It is puzzling that especially D. oliveri is missing, center of a “kurmi”. We conclude that more charcoal because today it is the most abundant and typical of types seem to represent gallery forest species (Tab. 4). all savanna woodland species of the Guinean savanna (Keay 1953). It is common in fallow regrowth and the Savanna woodland wood is valued as good-quality fuelwood. Moreover, In the vicinity of the site, savanna woodland was the wood anatomy of this taxon is very characteristic present and exploited as well. Due to the location in and easily recognizable (Höhn 1999). the valley and the proximity of the river, the site could have been located within the forest or at its margins, but savanna was visited and exploited frequently. The Middle Nok environment at the location of the This is documented by the prevalent Lophira spp. Janruwa sites (A & B) fragments in the assemblage and supported by the similarly prevalent Hymenocardia/Flueggea frag- Savanna woodland was present around the Janruwa sites ments. Hymenocardia acida and Flueggea virosa are at Middle Nok times as well as at least 500 years later; the most probable sources of the latter type as both are however, the character of the woodland exploited was known fuelwood resources. H. acida belongs to the different. In the Middle Nok samples of Janruwa A and B, typical species of the savanna woodland of the South- the strong presence of Caesalpinioideae I rather points ern Guinea Zone of Nigeria (Keay 1953). Another hint to Isoberlinia woodland as source of fuelwood as op- for savanna woodland is the almost constant presence posed to the woodland dominated by Lophira lanceolata of cf. Parkia spp. (Tab. 2). Parkia biglobosa is typical as part of the exploited vegetation at the post-Nok site of savanna woodland in the Southern Guinea zone, Janruwa C. Patches of Isoberlinia woodland occurred together with H. acida, Peterocarpus erinaceus and naturally on shallow soils in the Guinea-Congolia/ Terminalia glaucescens (Fairhead & Leach 1996). Sudanian regional transition zone in the past (White The presence of P. biglobosa in the fields is encour- 1983) and recent present (Keay 1953). They grow on aged by the farmers because of its manifold benefits rocky hilltops in the Southern Guinea zone. Isoberlinia as a useful tree: today the pods or parts of it are woodland is composed of species more typical and abun- worked into several foodstuffs, the foliage is browsed, dant in the drier Northern Guinea zone, e.g., Isoberlinia the litter improves the soil and different parts of the spp. and Uapaca togoensis and, at its best, this savanna tree have medicinal properties (Keay 1953, Burkill woodland has a thin but continuous canopy of I. doka and 1985–2000). Lophira lanceolata and H. acida thrive I. dalzielii together with some Pterocarpus erinaceus, in the fallows resulting from shifting cultivation. Uapaca togoensis and Lannea spp. (Keay 1953). Only Both tree species are capable of sucker growth and small fires pass through as the amount of combustible can easily be cut for fuel when preparing the field in material is low (Jones 1963). Besides the protection from an old fallow. The fragments of Rubiaceae I could severe fire on the hilltops, the reason for the growth of derive from Crossopteryx febrifuga, a typical species the fire-sensitive Isoberlinia on the rocky hilltops may of savanna woodland (Keay 1953) and with good be the drier conditions on the more shallow soils of the fuelwood qualities as well. C. febrifuga also grows hills. It is even possible that the Isoberlinia woodlands in the savanna adjacent to the forest outliers and is on the hilltops are the consequence of shifting cultivation one of the principal components of degraded savanna, and that Middle Nok Isoberlinia woodlands were already together with Hymenocardia acida, Butyrospermum the floristically impoverished state of light forest, the paradoxum (syn. Vitellaria paradoxa), Terminalia postulated “climax” vegetation (Keay 1953). avicennioides and Lophira lanceolata (Jones 1963). Still, there is no evidence in the assemblage for strong Caesalpinioideae I, interpreted as deriving from human impact through farming, because only few of Isoberlinia spp. in the case of Janruwa A and B, is not the represented species are listed as typical fallow con- unambiguous. Based on the results of the single samples stituents. From the dominant types of the assemblages of Janruwa A and B alone, it would have been difficult only the fragments of Lophira spp. and Flueggea/ to conclude from the presence of Caesalpinioideae I Hymenocardia could be derived from fallow trees to Isoberlinia woodlands. But because these assem- and shrubs (Tab. 4). Parinari curatellifolia, just like blages are quite similar to other Middle Nok samples Parkia biglobosa, is more typical of the Northern where the combination of types suggested Isoberlinia Guinea zone but occurs in the more southern savannas woodlands (Kahlheber et al. 2009), it is probable as well (Keay 1953). All these species are represented that Caesalpinioideae I stands for Isoberlina spp. in by charcoal types (Tab. 3 and 4). Janruwa A and B as well.

Journal of African Archaeology Vol. 14 (3) Special Issue, 2016 347

Downloaded from Brill.com10/10/2021 06:57:41PM via free access A. Höhn & K. Neumann

Figure 10. West Central Africa. Location of Janruwa C and palaeoecological sites mentioned in the text (Map: E. Eyub).

Environmental changes? What happened between the from Middle Nok to the first centuries CE. Thus the Middle Nok phase and early post-Nok times? only difference between the sites is their location, their material culture and subsistence strategy: Janruwa C The different assemblages, including other Mid - was located in the valley and the economy was based dle Nok sites in the research area ( Kahlheber et on a wider spectrum of crops as well as oil palm. al. 2009), indicate a shift from the exploitation of Isoberlinia woodland during the Middle Nok phase Due to the lack of suitable sediments, no pol- to the exploitation of forest in the vicinity of the len data are available for the Late Holocene in the watercourse and of savanna woodland with Lophira research area. However, palynological proxies from lanceolata in early post-Nok times. These vegetation other sites in southern West Africa indicate that no types are generally comparable to those described by significant vegetation changes occurred during the first Keay (1953) for the Southern Guinea vegetation zone. millennium BCE and the first centuries CE: Favorable Most of the charcoal types of the assemblages are even conditions for tree growth are recorded from the Da- today represented by species in the vicinity of the sites homey Gap at Lac Sélé (Benin) as well as from Lake (Tab. 4). The vegetation in the area probably did not Bambili (Cameroon) (Fig. 10). In the Dahomey Gap, change markedly between Middle Nok and the time a forest-savanna mosaic was continuously present be- of Janruwa C. Instead, the differences of the assem- tween 1400 BCE and 900CE, and pioneer forests only blages seem to be due to the locations of the sites: The disappeared long after Middle Nok times (Salzmann people collected wood in the vegetation type closest & Hoelzmann 2005). Around Lake Bambili, pioneer to their site — mainly Isoberlinia spp. on the slopes trees returned between 700 BCE and 500 CE after at Janruwa A and B versus several species from the having been absent for more than half a millennium savanna woodland and the adjacent diverse forest in (Lézine et al. 2013). At Nsukka (Fig. 10), just 300 km the valley in Janruwa C. We conclude that the general south of the research area, a palynological site shows distribution of Isoberlinia woodlands on the hilltops Late Holocene vegetation changes only during the first and Lophira woodlands in the valley and adjacent to half of the first millennium CE (Njokuocha 2012; the forests along the watercourse remained similar Njokuocha & Akaegbobi 2014).

348 Journal of African Archaeology Vol. 14 (3) Special Issue, 2016

Downloaded from Brill.com10/10/2021 06:57:41PM via free access The Palaeovegetation of Janruwa

However, lake level fluctuations at Lake Bosumtwi The climatic fluctuations responsible for lake level (Ghana) (Shanahan et al. 2006, 2008, 2013) indicate variations are obviously not reflected in vegetation changes that the climate in the first millennium BCE was not as in all parts of West Africa during this period and they did stable as the vegetation data indicate. Although precipita- not markedly affect the woody vegetation at the Janruwa tion resulting in high water levels is estimated to having sites. This can be explained by a strong resilience of the been only slightly higher (0.1–2.4 %) than today, temper- savanna vegetation in the Guinean zone, which is able to ature, cloudiness and seasonality variations might have buffer short-term variations in rainfall and seasonality. influenced evaporation and could have been responsible for century-long high stands of the lake level between 500 But there is strong evidence of erosion in the Nok BCE and 500 CE and a marked low stand around 400 region. Large-scale denudation must have taken place BCE — against the background of a progressive drying after or during the Nok period as evidenced by the up of the lake between 660 and ~1000 CE (Shanahan context of the first finds of terracotta sculptures. These et al. 2006, 2008). The climatic fluctuations responsible were discovered in tin-bearing gravel under sand and for lake level variations are obviously not reflected in several meters of alluvia of grey or blue clay during vegetation changes in most parts of West Africa during tin mining to the east of Janruwa (Fagg 1945, 1956; this period because local features such as soils, topogra- Bond 1956). The terracotta fragments had been washed phy and also internal biotic factors determine responses away from the original sites and deposited in erosion and resilience of African biota, especially savannas, to channels, which afterwards filled up with sandy and climate change (Willis et al. 2013). clayey sediments. Charcoal found in the terracotta- containing gravel layer as well as in the sandy layers The concomitance of a dramatic lowstand in Lake above gave very old radiocarbon dates between at least Bosumtwi (Shanahan et al. 2009, 2013) around 400 2000 BCE and the beginning of the first millennium BCE and the decline in the number of radiocarbon dates BCE (Barendsen et al. 1957). The dates are thus at at the end of the Middle Nok phase (Franke & Breunig least partly too old, which can be explained through 2014; Franke 2016) is striking. Evidence for climatic relocation of considerably older wood together with change that might have been responsible for the lowstand the terracotta figurines. Dating charcoal from riverine of Lake Bosumtwi and also involved in the decline of deposits can be misleading as the events leading to the Nok Culture can be found in the pollen diagram from charcoal deposition in valley-floor sampling sites may Nyabessan, ca 800 km to the Southeast of Janruwa C, in form lengthy histories (Frueh & Lancaster 2014). the rainforest of southern Cameroon (Fig. 10). While at Lac Sélé and Lake Bambili no vegetation changes could A terminus ante quem for the deposition of the be detected, the pollen diagram of Nyabessan shows a gravel deposit is provided by the radiocarbon date of the clear break at 400 BCE (Ngomada et al. 2009). The blue clay above the coarser material in the first centuries rainforest was opened and pioneer trees became suddenly CE (Barendsen et al. 1957). Even though the blue dominant. The opening of the forest is also visible in clay and the dated charcoals may have been relocated Lake Barombi Mbo (e.g., Maley & Brenac 1998; Ma- much later, a similar age delivered by charcoal from the ley 2001). Correlating the pollen evidence with diatom lower parts of alluvia to the east at Choribariki (H.-M. data from Lake Ossa (Fig. 10; Nguetsop et al. 2004) Peiter, pers. comm.) supports the hypothesis that the and sea surface temperature (SST) variations in the Gulf gravel deposit with the Nok figurines must have been of Guinea (Weldeab et al. 2005, 2007), Ngomanda et deposited either during or shortly after the Middle Nok al. (2009) claim that a different seasonality pattern was phase. Because there is no evidence for climatic change responsible for the vegetation changes. Due to a more nor for any cultural or environmental ruptures during southern position of the Intertropical Convergence Zone Middle Nok times, the most parsimonious hypothesis is (ITCZ) during north-hemispheric winter, the dry season that the erosion processes happened around or shortly was longer than today. Strong northern trade winds after 400 BCE, contemporaneous with the lowstand of brought dry air further to the south for a longer period of Lake Bosumtwi, and were related to distinct changes in the year. Furthermore, the sharp increase in SSTs reduced seasonality visible in the pollen diagram of Nyabessan. the stratiform cloud cover during north-hemispheric summer, leading to increased insolation during the Erosion after 400 BCE is also witnessed by an short dry season in regions with two rain maxima. The increase of the sedimentation rate in the upper part of combination of these effects, high evapotranspiration pollen core taken from Lake Tilla to the Northeast of Jan- during the short summer dry season and a prolonged ruwa C (Fig. 10; Salzmann 2000). Maley & Brenac winter dry season, could have also led to the lowstand at (1998) and Maley (2001) relate increased erosion in Lake Bosumtwi from 400 to almost 200 BCE recorded West Central Africa during the first millennium BCE to by Shanahan et al. (2013) and the lake level decrease the abrupt reduction of stratiform clouds and an increase at Lake Ossa, in spite of simultaneously higher annual of cumuliform clouds which are associated with more rainfall (Nguetsop et al. 2004). and heavier rain due to the changed SSTs.

Journal of African Archaeology Vol. 14 (3) Special Issue, 2016 349

Downloaded from Brill.com10/10/2021 06:57:41PM via free access A. Höhn & K. Neumann

However, in the Nok region the base for the later some “kurame” of the Abuja district, settlements within land erosion probably was laid during the Middle Nok the forest were still surrounded by defensive banks and phase for which intensive settlement activities are ditches 20 or 30 m inside the forest edge as a means attested by the high number of archaeological sites, of protection; their foundation reaching back probably including iron-smelting sites (Breunig & Rupp 2016). about 100 years (Jones 1963). These advantages, how- Clearing of Isoberlinia woodlands on the slopes for ever, had to be traded off with higher risks of illnesses plant cultivation and removal of wood for iron smelt- like trypanosomiasis and onchocercosis. ing could have opened the vegetation, without deeply changing its species composition, as Isoberlinia is able to resprout vigorously from the stump. When dry Conclusion seasons became longer and the rains were concentrated in shorter periods, the opened landscape facilitated the More and more a picture emerges that the people of the loss of topsoil. Especially at the beginning of the rainy Middle Nok period were remarkably conservative. This season, when the ITCZ passed through the region and is evident in their material culture, terracotta figurines, topsoil laid bare after a long dry period, the surface land-use systems and cultivated crops. Middle Nok material could have been washed off. On the eventually people seem to have avoided riverine environments. denudated slopes pearl millet cultivation was prob- Although oil palms have a long history of use in West ably still possible but with much lower yields. It is not Africa (Lavachery 2001; D’Andrea et al. 2006; Logan plausible that the stronger seasonality directly affected & D’Andrea 2012) and can be found in the gallery for- the yields because pearl millet and cowpea are savanna est of Janruwa even today, their endocarps are absent crops: they should have been able to cope with a shorter in Middle Nok archaeobotanical assemblages, in spite rainy season and a longer dry season. of their good preservation abilities.

In line with this hypothesis the presence of fonio in We conclude that the restricted land-use system of Janruwa C and other later sites points to a new system the Middle Nok people was not flexible enough to cope of crop cultivation. The cultivation of fonio could have with soil erosion after 400 BCE. Their narrow set of been a way of coping with the eroded soils on the slopes. crops, restricted to pearl millet and cow pea, prevented Fonio has a broad ecological tolerance; it can be grown adaptation to new environmental conditions. The impact on poor shallow soils unsuitable for other cereals, but of their shifting cultivation system on the vegetation also tolerates waterlogged soils (National Research was low as long as the climate remained stable. But the Council 1996: 59–75; Vodouhè & Achigan Dako system they were practicing successfully for at least 600 2006). Thus, it could have been cultivated on the shal- years was laying the base for land degradation when low soils of the slopes, maybe together with pearl millet climate changed around 400 BCE. which is present in the Janruwa C assemblage as well (Kahlheber et al. 2009; Kahlheber 2010), but also Based on the results from Janruwa C, we sup- on the deeper soils of the valley where waterlogging pose the following scenario for the decline of the Nok may occur. Culture around 400 BCE and its disappearance around the turn of the Common Era. Towards the end of the At Janruwa C the people settled in the valley dur- Middle Nok phase, the shift of the ITCZ, as witnessed ing the first centuries CE. They took advantage of the in Nyabessan, led to a prolonged dry season during the deep and rich soils ensuring good harvests. Probably north hemispheric winter. The woodland, opened for they also had a greater knowledge of forest and forest- cultivation and iron-smelting, was not able to prevent related resources than the Middle Nok people because soil erosion during the strong rainfall events concen- they made use of the oil palm growing in the gallery trated in a shorter rainy season. Due to reduced yields forest. As evidenced by the charcoal assemblages, they of pearl millet on the shallow soils, the number of sites collected their fuelwood in savanna and forest. declined at the end of Middle Nok. The charcoal sam- ples of Janruwa C and their comparison with Middle Living in the valley offered some additional ad- Nok assemblages indicate that, like at Lac Sélé and vantages, like the presence of water, fish and wildlife Lake Bambili, the woody vegetation did not change in the immediate vicinity of the site. Moreover, the markedly for several hundred years; the differences forest offers better protection from natural threats like between Middle Nok assemblages and those dating high winds, excessive heat and bush fires during the dry to the first centuries CE are only due to different site season; the latter being particularly important for houses preferences — on the hill versus in the valley. with a thatch roof that can easily catch fire from flying sparks (Fairhead & Leach 1996). The forest may even Between 100 and 400 CE, after the end of the Nok serve as a defense to assaults from enemies: In 1948 in Culture, people moved their settlement closer to the

350 Journal of African Archaeology Vol. 14 (3) Special Issue, 2016

Downloaded from Brill.com10/10/2021 06:57:41PM via free access The Palaeovegetation of Janruwa river. New pottery decorations and styles, the absence Breunig, P. & Rupp, N. 2016. An outline of recent studies on the of terracotta figurines, a wider variety of cultivated Nigerian Nok Culture. Journal of African Archaeology 14 (3), crops (fonio) and oil palm collected in the forest suggest 237–255, this issue. significant cultural changes - whether by the arrival of Bronk Ramsey, C. 2009. Bayesian Analysis of Radiocarbon populations from outside the Nok region or developed Dates. Radiocarbon 51 (1), 337–360. autochthonously after the breakdown of the former ritual concept and the associated life style, has to remain Burkill, H.M. 1985–2000.The Useful Plants of West Africa, 1–5. unanswered for the time being. Royal Botanic Gardens, Kew. Chrzazvez, J., Théry-Parisot, I., Fiorucci, G., Terral, J.-F. & Thibaut, B. 2014. Impact of post-depositional processes on Acknowledgements charcoal fragmentation and archaeobotanical implications: experimental approach combining charcoal analysis and bio- Funding was provided by the Deutsche Forschungsge- mechanics. Journal of Archaeological Science 44 (1), 30–42. meinschaft (DFG) within the frame of the long-term http://dx.doi.org/10.1016/j.jas.2014.01.006 project on the Nok Culture. We thank our Nigerian Dallwitz, M.J. 1980. A general system for coding taxonomic de- partners, especially the National Commission for Mu- scriptions. Taxon 29, 41–46. http://dx.doi.org/10.2307/1219595 seums and Monuments. We appreciate the provision with samples through the archaeological team, assisted D’Andrea, A.C. Logan, A.L. & Watson, D.J. 2006. Oil palm and by Fatima Nijimi. We would like to thank Jennifer prehistoric subsistence in tropical West Africa. Journal of Afri- Markwirth, Manfred Ruppel and Barbara Voss for tech- can Archaeology 4, 195–222. http://dx.doi.org/10.3213/1612- 1651-10072 nical assistance, Richard J. Byer for language editing, Eyub Eyub for geographical information and maps and Eggert, M.K.H., Höhn, A., Kahlheber, S., Meister, C., Neumann, K. Thomas Preuss for the map of the excavation. Thanks & Schweizer, A. 2006. Archaeological and archaeobotanical re- are also due to Barbara Eichhorn and Gabriele Franke search in Southern Cameroun. Journal of African Archaeology for fruitful discussions and comments on the text and 4 (2), 273–298. http://dx.doi.org/10.3213/1612-1651-10076 to Aline Garnier for the translation of the abstract into Fagg, B. 1945. A preliminary note on a new series of pot - French. tery figures from Northern Nigeria. Africa: Journal of the International African Institute 15, 21–22. http://dx.doi. org/10.2307/1156826

References Fagg, B. 1956. An outline of the Stone Age of the Plateau Mines- field. In:Proceedings of the Third International West African Abaje, I.B., Ishaya, S. & Usman, S.U. 2010. An analysis of rainfall Conference, Ibadan, Nigeria, 12-21 December 1949. Nigerian trends in Kafanchan, Kaduna State, Nigeria. Research Journal Museum, Lagos, pp. 203–222. of Environmental and Earth Sciences 2, 89–96. Fairhead, J. & Leach, M. 1996. Enriching the landscape: Social Arbonnier, M. 2002. Arbres, arbustes et lianes des zones sèches history and the management of the transition ecology in the d’Afrique de l’Ouest. 2nd ed. CIRAD, MNHN, Pont-sur- forest-savanna mosaic of the Republic of Ghana. Africa 66, Yonne. 14–36. http://dx.doi.org/10.2307/1161509

Avav, T.-R. & Uza, D. V. 2003. Agriculture. In: Atlas du Nigeria, Franke, G. 2014. When sherds speak – the pottery of the Nok Cul- Jeune Afrique – Jaguar, Paris, pp. 92–95. ture. In: Breunig, P. (ed.), Nok – African Sculpture in Archaeo- logical Context. Africa Magna Verlag, Frankfurt, pp. 168–177. Barendsen, G.W., Deevey, E.S. & Gralenski, L.J. 1957. Yale Natu- ral Radiocarbon Measurements III. Science 126, 908–919. Franke, G. 2015. Potsherds in Time – the Pottery of the Nigerian http://dx.doi.org/10.1126/science.126.3279.908 Nok Culture and its Chronology. PhD thesis, Goethe Univer- sity Frankfurt/Main. Beck, C. 2014. The clay’s the thing. In: Breunig, P. (ed.), Nok – African Sculpture in Archaeological Context. Africa Magna Franke, G. 2016. A chronology of the Central Nigerian Nok Verlag, Frankfurt, pp. 246–254. Culture – 1500 BC to the beginning of the Common Era. Journal of African Archaeology 14 (3), 257–289, this issue. Beck, C. 2015. The Value of Art: Studies in the Material Character of the Terracotta Figurines of the Nok Culture of Franke, G. & Breunig, P. 2014. How old is the Nok Culture. In: Central Nigeria. Unpublished PhD Thesis, Goethe University Breunig, P. (ed.), Nok – African Sculpture in Archaeological Frankfurt/Main. Context. Africa Magna Verlag, Frankfurt, pp. 130–136.

Bond, G. 1956. A preliminary account of the Pleistocene geology Frueh, W.T. & Lancaster, S.T. 2014. Correction of deposit ages for of the Plateau tin fields region of Northern Nigeria. In: Pro- inherited ages of charcoal: implications for sediment dynam- ceedings of the Third International West African Conference, ics inferred from random sampling of deposits on headwater Ibadan, Nigeria, 12-21 December 1949. Nigerian Museum, valley floors.Quaternary Science Reviews 88, 110–124. http:// Lagos, pp. 187–202. dx.doi.org/10.1016/j.quascirev.2013.10.029

Journal of African Archaeology Vol. 14 (3) Special Issue, 2016 351

Downloaded from Brill.com10/10/2021 06:57:41PM via free access A. Höhn & K. Neumann

Höhn, A. 1999. Wood anatomy of selected West African species Lézine, A.-M., Assi-Kaudjhis, C., Roche, E., Vincens, A. & of Caesalpinioideae and Mimosoideae (Leguminosae): A Achoundong, G. 2013. Towards an understanding of West comparative study. IAWA Journal 20, 115–146. http://dx.doi. African montane forest response to climate change. Journal of org/10.1163/22941932-90000672 Biogeography 40, 183–196. http://dx.doi.org/10.1111/j.1365- 2699.2012.02770.x Höhn, A. 2005. Zur eisenzeitlichen Entwicklung der Kultur- landschaft im Sahel von Burkina Faso. Untersuchungen Logan, A.L. & D’Andrea, A.C. 2012. Oil palm, arboriculture, von archäologischen Holzkohlen. PhD thesis, Goethe Uni- and changing subsistence practices during Kintampo times versity, Frankfurt. [http://publikationen.ub.uni-frankfurt.de/ (3600–3200 BP, Ghana). Quaternary International 249, volltexte/2005/2253/] 63–71. http://dx.doi.org/10.1016/j.quaint.2010.12.004

Höhn, A., Kahlheber, S., Neumann, K. & Schweizer, A. 2007. Maley, J. 2001. La destruction catastrophique des forêts d’Afrique Settling the rain forest – the environment of farming com- central survenue il y a environ 2500 ans exerce encore une munities in southern Cameroon during the first millennium influence majeure sur la répartition actuelle des formations BC. In: Runge, J. (ed.), Dynamics of Forest Ecosystems in végétales. Systematics and Geography of Plants 71, 777–796. Central Africa during the Holocene: Past – Present – Future. http://dx.doi.org/10.2307/3668717 Palaeoecology of Africa 28, 29–41. Maley, J. & Brenac, P. 1998. Vegetation dynamics, palaeoen- Hubau, W. 2013. Exploring Charcoal as a Natural Archive for vironments and climatic changes in the forests of western Palaeofire and Vegetation History of the Mayumbe, Demo- Cameroon during the last 28,000 years B.P. Review of cratic Republic of the Congo. PhD thesis, Ghent University, Palaeobotany and Palynology 99, 157–187. http://dx.doi. Ghent. org/10.1016/S0034-6667(97)00047-X

IAWA Committee 1989. IAWA List of Microscopic Features for National Research Council 1996. Lost Crops of Africa. Volume I: Hardwood Identification.IAWA Bulletin n.s. 10, 219–332. Grains. National Academy Press, Washington D.C.

InsideWood 2004-onwards, URL: http://insidewood.lib.ncsu.edu/ Natta, A.K. & Porembski, S. 2003. Ouémé and Comoé: forest search [access: 5/2013–2/2014]. savanna border relationships in two riparian ecosystems in West Africa. Botanische Jahrbücher für Systematik, Pflanz- Jansen, S., Robbrecht, E., Beeckman, H. & Smets, E. 2002. A engeschichte und Pflanzengeographie 124, 383–396. http:// survey of the systematic wood anatomy of the Rubiaceae. dx.doi.org/10.1127/0006-8152/2003/0124-0383 IAWA Journal 23, 1–67. http://dx.doi.org/10.1163/22941932- 90000288 Natta, A.K., Sinsins, B. & van der Maesen, L.J. 2002. Ri - parian forests, a unique but endangered ecosystem in Jones, E.W. 1963. The forest outliers in the Guinea zone of Benin. Botanische Jahrbücher für Systematik, Pflanzenge- northern Nigeria. Journal of Ecology 51, 415–434. http:// schichte und Pflanzengeographie 124, 55–69. http://dx.doi. dx.doi.org/10.2307/2257694 org/10.1127/0006-8152/2002/0124-0055

Joosten, H. & De Klerk, P. 2002. What’s in a name? Some thoughts Neumann, K., Kahlheber, S. & Uebel, D. 1998. Remains of on pollen classification, identification, and nomenclature woody plants from Saouga, a medieval West African village. in Quaternary palynology. Review of Palaeobotany and Vegetation History and Archaeobotany 7, 57–77. http://dx.doi. Palynology 122, 29–45. http://dx.doi.org/10.1016/S0034- org/10.1007/BF01373925 6667(02)00090-8 Neumann, K., Schoch, W., Détienne, P. & Schweingruber, F.H. Kahlheber, S., Höhn, A. & Rupp, N. 2009. Archaeobotanical 2001. Woods of the Sahara and the Sahel: an Anatomical studies at Nok sites: an interim report. Nyame Akuma 71, Atlas. Paul Haupt, Bern, Stuttgart, Wien. 2–14. Ngomanda, A., Neumann, K., Schweizer, A. & Maley, J. 2009. Kahlheber, S. 2010. Fonio at Nok sites – hungry rice or chief’s Seasonality change and the third millennium BP rainfor- food? Abstract of the presentation held in Dakar at the 13th est crisis in southern Cameroon (Central Africa). Quater- congress of the PANAF, 20th meeting of the Society of Afri- nary Research 71, 307–318. http://dx.doi.org/10.1016/j. canist Archaeologists. Congress publication, p. 151. yqres.2008.12.002

Keay, R.W.J. 1953. An Outline of Nigerian Vegetation. Govern- Nguetsop, V.F., Servant-Vildary, S. & Servant, M. 2004. Late Holo- ment Printer, Lagos. cene climatic changes in West Africa, a high resolution diatom record from equatorial Cameroon. Quaternary Science Reviews Keay, R.W.J. 1989. Trees of Nigeria. Clarendon, Oxford. 23, 591–609. http://dx.doi.org/10.1016/j.quascirev.2003.10.007

Killick, H.J. 1959. The ecological relationships of certain plants in Njokuocha, R.C. 2012. Holocene pollen deposits and recent vegeta- the forest and savanna of Central Nigeria. Journal of Ecology tion distribution of Obayi Lake, Nsukka. Quaternary Interna- 47, 115–127. http://dx.doi.org/10.2307/2257251 tional 262, 20–24. http://dx.doi.org/10.1016/j.quaint.2011.10.033

Lavachery, P. 2001. The Holocene archaeological sequence of Njokuocha, R.C. & Akaegbobi, I.M. 2014. A contribution to the Shum Laka Rock Shelter (Grass Fields, Western Cameroon). Holocene vegetation history of Nigeria: Pollen from Ohe African Archaeological Review 18, 213–247. http://dx.doi. Pond Nsukka, southeastern Nigeria. Quaternary International org/10.1023/A:1013114008855 338, 28–34. http://dx.doi.org/10.1016/j.quaint.2013.07.036

352 Journal of African Archaeology Vol. 14 (3) Special Issue, 2016

Downloaded from Brill.com10/10/2021 06:57:41PM via free access The Palaeovegetation of Janruwa

Porembski, S. 2001. Phytodiversity and structure of the Comoé Shanahan, T.M., Overpeck, J.T., Scholz, C.A., Beck, J.W., River gallery forest (NE Ivory Coast). In: Gottsberger, G. & Peck, J. & King, J.W. 2008. Abrupt changes in the water Liede, S. (eds.), Life Forms and Dynamics in Tropical Forests. balance of tropical West Africa during the late Quaternary. Dissertationes Botanicae 346, 1–10. Journal of Geophysical Research 113, D12108. http://dx.doi. org/10.1029/2007JD009320 Prior, J. & Gasson, P. 1993. Anatomical changes on charring six African hardwoods. IAWA Journal 14, 77–86. http://dx.doi. Shanahan, T.M., Overpeck, J.T., Anchukaitis, K.J., Beck, J.W., org/10.1163/22941932-90000579 Cole, J.E., Dettman, D.L., Peck, J., Scholz, C.A. & King, J.W. 2009. Atlantic forcing of persistent drought in West Africa. Sci- Reimer, P.J., Bard, E., Bayliss, A., Beck, J.W., Blackwell, P.G., ence 324, 377–380. http://dx.doi.org/10.1126/science.1166352 Bronk Ramsey, C., Buck,C.E., Cheng, H., Edwards, L.R., Friedrich, M., Grootes, P.M., Guilderson, T.P., Haflidason,H., Shanahan, T.M., McKay, N., Overpeck, J.A., Scholz, C.A., Heil, Hajdas, I., Hatté, C., Heaton, T.J., Hoffmann, D.L., Hogg, C.W. & King, J.W. 2013. Spatial and temporal variability in A.G., Hughen, K.A., Kaiser,K.F., Kromer, B., Manning, S.W., sedimentological and geochemical properties of sediments Niu, Mu., Reimer, R.W., Richards, D.A., Scott, E.M.,Southon, from an anoxic crater lake in West Africa: Implications for J.R., Staff, R.A., Turney, C.S.M. & van der Pflicht, J. 2013. paleoenvironmental reconstructions. Palaeogeography, Pal- IntCal 13 and Marine1326 Radiocarbon Age Calibration aeoclimatology, Palaeoecology 374, 96–109. http://dx.doi. Curves 0-50,000 Years Cal BP. Radiocarbon 55 (4), 1869– org/10.1016/j.palaeo.2013.01.008 1887. http://dx.doi.org/10.2458/azu_js_rc.55.16947 Vodouhè, S.R. & Achigan-Dako, G.E. 2006. Digitaria exilis (Kip- Rupp, N. 2014. Well hidden and densely spread: the sites of the pist) Stapf. In: Brink, M. & Belay, G. (eds), Plant Resources Nok Culture. In: Breunig, P. (ed.), Nok – African Sculpture of Tropical Africa 1. Cereals and Pulses. PROTA Foundation/ in Archaeological Context. Africa Magna Verlag, Frankfurt, Backhuys Publishers/CTA, Wageningen, pp. 59–63. pp. 138–149. Weldeab, S., Kölling, M. & Wefer, G. 2005. Holocene African Salzmann, U. 2000. Are modern savannas degraded forests? A droughts related nto eastern equatorial Atlantic cooling. Holocene pollen record from the Sudanian vegetation zone of Geology 33, 981–984. http://dx.doi.org/10.1130/G21874.1 NE Nigeria. Vegetation History and Archaeobotany 9, 1–15. http://dx.doi.org/10.1007/BF01295010 Weldeab, S., Lea, D.W., Schneider, R. & Andersen, N. 2007. 155,000 years of West African monsoon and ocean thermal Salzmann, U. & Hoelzmann, P. 2005. The Dahomey Gap: an evolution. Science 316,1303–1307. http://dx.doi.org/10.1126/ abrupt climatically induced forest fragmentation in West Af- science.1140461 rica during the Late Holocene. The Holocene 15 (2), 190–199. http://dx.doi.org/10.1191/0959683605hl799rp White, F. 1983. The Vegetation of Africa. A Descriptive Memoir to Accompany the Unesco/AETFAT/UNSO Vegetation Map Shanahan, T.M., Overpeck, J.T., Wheeler, C.W., Beck, J.W., of Africa. Unesco, Paris. Pigati, J.S., Talbot, M.R., Scholz, C.A., Peck, J. & King, J.W. 2006. Paleoclimatic variations in West Africa from a Willis, K.J., Bennett, K.D., Burrough, S.L., Macias-Fauria, M. record of late Pleistocene and Holocene lake level stands & Tovar, C. 2013. Determining the response of African biota of Lake Bosumtwi, Ghana. Palaeogeography, Palaeocli- to climate change: using the past to model the future. Philo- matology, Palaeoecology 242 (3–4), 287–302. http://dx.doi. sophical Transactions of the Royal Society B 368, 20120491. org/10.1016/j.palaeo.2006.06.007 http://dx.doi.org/10.1098/rstb.2012.0491

Journal of African Archaeology Vol. 14 (3) Special Issue, 2016 353

Downloaded from Brill.com10/10/2021 06:57:41PM via free access