Addressing the Wild Yam Question: How Baka Hunter-Gatherers Acted and Lived During Two Controlled Foraging Trips in the Tropical

ANTHROPOLOGICAL SCIENCE Advance Publication Vol. advpub(0), 000–000, 2012

Addressing the wild yam question: how Baka hunter-gatherers acted and lived during two controlled foraging trips in the tropical rainforest of southeastern Cameroon Hiroaki SATO1*, Kyohei KAWAMURA2, Koji HAYASHI3, Hiroyuki INAI4, Taro YAMAGUCHI5

1Faculty of Medicine, Hamamatsu University School of Medicine, Hamamatsu 431-3192, Japan 2Faculty of education Human Sciences, Yamanashi University, Kofu 400-8510, Japan 3Faculty of Humanities and Sciences, Kobe Gakuin University, Kobe 651-2180, Japan 4Graduate School of Asian and African Area Studies, Kyoto University, Kyoto 606-8501, Japan 5Graduate School of Health Sciences, Hokkaido University, Sapporo 060-0812, Japan

Received 13 September 2011; accepted 19 October 2011

Abstract We designed observational surveys of controlled foraging trips of Baka hunter-gatherers in Cameroon to verify the ‘wild yam question’—i.e. is it possible for human beings to live without agricultural products in a tropical rainforest?—and to examine their foraging lifestyle. We observed two 20-day trips during which no agricultural or commercial food except salt and pepper could be used. The first trip was conducted by six married couples in August, the short dry season, of the year 2003, and the second one by eight married couples in October, the rainy season, 2005. The Baka cooperators obtained 22 species and 43 vernacular names of food in all during both survey periods. No cooperators lost weight from any food shortage in both seasons. Energy intake per consumption-day was estimated at 2528–2865 kilocalories in the dry season, and at 2479–2777 kilocalories in the rainy season. Pro- viding more than 60% of estimated energy intake in both seasons, wild yam tubers proved to be an essential food to enable a foraging life in tropical rainforests. From this survey we could find no evidence that it is impossible to live independently of agriculture in a tropical rainforest although it seemed that the cooperators paid a high energy cost to secure food, especially wild yam tubers. This study implies that a Paleolithic foraging lifestyle in the African tropical rainforest was very likely, although not easy, and that Paleolithic foragers may have been the ancestors of the present ‘pygmy’ hunter-gatherers.

Key words: wild yam question, controlled foraging trip, Baka hunter-gatherers, African tropical rainforest, Paleolithic foraging lifestyle

Introduction such as wild yam tubers was poor. According to Vincent (1984), tropical rainforests have far less biomass of tuberous In the late 1980s, arguments arose among researchers of plants than savanna. Mammalian biomass also is reported to hunter-gatherer societies concerning the hypothesis of the be less than in drier environments (Barnes and Lahm, 1997). ‘wild yam question,’ i.e. is it possible for human beings to Another reason is the assertion of Bailey et al. that there are live without agricultural products in a tropical rainforest neither pure hunter-gatherer groups in tropical rainforests in (Headland and Bailey, 1991)? One of the main reasons for the present day, nor evidence that they lived there in the past the debate is that tropical rainforests are not rich in food re- (Bailey et al., 1989). Furthermore, according to Headland sources. Based on ecological surveys of the Ituri Forest, Hart (1997), modern-day hunter-gatherer societies have lived and Hart (1986) questioned whether hunter-gatherers en- with long histories of trade and interaction with the domi- tered the interior of rainforests as both animal and botanical nant societies around them. In contrast, Bahuchet et al. food resources were poorer than in its fringe or in the forest– (1991), Brosius (1991), Dwyer and Minnegal (1991), savanna ecotone. Headland (1987) argued that human be- Endicott and Bellwood (1991), and Stearman (1991) coun- ings could not live independently of agricultural products in tered the hypothesis based on their own fieldwork in Africa, tropical rainforests where the availability of starchy food Borneo, Papua New Guinea, the Malay Peninsula, and the Amazon in a special issue of Human Ecology on the ‘wild yam question.’ Elsewhere, from the ecological viewpoint,

* Correspondence to: Hiroaki Sato, Faculty of Medicine, Hamamatsu Colinvaux and Bush (1991) mentioned that the primary ha- University School of Medicine, 1-20-1 Handayama, Higashi-ku, Hamamatsu 431-3192, Japan. bitants in rainforests, especially the Amazon forest, were E-mail: [email protected] likely to have been hunter-gatherers. With regard to Africa, Published online 27 January 2012 giving an example of the Aka hunter-gatherers, Bahuchet et in J-STAGE (www.jstage.jst.go.jp) DOI: 10.1537/ase.110913 al. (1991) made the point that the tropical rainforests in the

© 2012 The Anthropological Society of Nippon 1 2 H. SATO ET AL. ANTHROPOLOGICAL SCIENCE west of the Republic of Central Africa were composed of di- oon and northern Republic of Congo, we made two 20-day verse vegetation and in some areas the productivity of wild trips in different seasons, during which no agricultural or yam tubers was sufficient to support a local group of hunter- commercial food except salt and pepper could be used, in the gatherers. In addition, Hladik and Dounias (1993) argued forest of southeastern Cameroon. that the forests in southern Cameroon have diverse yam In this paper we aim at testing the ‘wild yam question’ hy- plants and their productivity could possibly support a forag- pothesis, positive or negative, and at elucidating the critical ing lifestyle. Emphasizing the capability of a pure foraging factors determining whether the test is positive or negative, lifestyle in tropical rainforests, however, they did not go far examining what food resources the Baka cooperators ac- enough to entirely reject the hypothesis of the ‘wild yam quired and what cost they paid to get food, and, if negative, question.’ Since then, archaeological records, which have discussing whether the findings of this study can be applied the potential to disprove it, have been accumulated. In the to reconstruct a Stone Age foraging lifestyle in the African late 1990s the research team of Mercader excavated one site tropical rainforest. after another in the African tropical rainforests (Mercader et al., 2000, 2001; Mercader and Brooks, 2001; Mercader, Materials and Methods 2002; Mercader and Martí, 2003; Mercader, 2003). They found evidence that the sites in the Ituri Forest in the east of Survey area and cooperators the Democratic Republic of Congo were used sequentially All the Baka cooperators (Table 1) lived in a sedentary from 18800 years BP to 800 years BP and the sites in the for- small settlement near a farmers’ village, Ndongo, in south- ests of southwestern Cameroon and Equatorial Guinea from eastern Cameroon. The Ndongo villagers consist of the ma- 40000–30000 years BP to 9000 years BP. These forests did jority Bakwele farmers and the minority of Bamileke and not become entirely savanna vegetation but they left consid- Hausa merchants from northern Cameroon while the Baka, erable forested areas, i.e. they presented open forest environ- about 300 people, are the largest ethnic group around Ndon- ments, even in the Leopoldvillian cold climate period, and go Village. Significant numbers of the Baka have now set- they have presented dense forest environments since 10000 tled along a main road and keep banana fields and cacao years BP. Mercader’s work raised the possibility of a purely fields, as neighboring farmers do (Sato, 1992; Kitanishi, foraging lifestyle at least in African tropical rainforests. 2003). In addition, almost all of them get some money, local However, neither the archeological record, nor the ecologi- wine and food by helping local farmers. Nevertheless, they cal data on food resources in tropical rainforests is sufficient customarily enter the forest to catch animals, chiefly using to establish this definitively. Nor do we have clear ideas snares, and collect seed nuts such as wild mango seeds for about what enabled it or how hunter-gatherers could live in- one or two months a year, taking field crops, plantain banan- dependently of farmed agricultural products in tropical rain- as or cassava tubers. Thus the Baka people in this area main- forests. Further archeological records, ecological data of tain the knowledge and technology to live off the forest. The food resources, and behavioral data of foraging life in tropi- cooperators were no exception. cal rainforests are needed. And as Bailey and Headland The first trip was conducted in the dry season of August (1991) noted, experimental studies of the costs and benefits 2003 (Sato et al., 2006), and the second one in the rainy sea- of foraging are also necessary. son of October 2005. The Baka cooperators spontaneously Since 1995, in order to examine the potentiality of tropical selected the same place (14°45′20″ E, and 2°12′30″ N) at the rainforests as a human habitat, we have surveyed the distri- base of Mount Bek, about 600 meters above sea level, for bution and reserves of forest food resources, especially wild their campsite in both trips (Figure 1). People rarely visited yam tubers, in the tropical rainforest of southeastern Camer- the forest around the campsite as it took two days on foot oon. Our findings support Bahuchet’s view, showing that the from Ndongo Village. The vegetation around Mount Bek is forest of southeastern Cameroon could have enough wild semi-deciduous forest including small marshlands with an yams and yam-like plants to sustain a pure foraging lifestyle annual rainfall of less than 1600 mm (Letouzey, 1985). In in a tropical rainforest (Sato, 2001, 2006). Even if, however, the survey area there are four seasons: a minor rainy season, yam and yam-like tubers were plentiful, additional practical April–May, a minor dry season, June–August, a major rainy and concrete data is needed on the foraging life in a tropical season, September–November, and a major dry season, De- rainforest: how much forests can supply of the nutritional re- cember–March. As shown in Figure 2, at Moloundou, the quirements of foragers, what cost must be paid for procuring capital of this region, 50 kilometers east of Ndongo Village, them, how many people those starchy foods can feed, to the mean monthly rainfall in this area is 40 to <120 mm in what extent other food sources can contribute to foragers’ the major dry season, about 150 mm in the minor rainy sea- dietary needs, and so on. Furthermore pure foragers no long- son, around 100 mm in the minor dry season, and 130 to er exist anywhere at present (Bailey et al., 1989). In the Af- >200 mm in the major rainy season. The mean monthly tem- rican tropical rainforest we have only a few reports on perature is 25°C all year round. Mount Bek is rich in wild temporary foraging trips of hunter-gatherer groups who usu- yam tubers (Sato, 2006) and the Leke River running beside ally subsist on self-farming or farmers’ assistance other than the campsite was known as a treasure house of fish. Al- hunting and gathering (Kitanishi, 1995; Yasuoka, 2006). though this area was too far to be used for ordinary hunting Therefore we designed observational surveys of controlled or gathering trips, the Baka cooperators identified it as a foraging trips to verify the ‘wild yam question’ and to reveal bountiful forest and selected this place for the campsite. details of how they lived off the land. Gaining the coopera- The cooperators in the first trip consisted on six married tion of the Baka hunter-gatherers inhabiting southern Camer- couples, including four children under 7 years old, and those Vol. 120, 2012 OBSERVATIONS ON TWO CONTROLLED FORAGING TRIPS IN THE CAMEROON RAINFOREST 3

Table 1. Cooperators Couple no. Individual Couple no. Individual Sex Age estimated Child. Sex Age estimated Child. Aug. 2003 no. Oct. 2005 no. DF1 DF11 M 45–50 1 boy RF1 RF11 M 47–52 1 boy DF12 F 45–50 RF12 F 47–52 DF2 DF21 M 25–30 1 girl RF2 RF21 M 27–32 1 girl DF22 F 25–30 1 boy* RF22 F 27–32 1 boy* DF3 DF31 M 35–40 RF3 RF31 M 37–42 DF32 F 35–40 RF32 F 37–42 DF4 DF41 M 30–35 1 girl RF4 RF41 M 32–37 DF42 F 25–30 RF42 F 27–32 DF5 DF51 M 25–30 RF5 RF51 M 45–50 DF52 F 20–25 RF52 F 45–50 DF6 DF61 M 35–40 RF6 RF61 M 25–30 DF62 F 35–40 RF62 F 35–40 RF7 RF71 M 25–30 2 girls RF72 F 25–30 RF8 RF81 M 30–35 2 boys** RF82 F 25–30 DF1 and RF1, DF2 and RF2, DF3 and RF3, and DF4 and RF4 are the same families respectively. * an infant, ** including an infant.

Figure 1. Survey area. in the second trip eight married couples, including seven Baka language and providing for the foraging life. The fol- children under 10 years old. Since the cooperators did not lowing day the surveys started. From leaving the home set- habitually count their age, we estimated their age except for tlement to the morning of the beginning day the cooperators the children whose ages we could identify. They had a wide ate agricultural food such as plantain bananas and commer- age range from early 20s to early 50s but many of them were cial food, but from then to the morning of the final day they in the younger age group from 20s to 30s. Four among six depended on forest food resources except for salt and pep- couples in the first trip also participated in the second trip. per. During that time the research team members exclusively lived off agricultural and commercial foods. On the first trip Controlled foraging trip and observation period the observation survey began on the morning of 16 August In both trips it took three days to reach the campsite. On and ran through to the morning of 5 September 2003. On the the day of arrival the cooperators spent a lot of time in build- second trip, six families were observed for 21 days from the ing their traditional domical small huts called ‘mongulu’ in morning of 2 October to the morning of 23 October 2005, 4 H. SATO ET AL. ANTHROPOLOGICAL SCIENCE

Kitanishi (1995) and those for any others are from Standard Tables of Food Composition in Japan (Ministry of Educa- tion, Culture, Sports, Science and Technology, Japan, 2005). The value of energy and protein contents for yam tubers (item number, hereafter IN: 260 and 264), nuts (IN: 446), game meat (IN: 1134 and 1148, both are African beef as there is no data on wild games.), snake (IN: 1184), fish (IN: 1263 and 1251), snails (IN: 1423), termites (IN: 1190) and honey (IN: 1060) are derived from African food composition tables (Leung, 1968), that for leaves from Mialoundama (1993), and those for birds (IN: 11240) and crabs (IN: 10335) from Standard Tables of Food Composition in Japan (Ministry of Education, Culture, Sports, Science and Tech- nology, Japan, 2005). From above data we estimated the total energy intake of all participants in the camp and the per-capita daily energy intake (the total energy intake per consumption-day). The consumption-day was the adjusted Figure 2. Mean monthly rainfall in Moloundou, adapted from number of participants × observation days. Participants were Sigaha-Nkamdiou (1993). Although rainfall was recorded at the sta- tion from 1933 to 1991, 48 years of data was used. divided into two categories: adult persons and children (3– 10 years old), one person of which was converted to 70% of an adult person based on estimated BMR (basal metabolic while two other families were observed for 20 days from the rate) from body weight by sex and age grade (FAO, 2001). mornings of 3 to the morning of 23 October because their participation was one day late. Energy expenditure and steps walked A portable accelerometer (Kentz Lifecorder EX, Japan) Body weight was worn on the waist of every cooperator from morning We checked the cooperators’ body weight using a digital (6:00) to evening (18:30) every day to measure the energy weight scale (Tanita THD-652, Japan) every morning (6:00– expenditure for foraging activities. Since the accelerometer 6:30) before breakfast. The children’s body weight was also was not made to catch the multidimensional actions of for- checked in the morning of the starting day and ending day aging activities in the forest, however, we have not used the during the first trip and every morning during the second trip. data here. However, the accelerometer was equipped with a pedometer that counted the cooperator’s steps walked. We Time spent in foraging activities and activity observation use the pedometer data as a method to measure the cost ex- Everyday we timed every cooperator going out and re- pended for foraging activities. turning to the camp. Coming back, all the cooperators were asked what they did in the forest. From these data we esti- Outline of the controlled foraging trip mated the time spent in foraging activities by each coopera- In both trips the cooperators used their usual tools such as tor. In order to know their activity pattern, the time machetes, spears, iron fittings fixed on the point of wooden allocation, the division of labor by sex, and so on, the direct digging sticks, which were instantly prepared at the time of activities of every cooperator were observed from morning digging yam tubers, wire-snares, axes, hooks and lines, fish- (6:00) to evening (18:00) of one day, to which the observa- ing nets, carrying baskets, etc., for foraging activities. tion day was randomly allotted in both trips. In this survey During both survey periods the cooperators engaged in geographical information was also collected using a GPS re- collecting wild yam tubers, wild honey, termites, edible fun- ceiver (Garmin eTrex, USA). gi, nuts, etc.; trapping with wire-snares; and fishing. While honey collecting, trapping, net fishing, and hook and line Weighing food and estimation of dietary intake fishing were done exclusively by men, and dam and bail All food brought back to the camp was identified and fishing by women, collecting yam tubers, termites, nuts, and weighed using a digital hanging scale (Bonso model 393, edible fungi were done by both men and women (Table 2). HongKong), a digital kitchen scale (Tanita KD-174, Japan) The common activity pattern was that each couple went and a spring scale. All the discarded rotten food during out to collect wild yam tubers every two or three days and on survey periods and the remnant food at the end of survey the other days they engaged in collecting nuts, honey, ter- were also weighed. Animals were identified with the aid of mites, etc., or in fishing in the dry season. Most of the wire- Kingdon’s book (1997). As for the plants, we first recorded snares were set around the campsite and men made the their vernacular names and then, referring to Letouzey rounds of them every day before breakfast or on their way (1976), Hamon et al. (1995), and Dumont et al. (1994), we back from other foraging activities. cited their scientific names. Occasionally weighing the food Usually adult men had meals together, which their wives before and after cooking and, in the case of game, the left- cooked, at a meeting place called ‘mbanjo’ in the center of over bones after eating, we estimated the rate of edible por- the campsite both in the morning and evening, whereas wom- tions for yam tubers, nuts, leaves, games’ meat, snails, and en and their children had meals at the front of their own huts. termites. Those for snakes, fish, and honey are derived from Normal meals in the morning were only boiled yam tubers; Vol. 120, 2012 OBSERVATIONS ON TWO CONTROLLED FORAGING TRIPS IN THE CAMEROON RAINFOREST 5

Table 2. Foraging activities observed during the period of both surveys Worker Activity Target Tool used Brief discription male female Gathering yam ‘ngoso,’* machete Finding yam vines, they dig out yam tubers using an ‘ngoso’ or a machete. nut** machete Collecting fallen seeds of Panda oleosa, they break them with a machete and remove kernels. honey ax, machete Finding wild honeycomb on a tree, men cut it down with an ax and take the honey. termite machete Finding a termite hill, they dig out termites using a machete or a digging stick. fungus Fungi are normally picked during walking for other food getting activities. Hunting game spear Whenever men walk in a forest, they carry a spear and use it on game encountered. game wire snare All the snares set during research period were spring ones using wire for the loops. Fishing fish hook, line, rod Targets of fishing using a small hook and a rod are small fishes in streams. fish net Only a gill net was used during both research periods because it was expensive. big fish machete Finding the entrance hole of fishes on the side of a stream bank, they dig into it and catch fish. fish et al. pan, plant leaf*** Damming a stream with trees and clay, women bail water out of the downstream puddle and catch fish, crabs or shrimps. * an iron fitting fixed on the point of a digging stick. ** Almost all nuts were collected from Panda oleosa while very few irvingia nuts were found. *** Maranta leaves, especially those of Megaphrynium macrostachyum, are often used in place of a metal pan for bailing. the evening meal consisted of yam tubers and soup of meat, weight in the last half (Table 3). Among them the weight fish, or fungi with nut paste. As husbands shared the food at loss of one woman (DF52) in the dry season and one man the ‘mbanjo’ and wives often interchanged plates with each (RF81) in the rainy season appeared to be due to a feminine other, it was unlikely that any individuals lacked food. ailment and a digestive upset, respectively.

Statistical analysis Food brought back to the camp All statistical analyses were done using statistical package Food brought back to the camp during both survey peri- JMP (version 8.0.2). ods consisted of various food types: wild yam tubers (5 species), mammals (12 species), freshwater fish (18 vernacular Research ethic names), termites (one vernacular name), honey (one vernac- Before the two main trips, we made a preliminary 10-day ular name), nuts (2 species), and edible fungi (16 vernacular trip with six married couples in the forest near their settle- names). There were 22 species and 43 vernacular names of ment in August 2001 to determine what conditions were food in all (Table 4). Among these, the first ranking in needed to conduct each research trip safely. Based on this weight was wild yam tubers, and then mammals followed by preliminary trip, we informed the cooperators on the details fish in the dry season and by nuts in the rainy season of our research and the necessary conditions: staying in the (Table 5). In both seasons yam tubers and mammals occu- forest for about three weeks, subsisting only on forest food pied more than 90% in weight composition although there resources during the trip, being observed daily as to body were a few seasonal differences; the amount of yam tubers in weight, subsistence activities, meals, and so on, and being the rainy season was larger than in the dry season, whereas able to withdraw from the research whenever some problem the catches of mammals in the dry season were larger than in occurred. We asked each of the cooperators to voluntarily the rainy season. Catches of fish also were larger in the dry participate in this study and gained their consent. season. Although the lowered water level enabled large catches of fish in the dry season, the causes of large catches Results of mammals in the dry season were not identified but might have been attributable to the contingency in hunting. The Body weight large amount of yam tubers collected in the rainy season will On both trips the body weight of almost all the coopera- be discussed later. Among the five species of wild yam tu- tors varied little day to day. Comparing it between the first bers collected Dioscorea praehensilis (‘safa’), an annual half and the last half of the survey period, most of them kept plant, supplied the bulk of their food in both seasons. A large their weight consistently, or modestly increased their weight number of the mammals trapped were small-and-medium- in the last half, with the exception of a few women in the dry sized forestry duikers weighing 5–15 kg. While a small num- season and one man in the rainy season who slightly lost ber of large-sized catfish were found in their big catches, a 6 H. SATO ET AL. ANTHROPOLOGICAL SCIENCE

Table 3. Comparison of cooperators’ weight between the first half and latter half of survey period

Indiv. Indiv. Sex Mean ± SD MF ML BMR t-test Trend ± t no. no. Sex Mean SD MF ML BMR -test Trend Aug. 2003 DF11 M 36.0 ± 0.26 36.0 ± 0.32 36.1 ± 0.19 1286.1 ns DF12 F 36.4 ± 0.38 36.2 ± 0.45 36.6 ± 0.13 1141.4 * + DF21 M 51.6 ± 0.69 51.8 ± 0.82 51.4 ± 0.38 1469.1 ns DF22 F 47.7 ± 0.83 48.2 ± 0.72 47.1 ± 0.47 1193.4 *** − DF31 M 47.4 ± 0.35 47.4 ± 0.27 47.4 ± 0.44 1416.9 ns DF32 F 45.0 ± 0.59 45.1 ± 0.66 44.9 ± 0.55 1211.3 ns DF41 M 51.7 ± 0.54 51.5 ± 0.58 52.0 ± 0.31 1466.2 * + DF42 F 39.7 ± 0.29 39.8 ± 0.19 39.5 ± 0.30 1074.9 * − DF51 M 48.2 ± 0.51 48.2 ± 0.32 48.0 ± 0.60 1426.1 ns DF52 F 40.5 ± 1.10 41.4 ± 0.28 39.5 ± 0.58 1086.7 *** − DF61 M 47.2 ± 0.47 47.1 ± 0.28 47.3 ± 0.50 1414.6 ns DF62 F 35.9 ± 0.31 36.1 ± 0.30 35.8 ± 0.26 1137.3 ns Oct. ’05 1140.8 RF11 M 34.1 ± 0.54 34.1 ± 0.72 34.0 ± 0.22 1264.3 ns RF12 F 36.5 ± 0.36 36.4 ± 0.45 36.6 ± 0.16 1142.2 ns RF21 M 52.7 ± 0.86 52.3 ± 0.92 53.2 ± 0.41 1500.8 ** + RF22 F 47.8 ± 0.30 47.7 ± 0.32 47.9 ± 0.27 1194.9 ns RF31 M 46.3 ± 0.48 46.0 ± 0.42 46.6 ± 0.39 1404.3 ** + RF32 F 43.5 ± 0.43 43.3 ± 0.45 43.6 ± 0.39 1199.1 ns RF41 M 51.6 ± 1.04 51.1 ± 1.02 52.3 ± 0.45 1476.5 ** + RF42 F 39.2 ± 1.22 39.1 ± 1.69 39.4 ± 0.30 1164.1 ns RF51 M 43.6 ± 0.67 43.1 ± 0.35 44.1 ± 0.52 1373.3 *** + RF52 F 45.5 ± 0.64 45.2 ± 0.52 45.9 ± 0.62 1215.3 * + RF61 M 44.6 ± 0.43 44.3 ± 0.42 44.8 ± 0.31 1363.7 * + RF62 F 39.9 ± 0.79 39.4 ± 0.58 40.4 ± 0.67 1169.8 ** + RF71 M 59.3 ± 0.81 58.8 ± 0.54 59.9 ± 0.62 1585.1 *** + RF72 F 48.4 ± 0.32 48.3 ± 0.33 48.5 ± 0.29 1203.8 ns RF81 M 48.7 ± 1.84 50.0 ± 1.49 47.2 ± 0.65 1431.8 * − RF82 F 50.3 ± 0.46 50.4 ± 0.57 50.3 ± 0.29 1231.9 ns MF: mean weight and standard deviation during the first half of survey period. ML: mean weight and standard deviation during the latter half of survey period. BMR: basal metabolic rate estimated with the equations of FAO (2001: 37). Trend: Plus indicates weight’s increasing and minus weight’s decreasing from the first half to the latter half of survey period. * P < 0.05, ** P < 0.01, *** P < 0.001. great number of small fish constantly provisioned the diet of amount of acquired food among couples can be linked to the the camp, especially in the dry season. Almost all the nuts inequality of food intake, however, as not only game meat were Panda oleosa. Honey brought back to the camp repre- but also other foods such as boiled yam tubers were usually sented only a portion of the total collected as the cooperators shared by all cooperators at meals. There was no statistically usually consumed part of the honey at the collecting place. significant difference in the weight of game caught by couples between the first half and the last half of the survey pe- Stability and fluctuation of food procurement riod (Table 10); however, the weight of nuts in the dry The cooperators usually collected wild yam tubers in units season in the last half was larger than in the first half of married couples. Each couple left the camp with the inten- (Table 11). These things also suggest that there was no evi- tion to collect wild yam tubers on an average of 10 days at an dence of hunting and gathering pressure on game and nuts. interval of around a day during both surveys (Table 6). Henceforth we express the day when a couple collected wild Energy and protein intake yam tubers as the yam collecting day (YCD). The mean We did not measure individual food intake. Instead, we weight of yam tubers collected by each couple on YCD was derived the total food intake of cooperators and their chil- 9–12 kg per day in the dry season and 10–16 kg per day in dren from all food acquired less any leftovers, the bulk of the rainy season (Table 7). Although the weight of yam tu- which was preserved dry meat taken back to their settlement, bers collected by each couple slightly varied from day to and the amount of food discarded during each survey period. day, it did not correlate with the order of YCD with the ex- The cooperators usually left the yam tubers brought back be- ception of a couple that showed a negative correlation in the hind their huts, covered with leaves to prevent drying. Ac- rainy season: the weight of yam tubers decreased by the day cording the cooperators, yam tubers start going off two or because of the husband’s sickness. Neither in the dry season three days after being collected. On the second trip, about nor in the rainy season was there a statistically significant 5% of yam tubers collected were discarded because of rot. difference in the weight of yam tubers collected by couple Since we did not weigh the discarded tubers in the first trip, between the first half and the last half of the survey period in the estimation of food intake we assumed yam tubers (Table 8). This suggests that there was no evidence of gath- would have been discarded at the same rate as the second ering pressure on yam tubers, causing resource shortage, at trip. Some yam tubers were brought back to the home settle- least during the 20-day survey period. Although there was a ment and some used for packed lunches on the way back slight seasonal difference and daily fluctuations, the cooper- home from the camp after the surveys finished. Thus, ingest- ators consistently secured a considerable amount of yam tu- ed yam tubers were estimated at 1.4 kg per consumption-day bers in both seasons. Other than yam tubers, there was great in fresh weight in the dry season and 1.5 kg per consumption- variability among couples in the weight of food brought day in the rainy season. Nuts, game, and fish were dried and back in both seasons (Table 9). It is especially noteworthy preserved on their fires except for food that was cooked on that the quantity of captured game, which was of importance the day of foraging. A portion of these dried foods was also in terms of food supply, ranged widely from couple to cou- brought back to their home. Above all, a large amount of ple in both seasons. It is unlikely that the difference in the game meat was brought back since it was of particular value Vol. 120, 2012 OBSERVATIONS ON TWO CONTROLLED FORAGING TRIPS IN THE CAMEROON RAINFOREST 7

Table 4. The weight of food brought into the camp during each research period Aug. 2003 Oct. 2005 Food Total fresh n Total fresh n Scientific name or English name Baka name weight (kg) weight (kg) Wild yam Dioscorea praehensilis Benth. safa 636.6 1030.6 D. semperflorens Uline suma 10.9 27.0 D. mangenotiana Miége ba 7.7 22.4 D. burkilliana Miége keke 4.8 12.3 D. minutiflora Engl. kuku 0.7 0.6 Mammal Cephalophus callipygus Peters ngendi 85.4 6 45.1 3 C. dorsalis Gray ngbomu 74.3 4 132.4 7 C. leucogaster Gray mongala 27.0 2 C. monticola (Thünberg) ndengbe 5.5 1 26.8 6 C. silvicultor (Afzelius) bemba 19.9 1 Neotragus batesi De Winton samba 2.5 1 Hyemoschus aquaticus (Ogilby) akolo 96.6 9 45.2 5 Atherurus africanus (Gray) boke 3.2 1 Potamochoerus porcus (L.) famme 45.5 1 Bdeogale nigripes Pucheran buse 3.9 1 7.1 2 Crossarchus obscurus F. Cuvier ganda 4.9 2 3.0 Felis aurata Temminck ebbie 5.4 1 Reptile Bitis gabonica gabonica buma 3.6 1 Bird ? gbedi 0.1 1 ? elúá 0.1 1 eggs of guinea hen fafala-kanga 0.3 Fish electric catfish gbibi 12.8 8.4 catfish kannya 18.1 3.4 catfish ngolo 0.2 carp likamboka 4.1 0.8 characin popo 3.1 characin denge 0.4 0.0 characin mayanga 0.1 characin toshi 0.1 a characin jilelo 1.1 cichlid sale 0.3 cichlid toko 0.7 b characin jaseli 4.9 characin monjangbe 0.1 elephant nose fish mbose 1.3 African snakehead fish monga 1.4 ? mbongo 0.5 c ? mekpaso d ? misango 0.3 Crustacea crab kala 0.3 0.1 shrimp kanji 0.0 Mollusc African snail mbembe 0.8 1.7 Insect termite bandi 5.9 9.3 Honey stingless bee dandu 8.0 3.1 Plant seeds Panda oleosa Pierre kanna 22.9 44.3 Irvingia excelsa Mildbr. fayo 0.1 Plant leaves Gnetum spp. koko 0.0 8 H. SATO ET AL. ANTHROPOLOGICAL SCIENCE

Table 4. (continued) Aug. 2003 Oct. 2005 Food Total fresh n Total fresh n Scientific name or English name Baka name weight (kg) weight (kg) Fruit Anonidium manni (Oliv.) ngbe 12.0 Gambeya lacourtiana (De Wild.) bambu 1.6 Fungi ? mondongola 0.1 ? tulu-bela 0.0 ? tulu-sakusa 1.4 ? mawoluolu 0.0 ? tulu-bongo 2.0 ? tulu-kanga 1.3 ? tulu-baala 1.0 ? moselele 2.0 0.3 ? yofiyo 0.0 ? tulu-sakili 0.7 ? deddele 1.5 ? tulu-dengbe 0.0 ? tulu-timi 0.0 ? tulu-gbado 0.1 ? moleseko 0.4 ? asamoni 1.0 a + b + c + d = 0.338 kg

Table 5. Food brought into the camp Aug. 2003 Oct. 2005 Food type Total fresh weight Weight/CD Total fresh weight Weight/CD % % (kg) (kg) (kg) (kg) Yam tubers 660.7 59.4 2.75 1092.9 74.7 3.29 Seeds 23.0 2.1 0.10 44.3 3.0 0.13 Leaves 0.0 0.0 0.00 Fruit 13.6 1.2 0.06 Fungi 3.0 0.3 0.01 8.7 0.6 0.03 Mammal 343.1 30.9 1.43 290.6 19.9 0.88 Bird 0.2 0.0 0.00 Bird’s egg 0.3 0.0 0.00 Snake 3.6 0.3 0.02 Fish 49.5 4.5 0.21 12.9 0.9 0.04 Crustaceae 0.3 0.0 0.00 0.1 0.0 0.00 Snail 0.8 0.1 0.00 1.7 0.1 0.01 Termite 5.9 0.5 0.02 9.3 0.6 0.03 Honey 8.0 0.7 0.03 3.1 0.2 0.01 1111.8 100.0 1463.8 100.0 Weight/CD: Weight per cooperator per day. in the settlement where the cooperators could not have it similar in both seasons: yam tubers supplied more than 60% very often and could also sell it. Therefore, ingested meat of the total energy intake, game meat 15–20%, and nuts was estimated at 0.4 kg per consumption-day in fresh weight around 10% and the sum of these three food types more than in the dry season and 0.3 kg per consumption-day in the 90%. The daily total protein intake was estimated at 114.5– rainy season. 146.1 g per consumption-day in the dry season and 93.0– We estimated the daily total energy intake at 2528– 125.4 g per consumption-day in the rainy season. Although 2865 kcal per consumption-day in the dry season, and at around 60% of total protein intake was from animal protein 2479–2777 kcal per consumption-day in the rainy season in both seasons, its contribution in the dry season was larger (Table 12). Daily total energy intake per consumption-day in than in the rainy season due to larger catch of game and fish. the dry season was slightly larger than that in the rainy season. The contribution of food type to total energy intake was Vol. 120, 2012 OBSERVATIONS ON TWO CONTROLLED FORAGING TRIPS IN THE CAMEROON RAINFOREST 9

Table 6. Interval of yam collecting Aug. 2003 TYCD Interval Oct. 2005 TYCD Interval

Couple 0123Couple 0123 N N no. %%%%no. %%%% DF1 12 36.4 63.6 9.1 0.0 RF1 11 40.0 30.0 20.0 10.0 DF2 10 22.2 55.6 22.2 0.0 RF2 9 12.5 50.0 25.0 12.5 DF3 10 22.2 55.6 22.2 0.0 RF3 8 0.0 42.9 42.9 14.3 DF4 11 30.0 60.0 10.0 0.0 RF4 10 22.2 44.4 33.3 0.0 DF5 9 12.5 50.0 37.5 0.0 RF5 10 11.1 66.7 22.2 0.0 DF6 12 36.4 63.6 0.0 0.0 RF6 9 25.0 50.0 25.0 12.5 RF7 10 33.3 33.3 22.2 11.1 RF8 11 30.0 50.0 20.0 0.0 Mean 10.7 26.6 58.1 16.8 0.0 9.8 21.8 45.9 26.3 7.5 TYCD: Total of the yam collecting day when a couple aimed at collecting wild yam tubers, excluding days when they collected by chance wild yam tubers on their way of other foraging activities. Interval: interval of days between each contiguous yam collecting day.

Table 7. Daily fluctuation of the weight of yam tubers collected on the yam collecting day by couple Couple Order of the yam collecting day No. Aug. 2003 I II III IV V VI VII VIII IX X XI XII S kg M ± SD S/T % PC Ss DF1 10.2 11.2 6.3 12.3 6.4 12.3 11.5 8.6 4.6 10.5 4.6 7.9 106.4 8.9 ± 2.9 93.6 −0.387 ns DF2 9.5 7.1 8.5 15.9 8.2 9.2 13.2 8.3 11.5 6.9 98.3 9.8 ± 2.9 99.6 −0.005 ns DF3 7.3 5.7 11.9 8.6 13.1 9.3 11.5 8.7 2.5 11.3 89.9 9.0 ± 3.2 100.0 0.017 ns DF4 9.3 10.3 13.3 11.5 13.1 10.4 19.2 16.2 9.0 8.8 10.8 131.9 12.0 ± 3.3 99.0 0.041 ns DF5 15.8 6.1 10.5 15.4 5.4 11.0 7.9 9.2 17.0 98.3 10.9 ± 4.3 97.7 0.063 ns DF6 9.1 15.5 15.9 7.5 8.3 13.9 9.5 3.5 15.4 11.5 3.0 11.5 124.6 10.4 ± 4.4 100.0 −0.275 ns Oct. 2005 RF1 23.5 21.6 11.6 15.0 19.5 5.9 6.8 13.2 15.7 6.1 33.3 172.2 15.7 ± 8.4 99.7 −0.061 ns RF2 12.7 26.1 13.2 22.8 9.9 16.1 11.8 12.4 19.1 144.1 16.0 ± 5.5 98.5 −0.248 ns RF3 10.1 7.7 16.6 6.2 12.8 8.2 10.9 18.0 90.5 11.3 ± 4.2 95.9 0.467 ns RF4 15.7 7.9 25.0 6.3 11.3 16.5 11.2 10.1 13.4 11.7 129.1 12.9 ± 5.3 97.8 −0.198 ns RF5 19.2 20.4 13.2 14.5 4.2 11.3 15.3 18.7 19.2 23.7 159.7 16.0 ± 5.6 99.3 0.248 ns RF6 11.3 18.0 7.3 13.8 5.5 9.7 4.6 10.8 9.2 90.2 10.0 ± 4.2 98.3 −0.504 ns RF7 19.9 32.5 14.1 5.3 30.3 7.6 6.1 22.8 18.0 3.1 159.7 16.0 ± 10.5 92.5 −0.432 ns RF8 21.7 16.6 11.1 13.9 16.5 9.1 6.1 5.5 15.2 9.4 8.3 133.4 12.1 ± 5.1 100.0 −0.66 * S: Sum of the weight of yam tubers collected on the yam collecting days. S/T: Rate of S to the amount of yam tubers collected during research period. PC: Pearson’s product–moment correlation coefficient between the order of the yam collecting day and the weight of yam tubers collected on the yam collecting day. Ss: Statistical significance (ns, not significant), *P < 0.05.

Table 8. Weight variation of wild yam tubers collected between Intensity of foraging activities the first half and the last half of survey period We timed the periods the cooperators spent outside the Aug. 2003 First Last Oct. 2005 First* Last campsite, except for urination and carrying water and fuel for their fires. This time (working time) could be regarded as Couple no. kg/day kg/day Couple no. kg/day kg/day the time spent in the cooperators’ food-getting activities. We DF1 6.0 5.4 RF1 8.3 8.1 also recorded with pedometers the number of steps walked DF2 4.9 4.9 RF2 7.7 6.2 by cooperators everyday. Most cooperators presented a DF3 4.7 4.3 RF3 3.8 5.3 strong positive correlation between their daily working time DF4 5.3 4.7 RF4 7.5 4.9 DF5 6.8 6.5 RF5 6.6 8.8 and number of steps walked except for a few persons in both DF6 7.0 5.4 RF6 5.1 4.1 seasons (Table 13). The mean working time and number of RF7 7.3 8.9 steps walked per person per day had a seasonal difference; RF8 8.1 4.5 those in the dry season were larger than in the rainy season Wilcoxon’s test ns ns (Figure 3). There was a large daily variance in the working time and number of steps walked by individuals. One of the First: The weight per day of yam tubers collected in the first 10 days. Last: The weight per day of yam tubers collected in the last 10 days. reasons was the high intensity of activity on YCD * The first half of survey period includes 11 days except for RF6 (Table 14). The daily working time of couple on YCD, 800– and RF7. 1000 minutes, was 1.2–1.5 times longer than that on days 10 H. SATO ET AL. ANTHROPOLOGICAL SCIENCE

Table 9. Mean weight food brought into the camp by couple Family Couple no. Yam Game Fish Nuts Termite Fungi members Aug. 2003 kg kg kg g g g DF1 5.7 2.4 1.1 78.3 94.6 0.0 3 DF2 4.9 1.3 0.7 188.6 133.4 0.0 4 DF3 4.5 1.3 0.2 288.6 0.0 85.0 2 DF4 5.0 5.0 0.3 189.1 0.0 50.0 3 DF5 6.7 5.0 0.1 174.8 41.5 17.7 2 DF6 6.2 2.4 0.1 231.9 0.0 0.0 2 Mean 5.5 2.9 0.4 191.9 44.9 25.5 SD 0.8 1.7 0.4 69.5 57.2 35.1 Oct. 2005 RF1 8.2 0.0 0.0* 97.4 82.3 102.1 3 RF2 7.0 3.6 0.6 325.5 269.6 60.6 4 RF3 4.5 0.9 0.0 351.5 24.7 34.4 2 RF4 6.3 2.4 0.0 278.3 11.2 33.8 2 RF5 7.7 3.6 0.0 234.6 0.0 13.1 2 RF6 4.6 0.2 0.0* 167.1 28.4 37.9 2 RF7 8.1 1.9 0.0* 374.5 27.5 105.3 4 RF8 6.4 1.2 0.0 310.3 0.0 36.1 4 Mean 6.6 1.7 0.1 267.4 55.5 52.9 SD 1.4 1.4 0.3 95.6 90.3 33.9 * Extremely small amount.

Table 10. Weight variation of games caught between the first half Table 11. Weight variation of wild nuts collected between the first and the last half of survey period half and the last half of survey period Aug. 2003 First Last Oct. 2005 First* Last Aug. 2003 First Last Oct. 2005 First* Last Couple no. kg/day kg/day Couple no. kg/day kg/day Couple no. g/day g/day Couple no. g/day g/day DF1 1.2 3.2 RF1 0.0 0.0 DF1 63.2 93.4 RF1 109.4 84.2 DF2 1.1 1.5 RF2 1.4 6.1 DF2 122.9 254.3 RF2 295.3 358.8 DF3 2.0 0.6 RF3 0.0 2.0 DF3 247.5 329.6 RF3 381.8 318.2 DF4 1.8 8.2 RF4 2.0 2.8 DF4 134.1 215.5 RF4 293.7 261.3 DF5 3.1 7.0 RF5 1.4 6.0 DF5 118.7 259.5 RF5 145.1 333.1 DF6 1.9 2.8 RF6 0.0 0.4 DF6 167.3 296.5 RF6 162.4 171.7 RF7 3.3 0.6 RF7 332.0 416.9 RF8 2.4 0.0 RF8 284.5 338.7 Wilcoxon’s test ns ns Wilcoxon’s test P = 0.0313 ns First: The weight per day of games caught in the first 10 days. First: The weight per day of yam tubers collected in the first 10 days. Last: The weight per day of games caught in the last 10 days. Last: The weight per day of yam tubers collected in the last 10 days. *The first half of research period includes 11 days except for RF6 * The first half of research period includes 11 days except for RF6 and RF7. and RF7. other than YCD and the daily steps walked by couples on other than YCD. However, neither the working time and YCD, 20000–30000, were 1.4–3.0 times more than those on steps walked of couples on YCD, nor the weight of yam tu- days other than YCD in both seasons. We recorded many bers collected per unit of working time and per unit of steps cases where the steps walked by cooperators greatly de- walked by couples on YCD had a correlation with the order creased on the day after they took long walks. This is be- of YCD (Table 15). This suggests that during both survey cause the yam collecting area was concentrated around periods there was no gathering pressure to cause the cooper- Mount Bek, about 2.5 km from the camp. As shown in ators to increase the intensity of yam collecting activity, or Figure 4, any sites of foraging activities other than yam col- to decrease its efficiency by the day. Nor was there a differ- lecting were a shorter distance from the camp, although ence in the mean daily working time and number of steps there was a difference that nut collecting sites were widely walked by couples between the first half and the last half of dispersed whereas most snaring sites were within a radius of research period (Table 16). 1 km from the camp. On YCD, cooperators had to walk 3– 4 km along mountain roads back to the camp carrying heavy tubers of around 10 kg on their backs. Thus the cooperators would intentionally engage in lighter activities on the days Vol. 120, 2012 OBSERVATIONS ON TWO CONTROLLED FORAGING TRIPS IN THE CAMEROON RAINFOREST 11 5.4 1 0.7 0.9 0.9 0.5 0.7 0.6 0.3 0.4 + 0.68 0.85 Edible 8 20.6 × 1.5 11.8 15.3 22.5 14.4 13.9 17.7 20.4 Protein 0.7) 7 × 32 94 63 498 105 140 107 356 Energy 7 from Standard Tables of Food Tables from Standard shi (1995) and that for the others the for that and (1995) shi e last day was not weighed. Protein per 100 g per 100 g ratio couples and their five children 9.1 22.9–48.9 24.7–39.0 112–124 1.5–3.2 8), and those for the others are 005). because the breakfast on th nd honey are derived from Kitani 5 6 13.7–16.0 52.3–59.2 47.2–56.3 109–143 16.8–19.0 En kcal % g % kcal g Edible ratio /Consumption-day × 296 in Aug. 2003, and 15.5 (six = 4 0.6 consumption-day, 0.6 consumption-day, 20 R 2 × No 3 0.7) authors, that for snake, fish a Protein value × × 9 0.0 15.5–23.3 0.6–0.8 2.9–3.3 2.4–3.5 85–128 16.2–18.1 1 0.2 0.0 0.2–0.3 0.0 0.1 0.0–0.1 R)) + (D e, Sports, Science and Technology, Japan, 2 e, Sports, Science and Technology, − .0 1463.8 2479.4–2776.6 100.0 93.0–125.4 100.0 survey is calculated as 1.7 1092.9 50.6 113.6 1712.8–1896.3 68.3–6 (W ail, termite and honey are derived from Leung (196 = their four children + Table 12. Dietary energy and protein intake Protein W ce and Technology, Japan, 2005). ce and Technology, d rate recorded in the 2005 survey. snail and termite were measured by 425.1 in Oct. 2005. = 5 .4 16.9–19.6 65.8–74.5 51.0–57.5 290.6 95.8 339.7–445. En research days: 14.8 (six couples 19.6 on in Japan (Ministry of Education, Cultur × × Edible ratio/Consumption-day, Protein Edible ratio/Consumption-day, of the extremely small amount. kcal % g % kg D × Aug. 2003 (16 persons) Oct. 2005 (23 persons) 0.7) × rs, seeds, leaves, game’s meat, snake, fish, sn rs, seeds, leaves, game’s r a person in the evening on beginning day of Education, Culture, Sports, Scien intake per consumption-day. was not eaten in the camp. bers, seeds, leaves, game’s meat, bers, seeds, leaves, game’s 4 R tubers was estimated at the same discar 18.9 1618.9–1792.3 62.6–64.0 21.7–46.3 18.9–3 2 6 Energy value Energy No 3 Number of subjects × = R)) + 1 kg D W (D − 9 (W = Total fresh weight of food brought into the camp. Total Fresh weight of food which Food discarded because of rot. Remaining food at the end of survey. and protein Estimated energy The weight of discarded values for yam tube The energy The edible ratio for yam tu Gnetum leaves are excluded because A child was converted to 70% of an adult. survey one meal fo In the 2005’s Consumption-day En 1 2 3 4 5 6 7 8 9 Food type Yam tubersYam 660.7 30.6 (two couples and their two children NutsFungiGame meat 343.1Bird 23.0 eggBird’s 3.0Snake 172.5 0.3Fish 4.8 427.2–560 Crustaceae 0.0 3.6Snail 0.3 0.0Termite 49.5 275.6Honey 9.6–10.9 0.0Total 0.8 2.9 5.9 0.0 15.1 8.0 1.2 1111.8 8.5 0.1 59.3–89.3 0.0 0.0 5.8–7.4 8.0 2.3–3.1 0.0 0.2 0.0 11.3–12.6 44.3 0.1 7.7–11.0 0.3 2528.4–2865.1 71.0 0.0 1.2 0.1 12. 14.4 0.1 63.0 2.5–2.8 100.0 114.5–146.1 1.2 0.0 2.2–2.5 0.1 0.0 8.7 0.8–1.1 100 4.1 0.0–0.2 315.2 11.4–12.7 0.0 0.1 2.8–3.6 0.0–0.2 1.0 0.1 9.3 1.7 9.7 0.1 7.7–10.4 3.1 5.2 0.0 0.7 0.0 0.2 0.0 72.0 0.0 1.7 0.2 2.6–2.9 17.0 0.2–0.3 0.0 0.1 0.6–0.7 4.1 0.0 3.3–4.4 0.3 0.0 0.2–0.3 0.0 0.0 311 0.4 0.75 Composition in Japan (Ministry of of Food Compositi Tables are from Standard 12 H. SATO ET AL. ANTHROPOLOGICAL SCIENCE

Table 13. Work time and steps walked of cooperators

Individual Work time Steps walked 1 Individual Work time Steps walked Correlation 3 Correlation no. Sex per day (min.) per day (n) Ss no. Sex per day (min.) per day (n) Ss coefficient coefficient Aug. 2003 Mean ± SD Mean ± SDOct. 2005 Mean ± SD Mean ± SD DF11 M 489 ± 88 9577 ± 3000 0.2349 ns RF11 M 382 ± 142 7489 ± 4055 0.8392 *** DF122 F484± 87 26766 ± 6781 0.6958 *** RF12 F 378 ± 137 6915 ± 3967 0.8249 *** DF21 M 460 ± 82 10893 ± 4427 0.7740 *** RF21 M 409 ± 97 11654 ± 4423 0.6714 *** DF22 F 436 ± 78 9865 ± 6201 0.7478 *** RF22 F 373 ± 107 9175 ± 4741 0.6706 *** DF31 M 437 ± 77 9060 ± 4506 0.7288 *** RF31 M 356 ± 105 8101 ± 4412 0.5093 * DF32 F 435 ± 71 8032 ± 4906 0.7172 *** RF32 F 336 ± 98 5828 ± 3696 0.4328 ns DF41 M 459 ± 76 13457 ± 4080 0.5907 ** RF41 M 389 ± 108 9561 ± 4300 0.7411 *** DF42 F 436 ± 74 9998 ± 4901 0.5116 * RF42 F 331 ± 96 13490 ± 6483 0.6257 ** DF51 M 462 ± 69 15380 ± 4150 0.7374 *** RF51 M 437 ± 134 12770 ± 5092 0.7655 *** DF52 F 434 ± 76 9812 ± 4828 0.7624 *** RF52 F 389 ± 150 7147 ± 4510 0.7606 *** DF61 M 450 ± 93 12638 ± 4150 0.8044 *** RF61 M 381 ± 105 12862 ± 4564 0.7292 *** DF62 F 436 ± 93 13073 ± 6308 0.3841 ns RF62 F 310 ± 131 6727 ± 4234 0.7916 *** RF71 M 384 ± 116 8758 ± 3876 0.8377 *** RF72 F 342 ± 125 12364 ± 6184 0.8449 *** RF81 M 250 ± 166 5885 ± 5520 0.9131 *** RF82 F 309 ± 138 9276 ± 5219 0.9320 *** 1 Correlation coefficient between daily work time and number of steps walked during research period. 2 Because of leg pain this female cooperator took more steps than normal. 3 Statistical significance: *P < 0.05, **P < 0.01, ***P < 0.001.

forest food resources for about two months, from February to April 2002. Among diverse food items acquired, wild yam tubers were the most important as an energy source during the trip. In northeastern Republic of Congo, six groups of Aka hunter-gatherers, consisting of 15–75 persons, engaged in hunting trips for 6–27 days, during which they were almost independent of agricultural products, which supplied only 11% of their total energy intake (Kitanishi, 1995). Here also wild yam tubers were one of their major energy sources. In addition, the Late Pleistocene sites excavated by Mercader and Martí (2003) were slightly distant from the region of these surveys. These findings suggest a high likelihood of a hunting-gathering lifestyle at least in the forest at the northwest margin of the Congo Basin. However, some questions still remain to be solved.

Carrying capacity of the African tropical rainforest First we will discuss how many people can live off forest Figure 3. Mean daily work time and number of steps walked by food resources over many years, i.e. the carrying capacity of cooperators. The reason why one woman recorded the most steps was this forest. The tubers of yams and yam-like plants are the that she could not walk normally because of leg pain. most important foods for a foraging life in the forest and a key food involved in the ‘wild yam question.’ Referring to the little available data on yams and yam-like tubers in the Discussion forest of the Lobaye Region, Central African Republic (Bahuchet et al., 1991) and in the southern Cameroon forest 1. Answers to the ‘Wild Yam Question’ (Hladik and Dounias, 1993), we will examine this question The possibility of a hunting-gathering lifestyle in the African based on the research of standing stem density and tuber tropical rainforest productivity of yams and yam-like plants conducted in the Although the surveys suffered from certain restrictions, forest from Ndongo Village to Moloundou town using the such as a small population, short term, only two seasons, and line-transect method (Sato, 2001). As shown in Table 17, the a narrow forest area, there was no evidence against the pos- tuber biomass of yams and yam-like plants in six sites was sibility of a hunting-gathering lifestyle in a tropical rainfor- estimated at 5.3–17.0 kg per hectare. The vegetation at three est. Recently, a long-term foraging trip (‘molongo’ in Baka sites (total length of line-transects: 4 m × 6 km) showing language) of a different Baka group in a forest close to this higher values (15.3–17.0 kg/ha) was in the secondary forests survey area was reported (Yasuoka, 2006). According to including fallows, whereas that of the other three sites (total Yasuoka, about 100 Baka people stayed there and lived off length of line-transects: 4 m × 7 km), the productivity of Vol. 120, 2012 OBSERVATIONS ON TWO CONTROLLED FORAGING TRIPS IN THE CAMEROON RAINFOREST 13

Table 14. Difference in daily work time and steps walked between on the yam collecting day (YCD) and on days other than YCD by couple Couple no. NYCD TYCD SYCD Couple no. NYCD TYCD SYCD St St St St Aug. 2003 NO-YCD TOYCD min. SOYCD Oct. 2005 NO-YCD TOYCD min. SOYCD DF1 12 1016 ns 41161 *** RF1 11 951 *** 19820 *** 8 907 29116 10 554 8446 DF2 10 1004 *** 30689 *** RF2 9 913 ** 29888 *** 10 787 10827 12 683 14034 DF3 10 961 ** 25664 *** RF3 8 800 * 22895 *** 10 782 8519 13 626 8411 DF4 11 984 *** 29922 *** RF4 10 849 ** 30273 *** 9 788 17230 11 603 16486 DF5 9 960 ns 30680 *** RF5 10 982 ** 27088 *** 11 841 17543 11 684 13399 DF6 12 993 *** 30695 *** RF6 9 848 ** 26782 *** 8 724 18234 11 561 13704 RF7 10 890 *** 28996 *** 10 562 13248 RF8 11 704 ** 21970 *** 10 400 7672 NYCD: the number of YCD; NOYCD: the number of days other than YCD. TYCD: Mean combined daily work time of husband and wife on YCD. TOYCD: Mean combined daily work time of husband and wife on days other than YCD. SYCD: Mean combined daily steps walked by husband and wife on YCD. SOYCD: Mean combined daily steps walked by husband and wife on days other than YCD. St: Student’s t-test: *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 4. Plot of foraging activity sites recorded by GPS. This figure was made based on the daily activities of 12 cooperators (7 males and 5 females) recorded by GPS. A researcher recorded the geographical position of each new site by GPS every time a cooperator changed the site or the foraging activity. The number of sites plotted was 17 for yam collecting by 5 males and 2 females, 14 for panda nuts collecting by 3 males and 4 females, 5 for wire-snaring by 2 males, and 5 for other activities by 1 male and 3 females. which was 5.3–8.7 kg/ha, had few anthropogenic distur- consumption-day, respectively. Adopting the value in the bances. Here we will adopt the value of 5.0 kg/ha as the rainy season, we can calculate the annual requirements of tuber biomass of yams and yam-like plants in this region yam tubers at about 800 kg per person. We consider the uti- since that in a forest site, almost all of which had a micro- lization ratio of yam tubers biomass as one-fifth, referring to vegetation type, ‘manja,’ was 5.3 kg/ha. ‘Manja’ is a forest the data of Bahuchet et al. (1991) from the Central African with a closed canopy and occupied a large part of our survey Republic. Thus we can estimate the forest area that will area. Next we will calculate what weight of tubers of yams reserve 4000 kg of annual tuber requirements per person at 8 and yam-like plants an adult person requires per year based square kilometers. Carrying capacity in the survey area was on our survey. The fresh weight of yam tubers intake was 0.125 person/km2, which is similar to the 0.14 person/km2 estimated at 611.2 kg in August 2003 and 928.7 kg in in Yasuoka’s ‘molongo’ area (Yasuoka, 2006) and 0.10 October 2005, which were converted to 2.06 and 2.18 kg per person/km2 in the Lobaye Region (Bahuchet et al., 1991). It 14 H. SATO ET AL. ANTHROPOLOGICAL SCIENCE

Table 15. Kendall’s rank correlation coefficient between the order of the yam collecting day and the weight of yam tubers collected, work time, and steps walked on the yam collecting day (YCD) by couple Total of Couple no. Order vs. Yam weight Order vs. Work time Order vs. Steps walked Order vs. Y/WT Order vs. Y/SW YCD Aug. 2003TSsTSsTSsTSsTSs DF1 12 −0.2462 ns 0.2727 ns 0.0606 ns −0.3939 ns −0.3333 ns DF2 10 −0.0667 ns 0.2 ns −0.0222 ns 0.0667 ns 0.0222 ns DF3 10 0.0667 ns −0.1556 ns −0.1556 ns 0.1556 ns 0.0667 ns DF4 11 0.0556 ns 0 ns 0.3889 ns 0.0556 ns 0.0556 ns DF5 9 −0.0182 ns 0.0545 ns −0.0182 ns −0.0909 ns −0.0909 ns DF6 12 −0.1679 ns −0.1818 ns 0.1212 ns −0.1515 ns −0.3333 ns Oct. 2005TssTssTssTssTss RF1 11 −0.1636 ns −0.0734 ns −0.3818 ns −0.2 ns −0.0909 ns RF2 9 −0.1111 ns −0.3889 ns −0.1111 ns 0.1111 ns −0.0556 ns RF3 8 0.2857 ns −0.5455 ns −0.3571 ns −0.5 ns −0.2857 ns RF4 10 −0.0222 ns −0.1556 ns 0.1111 ns −0.0667 ns −0.2 ns RF5 10 0.2247 ns −0.2889 ns −0.2889 ns 0.4667 ns 0.4222 ns RF6 9 −0.3333 ns −0.1111 ns −0.3333 ns −0.2222 ns −0.2222 ns RF7 10 −0.3333 ns −0.2444 ns −0.1556 ns −0.2889 ns −0.3778 ns RF8 11 −0.5273 * −0.5273 * −0.4182 ns −0.0545 ns −0.0545 ns YCD: yam collecting days. Y/WT: The weight of yam tubers collected per 100 minutes of work time on YCD. Y/SW: The weight of yam tubers collected per 1000 steps walked on YCD. T: Kendall’s rank correlation coefficient. Ss: Statistical significance: *P < 0.05. is noteworthy that this value could indeed be modest. We re- ods such as line-transects, we could not obtain useful data for ferred to the value of tuber biomass, 5.3–8.7 kg/ha in three it immediately. However, the facts that the clump at Mount sites that had few anthropogenic disturbances in the calcula- Bek had a huge biomass of tubers and actually supplied a tion of carrying capacity. More than 70% of tuber biomass in large amount of tubers for consumption, and there were at these three sites was from perennial yam plants such as least several clumps around the survey area, suggest that tak- Dioscorea minutiflora, D. smilacifolia, and D. burkiliana, ing into account the tuber productivity of D. praehensilis, whereas the tubers of annual D. praehensilis hardly con- we can expect a much larger carrying capacity from this for- tribute to the biomass. As was observed in all six sites, est than that mentioned above. In addition, although no tu- these three species of perennial yam plants must be dis- bers of Dioscoreophyllum cumminsii and only a few tubers tributed evenly over any forest. This means that these three of D. semperflorens were used in this survey, those tubers perennial yam plants have the potential to sustain a carry- also could contribute to a foraging life in some areas as these ing capacity comparable to 0.08 person/km2. As shown in yams and yam-like plants, especially D. semperflorens, our survey and Yasuoka’s ‘molongo,’ however, the tubers of played an important role as energy sources in both annual D. praehensilis were indeed necessary for a pure Yasuoka’s ‘molongo’ (Yasuoka, 2006) and hunting trips of foraging lifestyle in this region. We should know how the Aka hunter-gatherers (Kitanishi, 1995). much productivity this plant has, but it is difficult to deter- As mentioned above, it is unlikely that the hypothesis of mine the tuber biomass because the plants clump and the the ‘wild yam question’ applies at least to this region, the clumps are scattered over the broad forest. As the yam col- southeastern part of Cameroon. We have insufficient data on lecting sites were concentrated in Mount Bek, there was a the species of Dioscoreaceae to extend our conclusion to any large clump of D. praehensilis on the slope and top of the other African tropical rainforest, whereas we can expand on mountain, the biomass of which had been estimated at their distribution. Although the annual species of 118 kg/ha by Sato (2006). According to the cooperators, Dioscoreaceae are commonly distributed in dry forest and there were three clumps of D. praehensilis around the camp woody savanna, D. praehensilis and D. semperflorens are other than at Mount Bek and they mentioned their names, particularly found in tropical rainforests (Hamon et al., Mboto, Mokinda, and Mokondo, all of which were the 1995). However, these two species cannot be seen in the two names of hills. Yasuoka (2006) also reported several clumps botanical reports of importance in the Ituri Forest (Hart and of D. praehensilis which were used during the long-term Hart, 1986; Tanno, 1981), presumably because its vegeta- expedition. Although Hart and Hart (1986) mentioned that tion type favors evergreen rainforests, where less productiv- the Mbuti in the Ituri Forest did not have clumped food such ity of D. praehensilis can be expected (Hladik et al., 1984). as tuberous plants in the Hadza and the mongongo nut in the As Headland (1987) pointed out, the ‘wild yam question’ Kalahari San, the Baka in the northwestern forest of Congo might be correct in evergreen tropical rainforests. According Basin could have many clumps of D. praehensilis. As it is to White (1983: 79), however, the type of drier peripheral difficult to determine clearly the tuber productivity of this semi-evergreen rainforest, a synonym for semi-deciduous plant, whose clumps are scattered about, with ordinal meth- forest, which covers the entire survey area, occurs in the Vol. 120, 2012 OBSERVATIONS ON TWO CONTROLLED FORAGING TRIPS IN THE CAMEROON RAINFOREST 15

Table 16. Comparison of work time and steps walked between the first half and the last half of research period Work time per day (min.) Steps walked per day Individual Sex Entire First half Last halfSt Entire First half Last half St no. Mean ± SD Mean ± SD Mean ± SD Mean ± SD Mean ± SD Mean ± SD Aug. 2003 DF11 M 489 ± 88 461 ± 112 517 ± 39 ns 9577 ± 3000 10710 ± 3718 8443 ± 1517 ns DF21 M 460 ± 82 438 ± 98 481 ± 56 ns 10893 ± 4427 10250 ± 5060 11536 ± 3852 ns DF31 M 437 ± 77 429 ± 84 444 ± 70 ns 9060 ± 4506 8695 ± 5114 9424 ± 4048 ns DF41 M 459 ± 76 429 ± 80 488 ± 60 ns 13457 ± 4080 13084 ± 4656 13830 ± 3625 ns DF51 M 462 ± 69 456 ± 85 468 ± 49 ns 15380 ± 4150 15089 ± 4952 15671 ± 3409 ns DF61 M 450 ± 93 437 ± 110 463 ± 73 ns 12638 ± 4150 13164 ± 4957 12112 ± 3821 ns DF121 F484± 87 451 ± 108 516 ± 39 ns 26766 ± 6781 24846 ± 8108 28687 ± 4809 ns DF22 F 436 ± 78 418 ± 88 454 ± 65 ns 9865 ± 6201 9841 ± 6720 9889 ± 5999 ns DF32 F 435 ± 71 432 ± 69 438 ± 74 ns 8032 ± 4906 7773 ± 5525 8292 ± 4486 ns DF42 F 436 ± 74 422 ± 84 450 ± 61 ns 9998 ± 4901 10025 ± 5291 9971 ± 4763 ns DF52 F 434 ± 76 423 ± 87 445 ± 63 ns 9812 ± 4828 8878 ± 5008 10559 ± 4807 ns DF62 F 436 ± 93 427 ± 106 445 ± 80 ns 13073 ± 6308 10192 ± 4017 15954 ± 7028 * Oct. 2005 RF11 M 382 ± 142 353 ± 180 413 ± 83 ns 7489 ± 4055 7382 ± 5158 7606 ± 2635 ns RF21 M 409 ± 97 415 ± 121 402 ± 66 ns 11654 ± 4423 11894 ± 4835 11390 ± 4166 ns RF31 M 356 ± 105 382 ± 120 327 ± 81 ns 8101 ± 4412 7977 ± 5197 8237 ± 3634 ns RF41 M 389 ± 108 370 ± 131 410 ± 78 ns 9561 ± 4300 9306 ± 4782 9842 ± 3940 ns RF51 M 437 ± 134 436 ± 166 438 ± 94 ns 12770 ± 5092 12509 ± 5433 13057 ± 4964 ns RF61 M 381 ± 105 362 ± 134 399 ± 67 ns 12862 ± 4564 12093 ± 5352 13631 ± 3744 ns RF71 M 384 ± 116 364 ± 121 404 ± 114 ns 8758 ± 3876 8608 ± 4510 8908 ± 3366 ns RF81 M 250 ± 166 314 ± 162 180 ± 147 ns 5885 ± 5520 7937 ± 6263 3628 ± 3651 ns RF12 F 378 ± 137 358 ± 170 404 ± 90 ns 6915 ± 3967 6653 ± 4849 7203 ± 2946 ns RF22 F 373 ± 107 373 ± 125 372 ± 89 ns 9175 ± 4741 8820 ± 5045 9565 ± 4620 ns RF32 F 336 ± 98 362 ± 117 307 ± 66 ns 5828 ± 3696 5824 ± 4133 5833 ± 3373 ns RF42 F 331 ± 96 315 ± 122 348 ± 58 ns 13490 ± 6483 10471 ± 6123 16811 ± 5325 * RF52 F 389 ± 150 385 ± 194 394 ± 91 ns 7147 ± 4510 7471 ± 5055 6791 ± 4067 ns RF62 F 310 ± 131 324 ± 129 295 ± 138 ns 6727 ± 4234 6768 ± 4817 6686 ± 3825 ns RF72 F 342 ± 125 317 ± 146 367 ± 102 ns 12364 ± 6184 10343 ± 6049 14385 ± 5922 ns RF82 F 309 ± 138 328 ± 163 289 ± 109 ns 9276 ± 5219 10105 ± 6032 8365 ± 4283 ns St: Student’s t-test: *P < 0.05. 1 Because of leg pain this female cooperator took more steps than normal.

Table 17. Estimated biomass of wild edible tubers by species by site (after Sato, 2001) Site S1 S2 S3 S4 S5 S6 Species g/ha g/ha g/ha g/ha g/ha g/ha Dioscoreophyllum cumminsii 2793 6584 6052 1197 1696 150 Dioscorea praehensilis 1110 278 D. mangenotiana 12 12 6 12 3 D. minutiflora 370 398 974 500 412 19 D. burkilliana 13575 8824 6788 6788 3054 7466 D. smilacifolia 206 213 353 248 107 131 “njakaka” ? Total kg/ha 17.0 16.0 15.3 8.7 5.3* 8.0 * The biomass of ‘njakaka’ is excluded because of the extremely small amount. form of two bands running transversely across Africa, to the and its high productivity of yam tubers. As Bailey and north and south of the moister forests. Large tuber produc- Headland (1991) drew attention to the anthropogenic distur- tivity of D. praehensilis and D. semperflorens to support a bance in tropical rainforests, e.g. the clearing of forests for foraging lifestyle could be expected within the broad bands farming, which was advantageous for light-demanding of this vegetation type. Botanical studies of D. praehensilis plants such as Dioscoreaceae, we also did not believe that and D. semperflorens in this vegetation type must therefore the forest of Mount Bek has been left untouched. In fact, we be undertaken. found pottery fragments around Mount Bek. However, as we Finally, we will comment on the vegetation of Mount Bek cannot identify who the users were, farmers or hunter- 16 H. SATO ET AL. ANTHROPOLOGICAL SCIENCE gatherers, when, and how long they lived there, neither we two hours on foot from their settlement. The forest around can determine the extent to which the inhabitants altered the the campsite, a semi-deciduous forest, was used daily for vegetation of Mount Bek, nor to whether the altered vegeta- snaring or for short hunting trips by the Baka people. Con- tion contributed to increasing yam tuber productivity. There cerned that this area was under high hunting-gathering pres- were several abandoned farmers villages, which were on the sure, and with the restricted use of wire snares by the old map issued in 1910, around many yam patches which cooperators to catch prey during the foraging trip, we decid- were used by the Baka people in Yasuoka’s ‘molongo’ ed to shorten the survey period. Here we will focus on the (Yasuoka, 2009). Yasuoka said that the activities of forest collecting activity of wild yam tubers, publishing the details people, including the Baka, might have influenced the distri- of this survey elsewhere. As shown in Table 18, 576 kg of bution of yam patches or the productivity of yam plants, es- fresh tubers of yams and yam-like plants were collected dur- pecially annual yams, but not directly; we agree with ing the 14-day survey period. The tubers of D. praehensilis Yasuoka’s view. Archeological and historical surveys accounted for more than 90% of collected tubers, as in our around Mount Bek need to be done. Here we will mention previous surveys. We also examined the weight of collected another possible factor linked to the high productivity of fresh tubers per consumption-day to compare the tuber pro- yam tubers. It is that the survey area was a hilly area. Much ductivity of yams and yam-like plants between four seasons; more sunlight can reach the surface of the top and slopes of the minor dry, major rainy, and minor rainy season in our hills than on the surface of flat forest. Such areas are suited survey and the major dry season in Yasuoka’s survey. Cal- to the genus Dioscorea of light-demanding plants, especially culating the consumption-day, we adopted the conversion annual species. All the places of D. praehensilis clumps that factors in Yasuoka’s study, i.e. boy and girl 12 or over, and the Baka cooperators named were hills. Extensive geomor- adult = 1, boy and girl between 2 and 12 = 0.5, and infant phological surveys can bring in new findings for the tuber under 2 = 0. The value of D. praehensilis in the minor rainy biomass and the distribution of yam plants. season (1.97) is slightly less than that in the minor dry season (2.27) and much less than that in the major rainy season The seasonality of D. praehensilis (2.66), but considerably larger than that in the major dry sea- The next question is the seasonality of D. praehensilis, son (1.36). These facts are different from the accepted which renews both stem and tuber annually (Hladik and knowledge on the seasonality of D. praehensilis. According Dounias, 1993). According to the cooperators, the amount of to Mckey et al. (1998), its biological cycle consists of a stem tubers they can collect will be slightly smaller in April–May, growth phase in April–May, a development phase in June– the minor rainy season, because some tubers of D. praehen- August, and a resting phase in September–March. Focusing silis sprout. Yasuoka (2011) described that the ‘molongo’ on the tuber development, Dounias described its biological finished in the middle of April because fibrous tubers or tu- cycle as follows: D. praehensilis, which renews its aerial bers with a bad taste became more common due to germina- stem and tuber reserve every year, uses its tuber as an energy tion. The minor rainy season, April–May, can be said to be source for the sprouting and growth of new stems from April the severest season for harvesting tubers of D. praehensilis. to July, stores tuber reserves from August to November, and In April 2010, we conducted a similar two-week survey with then keeps its tuber reserves at maximum for the next sprout- the Baka cooperators, eight couples and their four children. ing after March (Dounias, 2001). He also pointed out that The previous survey area could not be used because of new November–March is the best period for harvesting tubers. zoning laws for that forest area in Cameroon. Therefore the Generally agreeing, Yasuoka revised it to the period from cooperators established the campsite in a forest area about November to April from his observations of ‘molongo’

Table 18. Comparison of harvesting tubers of wild yams and yam-like plants Our studies Yasuoka’s study (2006) Study season Aug.–Sep. 2003 October 2005 April 2010 Feb.–Apr. 2002 (CD = 280) (CD = 388) (CD = 266) (CD = 3325) Tubers of yams and Weight Weight Weight Weight W/CD W/CD W/CD W/CD yam-like plants kg % kg % kg % kg % Dioscorea praehensilis 636.6 96.4 2.27 1030.6 94.3 2.66 525.1 91.1 1.97 4519.8 83.0 1.36 D. semperflorens 10.9 1.6 0.04 27.0 2.5 0.07 2.3 0.4 0.01 654.6 12.0 0.20 D. mangenotiana 7.7 1.2 0.03 22.4 2.0 0.06 31.0 5.4 0.12 61.8 1.1 0.02 D. burkilliana 4.8 0.7 0.02 12.3 1.1 0.03 16.5 2.9 0.06 184.7 3.4 0.06 D. minutiflora 0.7 0.1 0.00 0.6 0.1 0.00 0.8 0.1 0.00 22.7 0.4 0.01 Dioscoreophyllum cumminsii 0.7 0.1 0.00 3.1 0.1 0.00 Total 660.7 100.0 2.36 1092.9 100.0 2.82 576.4 100.0 2.16 5446.7 100.0 1.65 Weight: The fresh weight of tubers of wild yams and yam-like plants brought into the camp. CD: consumption-day. W/CD: Weight per consumption-day. The conversion factors of consumption-day in Yasuoka’s study were adopted to compare our studies with Yasuoka’s study. Conversion factors in Yasuoka’s study: boy and girl 12 or over and adult = 1; boy and girl between 2 and 12 = 0.5; infant under 2 = 0. Vol. 120, 2012 OBSERVATIONS ON TWO CONTROLLED FORAGING TRIPS IN THE CAMEROON RAINFOREST 17

(Yasuoka, 2006). In short, Dounias and Yasuoka say that yet, but many clumps of this plant are present in this region. April and August are not good seasons for harvesting the As a clump could supply a large amount of tubers to about tubers of D. praehensilis due to expending tuber reserves 20 persons for 3 weeks in this survey area, and several for renewing stems and the early phase of storing new tu- clumps about 100 persons for more than 1 month in the ber reserves, respectively. Our three surveys seems to sup- northern part of this region (Yasuoka, 2006), these clumps port their view: the weight of collected fresh tubers per could have sustained a Paleolithic foraging life. In addition, consumption-day in October (the major rainy season) was perennial wild yam plants also should have contributed to it the largest, followed by that in August (the minor dry season) because they had considerable biomass and were ubiquitous- and that in April (the minor rainy season) was the least. ly distributed in any forest (Sato, 2001); in fact, they sup- However the lowest value in April (1.97) was much larger plied a fair amount of tubers during the nomadic period of than that in the major dry season from February to April in ‘molongo’ (Yasuoka, 2006). Yasuoka’s survey (1.36). Although we were unable to di- The results of this survey and Yasuoka’s ‘molongo’ indi- rectly compare these values, as the circumstances of the sur- cated that various kinds of food other than wild yam tubers vey were not identical, the difference is too large to ignore. also could be important key foods that supported a Paleolith- How could the much larger amount of D. praehensilis ic foraging life. Among others, wild game, nuts, fish, honey, tubers collected in April and August than in the major dry and insects must have played complementary roles as energy season, December–March, or that a sufficient amount of D. and protein sources. Below we will mention the significance praehensilis tubers to enable a foraging life for 2–3 weeks of each food type and their issues with foraging activity. was collected in April and August, be accounted for? Al- (i) Game though the length of survey periods and the number of coop- Although the hunting gear used in this survey comprised erators is problematic, here we will lay out a likely wire snares and iron spears, all game were caught with wire hypothesis. As shown in Figure 2, the graph of rainfall in snares and almost all of them were forest duikers. Since this region has bimodal peaks, which we can also see in the Paleolithic hunters might have used only stone spears, it is rainfall graph at Moloundou in Dounias’s paper (Dounias, unlikely that the amount of catch per person or per unit of 2001: 138). In fact, this region, including the survey area, time spent could exceed that of this survey where efficient has two distinct rainy seasons, April–May and September– wire snaring was conducted. According to Yasuoka (2011), November, and two dry seasons, June–August and Decem- the Baka iron spear hunters caught red river hogs or dwarf ber–March. Thus, it is possible that the beginning of the ma- forest buffalos in six of ten cases of collective hunting forays jor rainy season in September is also another starting point including more than four hunters. Although Yasuoka noted for the biological cycle of D. praehensilis, although Dounias that spear hunting was not suited for forest duikers, if a hunt- considered the beginning of the minor rainy season, April– ing method such as the driving method which Mbuti archers May, as its starting point (Dounias, 2001). If it is correct, the (Harako, 1976) or net hunters (Tanno, 1976; Ichikawa, best period for harvesting the tubers of this type could be 1983) in the Ituri Forest adopted, i.e. driving the duikers hid- April–August and it is likely that the sympatric presence of den in the bush out into the open and shooting them or into two types of D. praehensilis, which have different biological nets, was used, Paleolithic spear hunters may well have cycles, account for the relatively large amount of collected hunted forest duikers. The problem is that Paleolithic hunt- tubers in August or for the lack of seasonal disadvantage in ers who were devoted to hunting would have had no time for the minor rainy season. Endicott and Bellwood (1991) men- collecting wild yam tubers. Game meat must have played an tioned the reason why wild yam tubers in the Malay Penin- important role for Paleolithic people, not only in a nutrition- sula could be obtained at any time of year was that the dry al aspect but also in an emotional aspect as pleasure food. season was so slight that it did not coordinate the tubering of (ii) Nuts different yam plants. Experimental research on how two dry Almost all nuts collected in this survey were P. oleosa. seasons a year can result in tubering of yam plants or annual Since this fruit has a hard-shelled seed, it takes eight months tubers productivity is needed. to three years for this fruit to germinate after falling (Vivien and Faure, 1996). Therefore many black seeds, the size of ta- 2. The implications of this survey to reconstruct the Paleo- ble tennis balls, usually can be seen on the surface around lithic foraging lifestyle in the African tropical rainforest this tree. Furthermore because this tree is common in the Essential food for the Paleolithic foraging life in the African survey area, it is not difficult to collect seeds anytime and tropical rainforest anywhere. Beside P. oleosa, several species of the genus Irv- Considering the high contribution of wild yam tubers to ingia are also common in the survey area. Although few irv- the food supply in our survey and Yasuoka’s ‘molongo,’ ingia seeds were collected in this survey, it is quite usual for they were the most dependable food for a purely foraging the Baka people to enter the forest to gather these seed nuts life in this region. Although the cooperators obtained vari- in the minor dry season, from June to August. Supplying a ous food items during the survey period, we could not find fair amount of dietary energy and enhancing the flavor of other reliable foods that had the potential to supply over 50% food as a source of dietary oil, nuts such as panda nuts and of dietary energy requirements. Among wild yam tubers, the irvingia nuts must have been of substantial importance for supply of the annual plant D. praehensilis was particularly Paleolithic hunter-gatherers. A particular problem is that Pa- notable. D. praehensilis is likely to have been an essential leolithic gatherers would have spent a lot of time and effort food for the Paleolithic foraging life in the African tropical shelling nuts after collecting seeds, and did so without the rainforest. We cannot know its productivity for certain as use of iron tools. In this survey couples usually collected 18 H. SATO ET AL. ANTHROPOLOGICAL SCIENCE seeds and shelled nuts in pairs using machetes. Judging from The Paleolithic foraging lifestyle in the African tropical our observation, there would have been few differences in rainforest was not easy working efficiency between iron tools and stone tools. Rath- The cooperators went out to collect yam tubers every 2 er, it needs to be considered whether Paleolithic women days in this survey. Such an activity pattern must have re- gatherers engaged by themselves in nut-collecting activity. sulted from the facts that (1) yam tubers spoil in a couple of Having been considered light work, this must have placed a days after being dug out, and (2) the yam collecting places considerable burden on Paleolithic woman gatherers. were not so close that they could visit them every day. (iii) Fish Whereas the gathering activities in hunting-gathering socie- In this survey a small but indispensable amount of fish ties heretofore have been described as women’s work, in this was procured in the dry season. Fish are indeed a valuable survey husband and wife commonly engaged together in animal food source in this area, where there are two dry sea- yam collecting activity. It must have been because wire- sons. The present Baka cooperators knew three fishing snaring, men’s work, was very economical in terms of work. methods—net fishing, hook and line fishing, and dam and Setting 10–20 wire-snares in a couple of days, men made the bail fishing—whereas Paleolithic hunter-gatherers would round of snares every morning and/or on the way back from have conducted only dam and bail fishing. Nevertheless, this other activities for half an hour at most. Wire-snaring is not last method is a reliable and effective means of catching fish. only economical but effective. In this survey wire-snares Engaging in it for one hour in a neighboring stream, even a were set in the forest where people rarely hunted and were so solitary woman can obtain a sufficient catch for her family. strong that they could provide a large amount of catch: 0.7– Furthermore this fishing method does not need iron tools. In 1.2 kg per consumption-day. However, Paleolithic hunters this survey, normal dam and bail fishing was a woman’s without wire-snares and other iron hunting gear could not task, whereas men with spears joined in a special type of have afforded the time to join in the yam collecting activity dam and bail fishing targeting big catfish hiding in holes on but were compelled to dedicate themselves solely to hunting the bank. Considering the ease of fishing activities and the activity. If so, a woman must have had to seek, dig out, and stability of fish procurement, fish must have been a contrib- carry back yam tubers of about 10 kg for her family every 2 utor to the typical Paleolithic foraging life beyond the num- days by herself. This would have been fairly hard work for bers in this survey. women. To solve this problem, Paleolithic hunter-gatherers (iv) Honey and insects must have built their camp as close to the clump of annual Although significant amounts of honey were not collected yam plants as possible. However, they could not always in this survey, it is well known that honey is a most attractive have done so because they had to secure other kinds of food food for all pygmy groups (Ichikawa, 1981; Bahuchet, 1985; and water, all of which might not have been close to the yam Kitanishi, 1995; Yasuoka, 2006). The Baka people, both clump. The reason why in this survey the cooperators did not men and women, pay special attention to seeking out honey- build their camp close to the yam clump was certainly that comb whenever they enter the forest. As Ichikawa (1981) re- they could not have done so. Since there was no watering ported, the Mbuti depend on honey for the bulk of their food place along the mountain roads between the camp and the in the rainy season. Honey must also have been a most valu- yam clump, they set their camp on the bank of the Leke Riv- able food for Paleolithic hunter-gatherers. The present Baka er where they could easily get both water and fish. Further- men usually cut down a tree with an iron ax to get honey, more, digging tubers would have been harder for Paleolithic whereas Paleolithic hunters would have had no choice but to gatherers since they did not have iron points for their wood- climb to get the honey. Climbing would have been much en digging sticks. The Baka people can make and use two riskier than cutting, no matter which was the more effective. types of wooden digging stick, ‘ngbafa’ and ‘bodunga.’ ‘Ng- Although there were various edible insects in this area, bafa’ is a simple digging stick, the end of which is sharply only termites were collected in this survey. Finding plump cut, and ‘bodunga’ is a special wooden stick, the end of termites gave the cooperators great pleasure. Many kinds of which is cut into four parts which are chamfered and fixed edible insects, such as beetles, lepidopteran larvae, and ter- with vines like a cone, to dig out tubers of D. praehensilis mites, which can be gathered with simple tools or even with- and D. semperflorens deep in the ground (see Dounias, out tools, must have been considered precious food as 1993: 628). The present Baka people rarely use these wood- animal protein sources for Paleolithic hunter-gatherers in en sticks now because they have iron points. Wooden sticks some seasons or in some areas. are suited for soft soil but not for soil and gravel, whereas a Beside these foods, other diverse food items were ob- wooden stick with an iron point attached can be used for tained during both survey periods. The longer the survey pe- both soil types. Furthermore, making a ‘bodunga’ is a time- riod, the more the variety of food items increases. Although consuming task, but the attachment of an iron point to the Hart and Hart (1986) and Bailey and Headland (1991) point- end of a wooden stick is fast and easy. According to Hurtado ed out the high cost in accessing food resources in small, and Hill (1989), Machiguenga women using wooden tools widely dispersed patches as one of challenges facing forag- spent two to three times as many minutes in digging and ers in tropical rainforests, we will cite the presence of large, peeling manioc tubers as women using machetes and knives. densely distributed patches of annual wild yam plants and An important consideration of a foraging lifestyle in trop- the tremendous resource diversity regardless of the sparse ical rainforests is the large energy expenditure required. We distribution of each food resource as factors that enhance the estimated the daily energy intake at 2528–2865 kcal per per- possibility of a Paleolithic hunting-gathering lifestyle in this son in the dry season and 2479–2777 kcal per person in the area. rainy season. Since there was no large fluctuation in the Vol. 120, 2012 OBSERVATIONS ON TWO CONTROLLED FORAGING TRIPS IN THE CAMEROON RAINFOREST 19 body weight of cooperators during both survey periods, the and low weight would have resulted from the adaptation for daily energy intake, i.e. total energy intake (TEI), could be heat dissipation in the hot and humid tropical rainforest considered as the daily energy expenditure, i.e. total energy (Cavalli-Sforza, 1986). That is, the adaptation to a hot and expenditure (TEE). We can estimate the mean BMR from humid forest environment, where convection, radiation, and the body weight of the cooperators at 1413 kcal/day in males evaporation cannot be expected to effectively dissipate body and 1141 kcal/day in females in the dry season and at heat, would have been accomplished by ‘miniaturizing’ the 1425 kcal/day in males and 1190 kcal/day in females in the body, short stature and low weight, resulting in decreasing rainy season (FAO, 2001). If the physical activity level heat production. Both are consistent in that the pygmy’s (PAL: TEE/BMR) were the same for both sexes, we can cal- physical features are genetic and the consequences of adap- culate approximate PAL value at 1.98–2.24 in the dry season tation to tropical rainforests. Although we basically agree, and 1.90–2.12 in the rainy season substituting total energy we still question Cavalli-Sforza’s argument. Should tropical intake for total energy expenditure. According to FAO rainforests be regarded as so hot and humid? We know that (2001), a lifestyle with a PAL value of 1.70–1.99 is classi- the maximum temperature reaches most 30°C during the day fied as an active or moderately active lifestyle, and that of in the tropical rainforest of the Congo Basin and before 2.00–2.40 as a vigorous or vigorously active lifestyle. dawn the temperature drops to 20°C; those who lived in the Leonard and Robertson (1992), who studied the bioenerget- Congo Basin may have felt rather chilly in the early morn- ics of primates from the evolutionary viewpoint, described ing. In general the humidity reaches 80–90% on chilly morn- the PAL values of two hunter-gatherer groups as 1.71 for ings whereas it drops to less than 60% as the temperature males and 1.51 for females of the !Kung and 2.15 for males rises during the afternoon, and it is rather cooler on rainy and 1.88 for females of the Ache. The PAL value of the Baka days. Could such climatic conditions have been a selective is higher than that of the !Kung and seems to be comparable factor for Paleolithic hunter-gatherers? to that of the Ache. It is noteworthy that both higher groups Both Hiernaux (1975a, b) and Cavalli-Sforza (1986) con- are tropical rainforest foragers. In this sense, as per Bailey et cluded that nutrition would not have been solely responsible al., such costs as searching and traveling for resources in for pygmies’ physical features. Nor do we think so, but, as small, widely dispersed patches are high in the tropical rain- mentioned already, the hardness of foraging life in the trop- forests, particularly for centrally placed foragers like the co- ical rainforest, i.e. the high cost paid to secure food require- operators in this survey staying in camps for a certain period ments, may have been a contributing factor to their short (Bailey and Headland, 1991). Yamauchi et al. (2000) report- stature and low body weight. For example, if a 25-year-old ed on the daily physical activity level of African popula- man weighing 60 kg and his twin weighing 50 kg lead their tions, including the Baka leading a village life in their lives at a PAL (TEE/BMR) value of 2.0, the former’s BMR sedentary settlement. Among more than ten populations, a is 1595.6 kcal and TEE 3191 kcal and the latter’s BMR farmer group in Gambia had the highest PAL value in both 1445.1 kcal and TEE 2890.1 kcal (see FAO, 2001: 37). The sexes: 2.02 for males and 1.97 for females, which are equiv- former needs 300 kcal of energy more than the latter. This is alent to that in this survey. In contrast, the Baka leading a equivalent to about 300 g in wild yam tubers. In tropical village life had a PAL value of 1.41 for males and 1.56 for rainforests, where the cost of exploiting resources is high, females, corresponding to sedentary or light activity lifestyle smaller food requirements could be of great advantage. Thus in the classification of FAO (2001). The hunting-gathering the small body weight of the pygmy would have been a like- life examined in this survey suggests a most active level of ly evolutionary development. daily life for human survival. Paleolithic hunter-gatherers Hiernaux (1975a: 117) thought that the pygmy’s physical without iron tools in tropical rainforests may have led a features were genetic in orgin and the period of 20000 years much more severe foraging life. indicated by prehistory would be long enough for selection to establish the present pygmy’s gene pool. Cavalli-Sforza Who were Paleolithic hunter-gatherers in the African tropi- calculated the separation between Khoisan, Mbuti, and other cal rainforest? African groups to be about 10000–20000 years based on 20 Finally we will address the question of who such Paleo- loci (Cavalli-Sforza, 1986: 414). By considering new genetic lithic hunter-gatherers in the tropical rainforest of the Congo information, he revised the separation time to be 18000 years Basin would have been. We believe they would most likely BP (Cavalli-Sforza et al., 1994). Recently it has been report- have been the ancestors of present pygmy groups for two ed that the separation between pygmy groups and Bantu reasons: one is that their physical features were unlikely to groups would have taken place less than 70000 years ago be established without close association with their environ- based on the analysis of mitochondrial (mt) DNA (Quintan- ment in tropical rainforests, and the other is recent genetic Murci et al., 2008). Furthermore resequencing 24 indepen- evidence. dent noncoding regions across the genome, Patin et al. Hiernaux argued that the physical characteristics of the (2009) inferred the separation time between pygmy hunter- Mbuti pygmy, shorter stature, low and wide nose and lighter gatherers and African farmers to be 60000 years BP and the skin color, represents their biological adaptation to the hot separation between the western and the eastern groups of and humid equatorial forest (Hiernaux, 1975a). Agreeing pygmy as 20000 years BP. As for the latter, elsewhere, the with Hiernaux, Cavalli-Sforza developed a convincing argu- separation time has been estimated at 18000 years BP based ment that a selective factor must have been whether the heat on mt-DNA analysis (Destro-Bisol, 2004). This genetic in- produced inside the body could be effectively dissipated in a formation indicates that the divergence between pygmy hot and humid environment and the pygmy’s short stature groups and other African groups was 70000–60000 years BP 20 H. SATO ET AL. ANTHROPOLOGICAL SCIENCE and the divergence of two Pygmy groups 20000–18000 a productive environment for human foragers? Human Ecol- years BP. If this is correct, it is likely that Paleolithic hunter- ogy, 19: 261–285. gatherers in the African tropical rainforest were the ancestors Bailey R.C., Head G., Jenike M., Owen B., Rechtman R., and Zechenter E. (1989) Hunting and gathering in tropical forest: of present pygmy groups considering their physical features. is it possible? American Anthropologist, 91: 59–82. Barnes R.F.W. and Lahm S.A. (1997) An ecological perspective on human densities in the central African forests. Journal of Conclusion Applied Ecology, 34: 245–260. This study, the observations on two 20-day controlled for- Brosius J.P. (1991) Foraging in tropical forests: the case of the aging trips in the minor dry season of August 2003 and the Penan of Sarawak, East Malaysia (Borneo). Human Ecology, 19: 123–150. major rainy season of October 2005 in southeastern Camer- Cavalli-Sforza L. (1986) African pygmies: an evaluation of the oon, found no evidence that it was impossible for human state of research. In: Cavalli-Sforza L. (ed.), African Pyg- beings to live independently of farmed agricultural products mies. Academic Press, Orland, FL, pp. 361–426. in a tropical rainforest environment. In both seasons a suffi- Cavalli-Sforza L., Menozzi P., and Piazza A. (1994) The History cient amount of food was constantly acquired, without any and Geography of Human Genes. Princeton University Press, food shortages, allowing the Baka cooperators to maintain Princeton. Colinvaux P.A. and Bush M.B. (1991) The rain-forest ecosystem their body weight during the survey period. Among more as a resource of hunting and gathering. American Anthropol- than 60 kinds of foods procured during the survey periods, ogist, 93: 153–160. wild yam tubers provided more than 60% of their estimated Destro-Bisol G., Coia V., Boschi I., Verginelli F., Caglíâ A., energy intake in both seasons, followed by game meat and Pascali V., Spedini G., and Calafell F. (2004) The analysis of wild nuts. These three major types of food supplied 90% of variation of mtDNA hypervariable region 1 suggests that their dietary energy. Wild yam tubers, especially the annual eastern and western Pygmies diverged before the Batu expan- sion. The American Naturalist, 163: 212–226. plant D. praehensilis were found to be an essential food to Dounias E. (1993) Perception and use of wild yams by the Baka enable a foraging life in the southeastern Cameroon rainfor- hunter-gatherers in south Cameroon. In: Hladik C.M., Hladik ests. The productivity of D. praehensilis in this survey area A., Linares O.F., Pagezy H., Semple A., and Hadley M. was sufficient for six to eight families of the Baka coopera- (eds.), Food and Nutrition in the Tropical Forest: Biocultural tors to maintain a foraging lifestyle for 20 continuous days, Interactions. The Man and the Biosphere Series, Vol. 15. with each marriage couple engaged in collecting yam tubers UNESCO, pp. 621–632. Dounias E. (2001) The management of wild yam tubers by the every 2 days. We also considered the implications of this Baka Pygmies in southern Cameroon. African Study Mono- study for the possibility of a Paleolithic foraging lifestyle ex- graphs, Suppl. 26: 135–156. isting in the African tropical rainforest. Dumont R., Hamon P., and Seignobos C. (1994) Les Ignames au In conclusion, we proved that the hypothesis of the ‘wild Cameroun. CIRAD, Montpellier. yam question’ could not be applied to all tropical rainforests. Dwyer P.D. and Minnegal M. (1991) Hunting in lowland tropical In the northwestern Congo basin, rich in wild yam tubers, a rain forest: towards a model of non-agricultural subsistence. Human Ecology, 19: 187–212. foraging lifestyle is possible although at a high energy cost Endicott K. and Bellwood P. (1991) The possibility of independent to secure sufficient food resources to maintain it. This im- foraging in the rain forest of Peninsular Malaysia. Human plies that a Paleolithic foraging lifestyle in the African trop- Ecology, 19: 151–185. ical rainforest may well have been possible, although not FAO (2001) Human Energy Requirements. Report of a Joint FAO/ easy, and that Paleolithic foragers may have been the ances- WHO/UNU Expert consultation. FAO, Rome. tors of the present ‘pygmy’ hunter-gatherers. Hamon P., Dumont R., Zoundjihekpon J., Tio-Touré B., and Hamon S. (1995) Wild Yams in West Africa: Morphological Characteristics. ORSTOM, Paris. Acknowledgments Harako R. (1976) The Mbuti as hunters: a study of ecological anthropology of the Mbuti Pygmies (1). Kyoto University We are extremely grateful to Dr Ngima of the University African Studies, 10: 37–99. of Yaoundé, Cameroon, for supporting our field surveys in Hart T.B. and Hart J.A. (1986) The ecological basis of hunter- Cameroon. We also express our gratitude to Mr Takanori gatherer subsistence in African rain forests: The Mbuti of Eastern Zaire. Human Ecology, 14: 29–55. Ohishi, Kyoto University for giving us useful documents. Headland T.N. (1987) The wild yam questions: how well could Special thanks go to our Baka friends in the survey area, es- independent hunter-gatherers live in a tropical rain forest pecially Mr G. Menyata, for their hospitality, kindness, and environment? Human Ecology, 15: 463–491. cooperation over the long period of this research. This study Headland T.N. (1997) Revisionism in ecological anthropology. was supported by Grant-in-Aid for Scientific Research (Nos. Current Anthropology, 38: 605–630. 06041046, 15405016, 20405014) from the Japanese Society Headland T. and Bailey R. (1991) Introduction: have hunter- gatherers ever lived in tropical rain forest independently of for the Promotion of Science. agriculture? Human Ecology, 19: 115–122. Hiernaux J. (1975a) The People of Africa. Weidenfeld and Nicol- References son, London. Hiernaux J. (1975b) Climate and the weight/height relationship in Bahuchet S. (1985) Les Pygmées Aka et la Forêt Centrafricaine. sub-Saharan Africa. Annals of Human Biology, 2: 3–12. SELAF, Paris. Hladik A. and Dounias E. (1993) Wild yams of the African forest Bahuchet S., Mckey D., and De Carine I. (1991) Wild yams revis- as potential food resources. In: Hladik C.M., Hladik A., ited: is independence from agriculture possible for rain forest Linares O.F., Pagezy H., Semple A., and Hadley M. (eds.), hunter-gatherers? Human Ecology, 19: 213–243. Food and Nutrition in the Tropical Forest: Biocultural Inter- Bailey R.C. and Headland T.N. (1991) The tropical rain forest: is it actions. The Man and the Biosphere Series, Vol. 15, Vol. 120, 2012 OBSERVATIONS ON TWO CONTROLLED FORAGING TRIPS IN THE CAMEROON RAINFOREST 21

UNESCO, pp. 163–176. 5th rev. and enlarged edn. Tokyo (in Japanese). Hladik A.S., Bahuchet S., Ducatillon C., and Hladik C.M. (1984) Patin E., Laval G., Barreiro L.B., Salas A., Semino O., Santachiara- Les plantes a tubercules de la foret dense d’Afrique centrale. Benerecetti S., Kidd K.K., Kidd J.R., Van der Veen L., Revue d’Ecologie (Terre et Vie), 39: 249–290. Hombert J., Gessain A., Froment A., Bahuchet S., Heyer E., Hurtado A.M. and Hill K. (1989) Experimental studies of tool effi- and Quintana-Murci L. (2009) Inferring the demographic his- ciency among Machiguenga women and implications for tory of African farmers and Pygmy hunter-gatherers using a root-digging foragers. Journal of Anthropological Research, multilocus resequencing data set. PLoS Genetics, 5(4): 45: 207–217. e1000448. doi: 10.1371. Ichikawa M. (1981) Ecological and sociological importance of Quintan-Murci L., Quach H., Harmant C., Luca F., Massonne B., honey to the Mbuti net hunters, Eastern Zaire. African Study Patin E., Sica L., Mougqiama-Daouda P., Comas D., Tzur S., Monographs, 1: 55–68. Balanovsky O., Kidd K.K., Kidd J.R., van der Veen L., Ichikawa M. (1983) An examination of the hunting-dependent life Hombert J., Gessain A., Verdu P., Froment A., Bahuchet S., of the Mbuti Pygmies, Eastern Zaire. African Study Mono- Heyer E., Dausset J., Salas A., and Behar D.M. (2008) Mater- graphs, 4: 55–76. nal traces of deep common ancestry and asymmetric gene Kingdon J. (1997) The Kingdon Field Guide to African Mammals. flow between Pygmy hunter-gatherers and Bantu-speaking Academic Press, London. farmers. Proceedings of the National Academy of Sciences, Kitanishi K. (1995) Seasonal changes in the subsistence activities 105: 1596–1601. and food intake of the Aka hunter-gatherers in northeastern Sato H. (1992) Notes on the distribution and settlement pattern of Congo. African Study Monographs, 16: 73–118. hunter-gatherers in northwestern Congo. African Study Kitanishi K. (2003) Cultivation by Baka hunter-gatherers in the Monographs, 13: 203–216. tropical rain forest of central Africa. African Study Mono- Sato H. (2001) The potential of edible wild yams and yam-like graphs, 28: 143–157. plants as a staple food resource in the African rain forest. Leonard W. and Robertson M.L. (1992) Nutritional requirements African Study Monographs, Suppl. 26, March: 123–134. and human evolution: a bioenergetics model. American Jour- Sato H. (2006) A brief report on a large mountain-top community nal of Human Biology, 4: 179–195. of Dioscorea praehensilis in the tropical rainforest of south- Letouzey R. (1976) Contribution de la Botanique au Problème eastern Cameroon. African Study Monographs, Suppl. 33: d’une Éventuelle Langue Pygmée. SELAF, Paris. 21–28. Letouzey R. (1985) Notice de la Carte Phytogeographique du Sato H., Kawamura K., Inai H., and Yamauchi T. (2006) A ‘pure’ Cameoun au 1:500000. Institute de la Recherche Agronomique foraging lifestyle in a tropical rainforest of southeastern Cam- (Herbier National), Toulouse. eroon: observations on a twenty-day trip of the Baka hunter- Leung W.W. (1968) Food composition table for use in Africa. gatherers in the short dry season. Journal of African Studies FAO Nutrition Division: US Department of Health Educa- (Japan), 77: 1–14 (in Japanese with English abstract). tion, and Welfare, Public Health Service, Bethesda. Sigaha-Nkamdiou (1993) Caractérisation et Fonctionnement Mckey D., Digiust B., Pascal L., Elias M., and Dounias E. (1998) Hydrochimique d’un Bassin Versant en Milieu Forestier Stratégies de croissance et de defense anti-herbivore des Équatorial Humide: L’exemple de la Ngoko Moloundou ignames sauvages: leçons pour l’agronomie. In: Berthaud J., (Sud-est du Cameroun). Thèse de docteur de l’Université Bricas N., and Marchand J.L. (eds.), L’igname, plante sécu- Paris XI, Orsay. laire et culture d’avenir. CIRAD, Montpellier, pp. 181–188. Stearman A.M. (1991) Making a living in the tropical forest: Mercader J. (2002) Forest people: the role of African rainforests in Yuqui foragers in the Bolivian Amazon. Human Ecology, 19: human evolution and dispersal. Evolutionary Anthropology, 245–260. 11: 117–124. Tanno T. (1976) The Mbuti net-hunters in the Ituri forest, eastern Mercader J. (2003) Foragers of the Congo: the early settlement of Zaire: their hunting activities and band composition. Kyoto the Ituri Forest. In: Mercader J. (ed.), Under the Canopy. Rut- University African Studies, 10: 101–135. gers University Press, Rutgers, NJ, pp. 93–118. Tanno T. (1981) Plant utilization of the Mbuti Pygmies: with spe- Mercader J. and Brooks A.S. (2001) Across forests and savannas: cial reference to their material culture and use of wild vegeta- later stone age assemblages from Ituri and Semliki, Demo- ble food. African Study Monographs, 1: 1–53. cratic Republic of Congo. Journal Anthropological Research, Vincent A.S. (1984) Plant foods in savanna environments: a pre- 57: 197–217. liminary report of tubers eaten by the Hadza of northern Tna- Mercader J. and Martí R. (2003) The Middle Stone Age occupa- zania. World Archaeology, 17: 131–148. tion of Atlantic central Africa: New evidence from Equatrial Vivien J. and Faure J.J. (1996) Fruitiers Sauvages d’Afrique Guinea and Cameroon. In: Mercader J. (ed.), Under the Can- (Espèces du Cameroun). Ministère Francais de la Coopéra- opy. Rutgers University Press, Rutgers, NJ, pp. 64–92. tion et du CTA, Paris. Mercader J., Runge F., Vrydaghs L., Doutrelepont H., Ewango White F. (1983) The Vegetation of Africa. UNESCO, Paris. C.E.N., and Juan-Tresseras J. (2000) Phytoliths from archae- Yamauchi T., Sato H., and Kawamura K. (2000) Nutrition status, ological sites in the tropical forest of Ituri, Democratic activity pattern, and dietary intake among the Baka hunter- Republic of Congo. Quaternary Research, 54: 102–112. gatherers in the village camps in Cameroon. African Study Mercader J., Garralda M.D., Pearson O.M., and Bailey R.C. (2001) Monographs, 21: 67–82. Eight hundred-year-old human remains from the Ituri tropical Yasuoka H. (2006) Long-term foraging expeditions (Molongo) forest, Democratic Republic of Congo: the rock shelter site of among the Baka hunter-gatherers in the northwestern Congo Matangai Turu Northwest. American Journal of Physical basin, with special reference to the ‘wild yam question.’ Anthropology, 115: 24–37. Human Ecology, 34: 275–296. Mialoundama F. (1993) Nutritional and socio-economic value of Yasuoka H. (2009) Concentrated distribution of wild yam patches: Gnetum leaves in central African forest. In: Hladik C.M., Historical ecology and the subsistence of African rainforest Hladik A.O., Linares O.F., Pagezy H., Semple A., and Hadley hunter-gatherers. Human Ecology, 37: 577–587. M. (eds.), Food and Nutrition in the Tropical Forest: Biocul- Yasuoka H. (2011) Ecological Anthropology of the Baka Hunter- tural Interactions. The Man and the Biosphere Series, Vol. 15, Gatherers: Rethinking Hunter-gatherer Subsistence in the UNESCO, pp. 177–182. African Rainforests. The Center for African Area Studies, Ministry of Education, Culture, Sports, Science and Technology, Kyoto University (in Japanese). Japan (2005) Standard Tables of Food Composition in Japan,