Introduction

Considerable strides have been made in understanding the patterns and causes of bushmeat consumption in west Africa [e.g., 1, 2] and central Africa [3,4], but studies in East Africa are rather less advanced (Kenya [5], Uganda [6,7], Tanzania [8,9]). In Tanzania, for example, research on bushmeat hunting, defined as wildlife hunted for human consumption [10], has focused almost entirely on a single area, the Serengeti National Park [11]. Studies conducted there show that illegal hunting occurs mostly in the dry season when the huge herds of migratory herbivores arrive [12,13]; that people hunt with dogs and snares [14]; that bushmeat is substantially cheaper than domestic meat [15]; that hunting is principally a means of generating cash although it is also used in household consumption [8]; that cultural and social requirements affect consumption [16]; and that people who have access to alternative incomes are less likely to hunt [8,17].

While some studies in East Africa acknowledge that different ethnic groups may exploit wildlife in different ways [e.g., 18], and also model the potential for zonal variation in wildlife exploitation [19], the vast majority of studies still treat bushmeat consumption as a uniform activity. This results in policies based on the assumption that the economic and ecological constraints of people living around one protected area are the same, and that one solution therefore fits all. Only one empirical study, again around Serengeti, has compared bushmeat hunting originating from multiple areas, at least as indexed by ethnicity [20]. It found that non-immigrants and households with above average livestock holdings are typically more involved in hunting than recent migrants and livestock-poor households. It also showed that ethnicity affected hunting preferences and reasons for hunting: the Maasai living to the east of the park preferred small-sized mammals and hunted only for protein, whereas farmers and foragers (comprising principally Natta, Sukuma, Ikoma, Kuria, Taturu and Ikizu ethnicities) to the west, preferred larger-bodied mammals and hunted additionally for income. It is not clear how general such comparative findings may be, especially since the Serengeti is a migratory ecosystem, has high wild mammal biomass, and has a “hard” western boundary with a very high human population density outside [21], but the diversity of threats around Serengeti suggested that a similar approach might usefully be taken elsewhere.

In an effort to make more general comparative statements about bushmeat consumption in rural Tanzania, and to draw the attention of Tanzanian authorities to bushmeat consumption as a national issue, we examined protein consumption in another area. We chose the Katavi-Rukwa ecosystem, an area of western Tanzania where illegal hunting is well documented [22] and may be driving wild mammal population declines [23,24], and we focused on two different communities. The specific aim of this paper is to document, through quantitative analysis, the similarities and differences in bushmeat consumption and hunting activities of different people living adjacent to the same protected area. This enables us to provide not only specific on-site advice but also general recommendations for reducing bushmeat consumption., By considering bushmeat hunting in other parts of Tanzania, we use our data to make general statements about bushmeat hunting in East Africa.

Methods

Study Site

This study was carried out in Mpanda District of Rukwa Region of western Tanzania in two areas around Katavi National Park (KNP) and Rukwa Game Reserve (RGR) that comprise the Katavi-Rukwa ecosystem. KNP, the third largest national park in the country [25], lies in the Rukwa Valley and is 4471 km² in area. KNP is managed by the Tanzania National Parks Authority (TANAPA), and photographic tourism is permitted within the Park. RGR (4323 km²) borders KNP to the southeast (Fig. 1) and is managed by the Wildlife Division (WD); tourist hunting is allowed there from July to December. Both areas consist of miombo woodland, dry forest habitat characterized by Markamia, Grewia, Terminalia, Syzygium, Acacia and Combretum tree genera [2627–28], but KNP additionally has four large floodplains that attract high concentrations of wildlife in the dry season [29]. The Katavi-Rukwa ecosystem receives 600–1200 mm of rain per year [30].

Fig. 1.

The location of the two sites (Katumba Ward and Mpimbwe Division) outside Katavi National Park and Rukwa Game Reserve. Msaginia Forest Reserve lies immediately north of Katavi National Park but its borders are ill-defined. Black dots show the location of the villages where interviews were conducted (except for Mpanda, the District capital). Inserts shows the location of the study area in Tanzania.

In Tanzania, wildlife hunting requires a permit issued to residents by district or regional game officers and to foreign hunters by the Wildlife Division. The vast majority of bushmeat hunting is carried out by local people without any license and is therefore illegal. Selling of bushmeat is illegal, too. Our study took place in communities outside two protected areas. The first, Katumba Ward, which lies north of Msaginia Forest Reserve that borders KNP on the north (Fig. 1), has an estimated population of 96,451 (Tanzania Ministry of home affairs, 2008) and is principally occupied by Hutu refugees from Burundi who practice intensive agriculture. Few native Tanzanians live here, as it was originally a forest reserve before becoming a refugee camp. Poaching of animals and forest products by refugees is often claimed to be rampant in the area [31]. The second area is Mpimbwe Division (estimated population 100,000, E. Carabine and M. Borgerhoff Mulder, unpublished data) which lies immediately to the south of KNP and RGR (Figure 1) and is occupied by native Tanzanian Pimbwe, Fipa and Sukuma (but not Hutu), who farm and keep livestock; like the Hutu, the Pimbwe and Fipa have a history of hunting and fishing [25]. The main species taken are impala (Aepyceros melampus), common duiker (Sylvicapra grimmia), warthog (Phacocherus africanus), buffalo (Syncerus caffer), bushbuck (Tragelaphus scriptus), bushpig (Potamochoerus porcus) and zebra (Equus quagga), and also include another 12 medium to large-sized mammals [24].

Results

Importance of bushmeat

Our study revealed a significant difference in the reported consumption of bushmeat between the two communities (Table 1). Eighty-five percent of respondents in Mpimbwe and 52% in Katumba reported that they used bushmeat (n = 193 respondents, X² = 24.62, df =1, P <0.0001). Amongst consumers, bushmeat was reported to be consumed more frequently in Mpimbwe than in Katumba (n = 141 respondents, X² = 22.77, df =1, P <0.0001). Respondents also reported the use of a variety of other kinds of protein, with inhabitants of Mpimbwe showing significantly lower consumption of eggs and pork (Table 2).

Seventy-five percent of respondents in Mpimbwe and 42% of respondents in Katumba claimed that hunters hunted principally for selling to others rather than obtaining food for themselves. This was corroborated by the finding that only a very small proportion of those who reported consuming bushmeat also reported that they hunted themselves (2% and 0% from Mpimbwe and Katumba respectively); more commonly, consuming households said that their meat came from middlemen (26%, 4%), from hunters (35%, 36%), or from both hunters and middlemen (37%, 60%) in Mpimbwe and Katumba respectively. In Mpimbwe most of the bushmeat was sold in the village (as reported by 85%) or in nearby villages (77%), or outside the area (43%, n=120 respondents in all cases). In Katumba bushmeat was sold within the refugee settlement (81%), in nearby villages (68%), or away from the Katumba refugee settlement area (16 %, n=73).

Table 1.

Number (and percentage) of respondents reporting rates of bushmeat consumption in Mpimbwe (n=120 respondents) and Katumba (N=73 respondents).

Price of bushmeat

Despite the similarity in income in the two sites (mean household monthly income was 25,100 Tsh (US$20.08) and 24,200 Tsh (US$19.36) for Mpimbwe and Katumba respectively), bushmeat was more expensive to purchase in Katumba than in Mpimbwe (Table 3), in both absolute terms and relative to domestic meat prices. The reasons for this difference are not well understood, since there is no apparent shortage of protein in Katumba that might drive up prices for both domestic meat and bushmeat due to scarcity; indeed, Table 2 shows a greater consumption of pork and eggs in Katumba. Greater variations in wealth in Katumba than Mpimbwe could be forcing Katumba meat prices up as a result of the rich monopolizing the meat supply and thereby escalating demand in relation to supply, but the data suggest wealth variation is actually greater in Mpimbwe than in Katumba (SDs = 16,300 Tsh (US$10.80) and 8200 Tsh (US$5.40) respectively).

Table 2.

Number (and percentage) of respondents in Mpimbwe (n = 120 respondents) and Katumba (n = 73 respondents) reporting consumption of different sources of protein in 2008 and 2009.

Pattern of hunting

Hunting was reported to occur at different times of year in the two areas. Mpimbwe residents were reported to hunt more in the wet season than in the dry season (73% and 13% respectively) whereas Katumba residents hunted more in the dry season than wet season (45% and 30% respectively, (n=193 respondents X² = 31.09, df =1, P <0.0001). Fifty-four percent (n=32 respondents who mentioned wet season) in Mpimbwe reported that wet season hunting occurred because patrols cannot enter the area due to dirt roads being impassable, although 37% also mentioned that thick bushes help to hide poachers from rangers and game scouts; other reasons mentioned included ease of locating animals near farms, and economic stress during the rainy season. Katumba residents principally listed thick bushes (59 %) and seasonal economic hardship 41 % (n=22 respondents who mentioned wet season) as major reasons for hunting in the wet season. As for the dry season in Katumba, 66% mentioned it is because there is no rain in the bush and 18% because of no farming activities (n=33 respondents who mentioned dry season). Responses regarding whether hunting occurs at night or during the day, hunting techniques, and location of hunting activity showed no significant differences between the two areas.

Table 3.

Modal, minimum and maximum meat prices in Tanzania shillings in the two areas.

Discussion

Our study extends the number of studies of bushmeat consumption in Tanzania and importantly provides comparative data on two groups “mining” the same wildlife populations from the same ecosystem [see also 32,33]. Our principal conclusions are that bushmeat consumption is significant in this part of Tanzania, but that the pattern of consumption differs between areas. Ancillary conclusions – that most bushmeat consumed comes either directly from hunters (for cash) or through middlemen, and the low number of consumers who report that they themselves hunt - may reflect a preference among consumers to avoid disclosure of their own hunting exploits. Such natural reluctance emphasizes the utility of focusing on consumption rather than harvesting in bushmeat research; although both the harvest and the consumption of bushmeat are currently illegal in Tanzania, the former is associated with more stigma and fear of arrest than the latter, and is therefore far more likely to be underestimated. While our sample sizes are modest, representing only 2% of households in Mpimbwe and an estimated 0.8% of people in Katumba (assuming 10 people per household), we are confident that our focus on consumers provides a robust indication of the scale and the variability in western Tanzania's bushmeat problem, corroborated by a different but parallel study that focused on Mpimbwe hunters themselves [24, 34].

Our comparative data show that in some respects, the nature of illegal hunting is very similar between the two areas that we sampled. Mpimbwe and Katumba residents hunted using the same sort of weapons, muzzle loaders; and most of the hunting was carried out in KNP. In addition, residents of both areas are in agreement that hunting was declining. However, there were major differences between the two sites surrounding the Katavi-Rukwa ecosystem in that more than four-fifths of the population of Mpimbwe ate bushmeat compared to about half of the population of Katumba. Furthermore, the frequency of eating bushmeat was higher in Mpimbwe than Katumba. This is interesting in its own right, as it contradicts many informal (and often prejudicial) statements made by Tanzanians about Hutu refugees being the most important threat to wildlife in western Tanzania.

There are at least four possible (and non-exclusive) explanations for the difference in bushmeat consumption between these areas: (i) people in Katumba do not like bushmeat as much; (ii) there is a shortage of alternative protein in Mpimbwe; (iii) people in Katumba are poorer than Mpimbwe, or (iv) there is a shortage of bushmeat in Katumba. The first reason addresses cultural preferences, but we have no evidence that the people of Katumba are culturally less predisposed to eating bushmeat than the Pimbwe; to the contrary, interviews suggested the residents of Katumba like bushmeat very much, and this preference may indeed be driving up the price. Furthermore, Hutu are renowned for eating bushmeat in Burundi [35]. The second reason (shortage of alternative protein in Mpimbwe) is supported by the comparatively low intake of pork and eggs in Mpimbwe, and long term studies reveal that Pimbwe people are indeed short of domestic protein: they have insufficient cash to buy goats and their chickens die of disease [36]. While bushmeat is cheaper than domestic meat in Mpimbwe, it is typically only land-wealthy households in Mpimbwe who have the cash available for bushmeat purchases [22]; bushmeat is a luxury good, by which we mean not that its consumption is rare, but that it is enjoyed primarily by the wealthy.

The third and fourth reasons may also contribute jointly to the greater consumption of bushmeat in Mpimbwe than Katumba. While Katumba households are no poorer on average (in terms of household income) than Mpimwbe households, the relative cost of bushmeat (compared to domestic protein alternatives) is higher in Katumba. One possible reason for the price difference is our fourth reason - scarcity of supply. Katumba residents live farther from high densities of wildlife than do people in Mpimbwe Division. The former have to leave Katumba and travel on foot or bicycle a distance of 40 km through Msaginia Forest Reserve, which has low densities of large mammals [36], before they enter KNP. In contrast, Mpimbwe residents live much closer to high wildlife densities in KNP (2–19 km) – an easier walk or bicycle ride – and a parallel study indicates that more wild animal carcasses enter Mpimbwe villages situated closer to protected areas [22]. The effort to acquire bushmeat is therefore greater in Katumba. This is a case study in the benefits of national parks having a buffer zone (Msaginia Forest Reserve) rather than a hard boundary (Mpimbwe, which abuts the National Park directly).

We documented that Mpimbwe residents hunted principally in the wet season because park patrols were unable to conduct effective patrols there: swollen rivers prevent them from reaching the southern boundary of KNP. In the wet season illegal hunters have virtually unlimited access to the animals in KNP. Here our recommendation is to increase KNP investments in wet season patrols, including starting regular foot and bicycle patrols. Katumba residents, on the other hand, hunted all year, although slightly more in the dry season, when patrols by TANAPA and the Wildlife Division are more frequent. Our recommendation is to patrol along well-worked routes that poachers use from Katumba to KNP. Our study did not find noticeable differences in the species targeted or where hunting occurred (data not shown), but it did uncover marked differences in consumption between Katumba and Mpimbwe, which we attribute to the relative scarcity of bushmeat at Katumba, which drives higher prices. To our knowledge, only one other study has compared bushmeat consumption of different populations living around the same national park in Tanzania. Mfunda and Roskaft [20] compared Ikoma and Natta hunter-gatherers, and Sukuma, Taturu and Ikizu agro-pastoralists (all in the west), with Maasai pastoralists to the east of Serengeti. They found that Ikoma, Natta and Maasai were more involved in hunting than other groups. Maasai preferred to hunt small species around villages and bushmeat constituted just a small proportion of protein for Maasai; in contrast, the other ethnic groups hunted large to mid-sized species in more heavily protected areas, and bushmeat made up nearly two thirds of their protein intake. Despite these differences, the authors make similar recommendations for both zones: more investment in community-based conservation schemes.

At least ten studies of bushmeat hunting have been conducted in Tanzania (Table 4) using different methodologies. Arrest records, self-reporting about illegal activities, and asking about bushmeat consumption can lead to different conclusions, although the first and last methods are sometimes congruent [41]. Despite differences in methodologies, however, current information shows that Tanzanians illegally hunt a large variety of mammalian species, ranging in size from elephants Loxodonta africana to elephant shrews; that they use snares, traps and guns to catch wild animals; and that they hunt for both direct consumption and for selling meat. In some places, bushmeat constitutes the chief source of animal protein. These studies indicate that bushmeat consumption is commonplace in Tanzania and supports a growing awareness in the country that illegal hunting is rife [e.g., 38,43] and constitutes a threat to wildlife populations in some areas [23,35]. Nonetheless, Table 4 shows that illegal hunting in Tanzania appears principally geared for local consumption either through local sales or by the hunter and his family. While there is some evidence of sale of meat to nearby cities, for example, such as Mpanda in our study (J. Gara, Personal Communication 2009), this does not seem to occur on such a large scale as in West Africa [e.g., 44], and certainly little appears to be exported abroad [45].

Table 4:

Summary of bushmeat studies in Tanzania.

Implications for conservation

Two factors affect patterns of bushmeat consumption in this area of western Tanzania: a scarcity of wild mammals near Katumba, and a shortage of affordable domestic protein in Mpimbwe. We concur that ecological and economic rather than cultural factors are responsible for the observed differences between sites and ethnic groups [46]. On a practical level, given the availability of bushmeat in Mpimbwe and the limited availability of alternative sources of protein, we support recent interventions that encourage raising chickens and pigs there. In Katumba, bushmeat consumption can be lowered by reducing the price of domestic meat (through subsidies or enhanced production). In both areas, but especially in Mpimbwe, increasing protection activities, which are the Government mandated responsibility of TANAPA (KNP) and the Wildlife Division (RGR), would help to reduce poaching. In the longer term, education about the value of ecosystem services provided by the Katavi-Rukwa protected areas may help. In addition, future revenues from tourism and controlled exploitation under proposed Wildlife Management Areas outside these protected areas might increase appreciation of the environment.

By Tanzanian law, killing large and medium-sized wild mammals is illegal without a license, but a number of studies now show it to be widespread in the country. From a practical standpoint, Tanzania needs to tackle the problem of bushmeat consumption using several strategies in tandem [47]: increasing livestock production so as to reduce the price of domestic meat [48], increasing protection of protected areas through a greater commitment to on-the-ground enforcement, and through education. The importance of these implementation strategies is likely to vary by location as shown by our comparative study around KNP and RGR.